PDA

View Full Version : Is the prop-efficiency component of thrust modelled as a constant in IL-2?



Jetbuff
02-13-2006, 05:10 PM
The answer would be much appreciated from anyone in the know. Does propellor thrust in the game's FM vary with speed or is it constant?

Edited thread title to better reflect current state of the discussion, namely: is a dynamically calculated, speed dependent value for prop-efficiency factored into the determination of thrust?

Bearcat99
02-13-2006, 05:49 PM
I would imagine that would be a matter of pitch and engine RPM. Just what do you mean by prop thrust... If the pitch is more co**** then the thrust will be greater... LOL.. the filter took out co****... probably because it thought i ws saying ****.. LMAO... this cracks me up....

WWMaxGunz
02-13-2006, 05:50 PM
Tricky question that one.

I might say:
With the different kinds of props, how do you know?

And that may not be right to even ask as I'm not sure it addresses the question.
IE, I'm a bit stumped what's really involved.

You want to do some level acceleration runs and devicelink the data then spreadsheet?
Would that tell you what you want?

WWMaxGunz
02-13-2006, 05:55 PM
BC, props are wings so there's prop AOA IRL prop thrust. Matching revs to speed is serious
business though with CSP type self-feedback systems it's got to be easier than without.

There's also things like prop tips going supersonic with high revs at very high speed but
is it modelled as well? I wouldn't put it past Mr. Maddox and Team, not at all.

Jetbuff
02-13-2006, 09:42 PM
BC, a prop, as Neal pointed out, is a set of aerofoils. Its efficiency varies depending on AoA, (pitch angle relative to relative airflow) rpm and the aircraft's forward velocity. The first two are covered in a CSP system allowing for a much flatter thrust curve. By varying the pitch of the blades/rpm it can maintain more constant thrust levels throughout a wider range of speeds than otherwise possible.

Still, even a CSP, cannot produce constant thrust irrespective of aircraft speed. Though I'm fuzzy on the theory so I'm probably not the person to explain it, I'll give it a shot. A helpful term to look up though would be 'prop advance ratio'. This is the ratio of how much air the prop 'bites through' versus the airplane's forward speed and is often encountered in discussions about prop efficiency.

From what I understand, prop efficiency drops off both at very low speeds and very high speeds. At low speeds, problems appear to arise out of turbulent inflow. It is very reminiscint in my mind of airflow separation over an aerofoil.

At really high speeds, you actually reach a point where angle of attack is too low, even at max blade pitch, relative to the inflow and you actually can end up with negative angle-of-attack and negative thrust. (think windmilling prop even with the engine on)

Neither transition is abrupt from what I understand, and props have a range of maximum efficiency values corresponding to various aircraft speeds. Where exactly max efficiency is achieved and for how wide of a speed range it can be sustained is dependent on the prop design as well as the mechanical engine rpm and prop-pitch limits.

Now, if thrust is modelled as a constant in IL-2, then fudges in other parameters might be necessary for outwardly realistic aircraft performance. e.g. to reach for example the published values for top speeds of a real aircraft, parasitic drag may have to be slightly increased to account for higher average thrust or the thrust may itself be lower than the max achievable by that plane. In addition to the adverse effects of these fudges on the less obviously contested areas, (like E-bleed, optimum glide speeds and angles, etc.) the constant thrust itself would possibly significantly influence aircraft performance at the speed extremes. e.g. all planes might have higher than expected acceleration at low speeds or it may detract from the advantage of heavier planes at high speeds skewing the dive/zoom equation more in favour of high thrust-to-weight ratio aircraft than expected.

For a graphic example:
Take the P-47 and the Spitfire. P-47 is a lot heavier, so like a big rig, once it gets going it's harder for the air to stop it. (F=m*a where a=drag) Yes, the drag on the P-47 is higher, but I doubt the ratio is higher than that for the weight difference. Finally, the P-47 also has a poorer thrust-to-weight ratio too so it takes a lot to get it going in the first place. In contrast, the spitfire is like a sports car. It can do 0-60 a lot faster and can stop on a dime. So far so good.

Given that, one would assume that in a dive, the Spitfire will first pull away quite significantly and then the P-47 will quickly catch up to it and overtake it. However, if thrust remains constant, even after the Spit reaches high speeds (where the real plane would see it's available thrust fall) it still has the benefit of its higher thrust-to-weight ratio while the P-47 is handicapped by its numbers. Furthermore, the same applies to the first part of the dive if commenced at a speed where the prop would otherwise be at less than peak efficiency. Constant thrust in those areas would emphasize the advantage of a higher thrust-to-weight ratio beyond other considerations.



Caveat: my grasp of the details is tentative at best at this time and thrust variation based on airspeed may already be modelled. Since the former will take me some time to wrap my head around, I thought I'd at least verify that it was something worth pursuing in the first place.

Of course, some more learned aero-buff could save me the trouble and demonstrate the probably obvious mistakes I've made, sending me to stand in the corner! :vhappy:

Jetbuff
02-15-2006, 12:36 AM
Anyone?

Viper2005_
02-15-2006, 05:35 AM
Prop thrust varies with speed. It's got to, otherwise you break the laws of physics quite impressively.

Power = thrust * velocity

1 hp = 1 lbf * 550 ft/sec (or 375 mph if you prefer)

This is almost certainly simulated as otherwise we'd see <span class="ev_code_red">massive</span> performance errors.

The only way of "fudging" things otherwise would be to tweak the drag curves so that parasite drag increased with the cube of IAS instead of the square of IAS.

The consequences of this would be testable; all you've got to do is kill your engine, fly constant IAS and measure rate of descent for a variety of speeds and aircraft. You can then plot rate of descent against IAS. If you get a quadratic relationship then all is right with the world. If you get a cubic relationship then something fishy is going on!

I suggest airstarting in QMB with 25% fuel (the lighter you are the lower your minimum drag speed will be in relation to the structural limits of the aircraft). Set radiators open and prop pitch 100%; this will increase your parasite drag and thus reduce your minimum drag speed such that your data should more closely approximate the ideal case (pure parasite drag).

At present I am rather too busy to carry out such testing (I am concentrating on the level speed performance of the Fw-190, and on revising for my PPL Nav exam). But it's simple enough if you're prepared to put the effort in.

It's probably best to measure the time it takes you to descend through a set height (1 km for example) as your rates of descent will mostly be off the clock as far as cockpit instruments are concerned.

Since this is effectively all about parasite drag you should probably start at 300 km/h IAS and work up in increments of 100 km/h IAS until things start to come unglued. For the higher speeds you may need to airstart a couple of km above your starting "gate" altitude in order that you're established on a constant IAS profile for the entire measuring period.

BTW: If you will use such c*o*a*r*s*e* language Bearcat you've got to expect censorship! http://forums.ubi.com/images/smilies/88.gif

WWMaxGunz
02-15-2006, 05:53 AM
Jetbuff, which has the higher recorded dive speed, a Spitfire or a Lightning?

Okay, it *was* a late model Spitfire.

Stafroty
02-15-2006, 05:58 AM
almost made too fast thinking http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif how constant it is in real life? is the friction eating the speed at certain speeds? if there would not be drag or friction, plane would go as fast as pilot would have balls or engine and gearing would last, or how long propellor would stay in one piece before it tears apart by forces i dont know to say in english http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif

DaimonSyrius
02-15-2006, 06:22 AM
Originally posted by Stafroty:
if there would not be drag or friction

If there were no drag, no friction, the propeller just wouldn't work at all. As in it couldn't propel anything forwards http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

The propeller needs air to work on, and air has viscosity, it needs to create friction. Without it (if we imagine an 'ideal' air) the propeller blades would just turn through air without pushing the molecules, it wouldn't transfer energy to them (I believe).

Cheers,
S.

P.S.: Jets would be a different thing, of course, like in outer space... No friction (or extremely low).

Stafroty
02-15-2006, 06:38 AM
pretty damn good point man http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

Jetbuff
02-15-2006, 12:34 PM
Originally posted by Viper2005_:
Prop thrust varies with speed. It's got to, otherwise you break the laws of physics quite impressively.
Agreed.

This is almost certainly simulated as otherwise we'd see <span class="ev_code_red">massive</span> performance errors.
I disagree. If thrust were modelled as a constant (not power, thrust, i.e. the effective force causing acceleration) it could theoretically be possible to match most of the obvious performance parameters relatively well by manipulating drag coefficients, aircraft weight, etc.

The only way of "fudging" things otherwise would be to tweak the drag curves so that parasite drag increased with the cube of IAS instead of the square of IAS.
Do you know this for a fact? I don't know that either (a) it is necessary to have a cubic relationship (why not just higher base drag coefficients?); nor (b) if we do have a proper squared relationship in IL-2. The latter would require some exhaustive testing at multiple speed points.

Just to be clear, I am not saying you're wrong and that thrust is indeed a constant - afterall, that's the whole point of this post, to find out whether or not it is. What I am saying is that your argument, whether through lack of understanding on my part or yours, doesn't have me convinced yet.

NonWonderDog
02-15-2006, 01:41 PM
Thrust should be (almost) constant for the jets, power should be (almost) constant for the prop planes. This is freshman stuff, there's almost <span class="ev_code_RED">ZERO</span> chance of a professional engineer such as Oleg getting that wrong.

Again, there is no chance of performance figures being even close to realistic if this is wrong... but noone who passed freshman aero classes would get this wrong.

Jetbuff
02-15-2006, 01:43 PM
Not getting it wrong NonWonderDog, but perhaps it was a necessary compromise for performance concerns. Afterall, the game has to run on relatively low hardware capabilities. e.g. gyroscopic effects were only recently added, was that an error or just an increased fidelity of the FM thanks to increasing hardware capabilities?

Viper2005_
02-15-2006, 02:00 PM
Originally posted by Jetbuff:
Just to be clear, I am not saying you're wrong and that thrust is indeed a constant - afterall, that's the whole point of this post, to find out whether or not it is. What I am saying is that your argument, whether through lack of understanding on my part or yours, doesn't have me convinced yet.

If we had higher base drag coefficients then things would come seriously unglued at low speed; we'd have way too much drag. The larger the IAS range of the aircraft concerned, the bigger the problem would be. The Mustang III would be hit the hardest since it's the fastest non-jet in the planeset (jets have pretty constant thrust at any given altitude when flying at subsonic speed).

As I said in my original post, flight testing would be required to ascertain the shape of the drag polar for a variety of aircraft; this would be a time consuming business and might easily take 10 hours of hard work.

But my experience of general flying in the game would lead me to believe that it's probably got a reasonably accurate drag relationship.

At the end of the day the amount of effort required to fudge things would probably be greater than that required to just put in some accurate mathematics. As for performance, I would guess that the AI and graphics are the most expensive items on the menu...

AFAIK the only major problem with the FM is that it doesn't simulate Mach number effects.

Jetbuff
02-15-2006, 02:06 PM
Ah, but at low speeds drag would be much less of an issue, quadratic relationship and all. i.e. it can be overcome, or the discrepancy would be too small to matter materially a lot of the time. Then again, I could climb the Spit IXe, K-4 and La7 at below their level stall speed at an appreciable rate.

Your argument though that it the fudges might be more trouble than coding it correctly in the first place does make a lot of sense though. Still would be nice if someone from Maddox Games would state definitively that the necessary thrust modulation based on speed is included eh?

Viper2005_
02-15-2006, 03:15 PM
It'd certainly save a lot of testing!

Jetbuff
02-15-2006, 03:42 PM
For those who are still not on board, this is what I'm talking about:
http://www.epi-eng.com/Prop-Selection.htm

Props usually provide their best thrust (due to highest efficiency) at low speeds. This thrust drops off above and below this speed and CSP systems maintain the plateau of maximum efficiency over a wider speed range. My question is, is this modelled in IL-2's FM or do we have a constant thrust value throughout the speed range?


Viper, in the event that Oleg is a sadistic bastage ( http://forums.ubi.com/images/smilies/16x16_smiley-wink.gif ) and we don't get the answer we're looking for, how would we go about verifying the variability of thrust in the FM?

PS: I thought thrust from jet engines also varied considerably even at subsonic speeds? My understanding was that jet engines generally tended to provide their best thrust at much higher speeds than the average prop. Am I wrong?

Viper2005_
02-15-2006, 04:07 PM
Your graph shows prop efficiency not thrust; they are very different things. Power = thrust * velocity.

If you want to throw in efficiency terms then that's great but you need to be careful as prop efficiency is often defined as a fraction of the ideal efficiency which is itself less than 100% (and varies as a function of disk loading).

Jets provide roughly constant thrust at subsonic speeds (actually it tends to fall off from 0-300 mph and then increase again from 300 mph up; again this is very rough and depends more than anything upon the installation).

I provided a test methodology for drag curve shape above. If the drag curve is the right shape then the thrust curve is most likely the right shape because otherwise considerable tweaking would be required. The easiest way to measure the thrust curve is to compare climb rates at different speeds with the drag polar at different speeds, remembering that energy is conserved.

Jetbuff
02-15-2006, 05:32 PM
But doesn't prop efficiency does correlate with thrust?

Prop efficiency = Ct/Cp*V (or something like that)

i.e. prop efficiency determines the rate of conversion of available power to usable thrust. At low efficiency, much of the blades' work (aka engine power) is wasted, no?

The-Pizza-Man
02-15-2006, 06:36 PM
Thrust for a prop is given by

F=0.5*r*A*(Ve^2-Vs^2)

r=air density, A=prop disc area, Ve=exit velocity, Vs=Velocity of the free stream(the same as the velocity of the plane)

As you can see thrust is zero when Vs=Ve. You'll actually reach a point before that where drag=thrust and you won't be able accelerate any further, aka top speed http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif . Jets can fly at higher speeds because their Ve is higher than for a propeller.

The way to test it would be to do what Viper said. Don't test thrust, test if drag increases faster than it should in order to fudge a constant thrust.

Jetbuff
02-15-2006, 07:34 PM
Originally posted by The-Pizza-Man:
Thrust for a prop is given by

F=0.5*r*A*(Ve^2-Vs^2)

r=air density, A=prop disc area, Ve=exit velocity, Vs=Velocity of the free stream(the same as the velocity of the plane)

As you can see thrust is zero when Vs=Ve. You'll actually reach a point before that where drag=thrust and you won't be able accelerate any further, aka top speed http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif . Jets can fly at higher speeds because their Ve is higher than for a propeller.
Thanks. I thought so.

The way to test it would be to do what Viper said. Don't test thrust, test if drag increases faster than it should in order to fudge a constant thrust.
Well, what if the fudge isn't in the rate of increase of drag but a base increase in Cdo. i.e. because thrust is higher than it should be, so is the drag coefficient.

Warning: wild assumptions follow:

To compensate and get proper Vmax for various aircraft, the Coefficient of parasitic drag is slightly higher inducing progressively higher drag (related to square of speed) until aprropriate Vmax is achieved. Of course this fudge would upset the relationships at other points of Vs too. However, several possibilities arise to mitigate this:

1. Thrust is still higher at very low speeds than it should be hence the extra drag is actually overcome by the higher thrust.

2. Other parts of the flight envelope may not be as affected by Cdo as Vmax.

3. Cdi, the other component of drag, is also manipulated to compensate.

4. Finally, weights, wing-area, etc. may be tweaked to get the desired results.

This kind of thing could explain some of the rather surprising behaviour at low speeds or the very low best glide speeds (Vbg) for all planes. Afterall, the Lagg got by with a 300kg diet for how many patches? http://forums.ubi.com/images/smilies/16x16_smiley-tongue.gif

Of course, I could be grasping at straws here and I'm not about to sell myself on the idea without passing it by some of the more knowledgeable folk here.

Along the latter lines, here's an idea:

Say we use a light plane like the P-11, which also has a fixed blade prop to boot, (narrow prop efficiency band) and dive it with invulnerability on to prevent it exploding, there might be a chance it will reach terminal velocity right? (no more acceleration)

At that point, thrust would likely be negligible (Vs probably having long exceeded Ve for that fixed prop) and drag would equal the force of gravity correct? (hence no acceleration) Now, assuming the mass of the plane is historically correct, we could substitute m * g = Drag and then solve the drag equation for Cdo. Now knowing Cdo could we not estimate thrust at various points during that dive?

Of course, just so I don't get locked into this single train of thought, what other factors might be influence E-retention for heavier planes?

WWMaxGunz
02-15-2006, 07:48 PM
Right. Go through all that to save on a floating point divide?

Jetbuff
02-15-2006, 08:29 PM
Neal, I know it seems unlikely. I've said as much. But then again, whereas these fudges would be static values, properly calculating prop efficiency and thence thrust dynamically may be quite the load on the CPU. A shortcut may have been deemed necessary. It may not even be a full shortcut - perhaps thrust is precalculated across the entire speed range beforehand.

Trust me, I would be extremely happy to be proven wrong by a simple statement from Oleg to the contrary. I'm too lazy to go figure it out on my own. http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif

The-Pizza-Man
02-15-2006, 09:29 PM
Along the latter lines, here's an idea:

Say we use a light plane like the P-11, which also has a fixed blade prop to boot, (narrow prop efficiency band) and dive it with invulnerability on to prevent it exploding, there might be a chance it will reach terminal velocity right? (no more acceleration)

At that point, thrust would likely be negligible (Vs probably having long exceeded Ve for that fixed prop) and drag would equal the force of gravity correct? (hence no acceleration) Now, assuming the mass of the plane is historically correct, we could substitute m * g = Drag and then solve the drag equation for Cdo. Now knowing Cdo could we not estimate thrust at various points during that dive?

Of course, just so I don't get locked into this single train of thought, what other factors might be influence E-retention for heavier planes?

I don't think you could solve for Cdo without another equation. If you were to do a power off dive, you would have Cdo=D/(0.5*A*r*V^2), we know V and D because the dive is at terminal velocity, we could assume that air density is what it normally is, but that still leaves us with the reference area. So Cdo=2*x/A. We need another equation to solve for Cdo and A. If we could solve that we could calculate what a simple thrust-velocity curve should look like and what it is in the game for the aircraft we tested.

Jetbuff
02-16-2006, 09:46 AM
I was afraid it was going to get even more complicated. http://forums.ubi.com/images/smilies/16x16_smiley-sad.gif

WWMaxGunz
02-16-2006, 09:47 PM
Originally posted by Jetbuff:
Neal, I know it seems unlikely. I've said as much. But then again, whereas these fudges would be static values, properly calculating prop efficiency and thence thrust dynamically may be quite the load on the CPU. A shortcut may have been deemed necessary. It may not even be a full shortcut - perhaps thrust is precalculated across the entire speed range beforehand.

Trust me, I would be extremely happy to be proven wrong by a simple statement from Oleg to the contrary. I'm too lazy to go figure it out on my own. http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif

S! Do I know you from back in the Delphi days? I had a bit of a hard time in 2000 and
sometimes I'm not sure of everything. Ah, you know me on SimHQ for sure.

Only proof will be from Oleg or a team member I'm afraid. Just what they do is rightfully
a company trade secret.

I could posit a small table per plane and interpolating values between points but really
as of CEM with variable pitch props... well not a tiny table but still doable and the
difference between output and longer detailed calcs would be finer than you could tell.
I could fit such a table in 1/4 k maybe and it would save on CPU load.

What can I say? I started coding on 3mz 8085 machines running interpreter basic to do
engineering and business code with a whopping 32k RAM. You learn to do tricks working
like that.

DaimonSyrius
02-16-2006, 11:07 PM
I have been trying to track down what Oleg has posted here lately about modelling of physics and departures thereof. Not an easy task... I mean collecting info isn't an easy task here, the signal/noise ratio in this place is appalling. http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

Besides the general background noise, Olegish (as he himself calls it http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif) is a noisy, or fuzzy, language by itself, and on top of that, there's the factor of his very understandable obscurity about details (for business and forum-sociology reasons), which are what we would like... Details.

I've come across an 'official answer' he posted after an uproar about some 'leaked info' http://forums.ubi.com/groupee_common/emoticons/icon_rolleyes.gif concerning the plans for the FM in BoB, that was in Dec 2004. It's quite detailed about how the changes, and particularly the development of inertial effects, should affect the way planes fly, at high/low airspeeds... but not really many specific details about which parameters are being 'cooked', so to say, and by how much. Again, I understand why they're not releasing every bit of info, more so considering the predatory environment around here http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif
http://forums.ubi.com/eve/forums/a/tpc/f/63110913/m...631080852#6631080852 (http://forums.ubi.com/eve/forums/a/tpc/f/63110913/m/6631080852/r/6631080852#6631080852)
I liked this bit:

Originally posted by Oleg_Maddox:
Everything will be even more precise and true to real life when speaking about aerodynamics, which in the end is the law that governs absolutely everything, even including changes of inertia in different modes of flight.

Also the much-mentioned (and seldom-quoted) bit on up-to-15% departures from 'correct' values for some parameters in some planes, in the course of 'tuning' performance; this was Aug 2005
http://forums.ubi.com/eve/forums/a/tpc/f/63110913/m...801013053#1801013053 (http://forums.ubi.com/eve/forums/a/tpc/f/63110913/m/8371080943/r/1801013053#1801013053)

And also this post commenting on accuracy of calculations for G-forces, in reply to the usual whining/moaning, on Nov 2005.
http://forums.ubi.com/eve/forums/a/tpc/f/63110913/m...411075832#7411075832 (http://forums.ubi.com/eve/forums/a/tpc/f/63110913/m/9421058732/r/7411075832#7411075832)
There was next to no specific data in that one, but it contains a couple of pearls like:

Originally posted by Oleg_Maddox:
Interesting input of the user that know nothing and just read these than know just bit more.
And a winner:

You has wrong logic. http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif

All in all, nothing actually new, of course, I had read them all before, but one gets a better picture when putting them together.

What I would like to know (as everyone would, I guess) is how detailed, how complete, up to what point in accuracy, the 'universe physics' is in IL2, and also which are the specific tweaks or allowances made, regarding that universal physics and regarding the specifications for individual planes, in order to 'tune' their performance. So far, my impression is that the 'universe physics' is made as accurate as it can get, but that more flexibility is allowed in the process of fitting the individual planes within that physics model, because the set of parameters for an aircraft is limited in number (amount of different parameters), so they have to be 'helped' in order to yield an acceptable performance.

Still, that's only my guess.

Cheers,
S.

Jetbuff
02-17-2006, 11:14 AM
Cheers for the Olegisms. One of my favourites:

Oleg Maddox wrote:
Don't tell me [over/undermoddelling discussions] won't happen - always has, always will.

User psychology is most often impossible to change. How soon they all forget what I have actually said, and instead misinterpret and misquote me to claim many things I have never said.
The man's a friggin' philosopher! http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif

But you hit the nail on the head, we are not 100% sure what shortcuts were taken (and where) in the game's aerodynamic model. Without that info, we can only speculate and not very accurately. All we do know for sure is that there are shortcuts and fudges to make the planes behave accurately inspite of the shortcuts.

I also respect Oleg's right to protecting the game's secrets, but I suspect that a simple yes/no answer to my original question does not really expose much.

Without that answer, it's a long uphill struggle to find out.

Holtzauge
02-17-2006, 02:07 PM
It is the prop efficiency modelling that is off. In this sim it is possible to get decent performance going slower than IAS for best climb. This was punished IRL and why there is always a table specifically stating best climb speed IAS in flight manuals.

We had a thread about this some time ago and climb tests posted in it and comparisons with real life showed that in addition to having to high absolute climb values across the board, that climb performance at slow speed is significantly overmodelled.

That prop efficiency is a complex issue can be deduced from the numerous NACA reports available on the subject. One problem with WW2 fighters was that as power incresed on a certain model, the ground clearance did not (same landing gear!). The restriction in prop dia made the disc loading Cp high. A high disc loading can to a certain extent be offset by increasing the solidity (width of the blades). Even so, most high perf prop fighters in WW2 had due to the high disc loading low efficiencies at low prop advance ratios (speeds) which reduced efficiency as the speed dropped below optimum climb speed.

This is not correctly modelled in the sim and this is why we see such remarkable prop hanging climbs! http://forums.ubi.com/images/smilies/winky.gif

Jetbuff
02-17-2006, 05:33 PM
I wouldn't go so far as to make such a definitive conclusion just yet Holtzauge. For one thing, for most props, particularly CSP-governor equipped ones, the speeds at which prop efficiency falls off are really low. Much lower than best climb speed.

Best climb speed has more to do with the total drag figures, as a sum of induced and parasitic drag. Whereas the former increases with angle of attack (and hence lower speeds), the latter increases with the square of velocity. Since both curves are exponential, there is a midpoint where the sum of both is lower than anywhere else on the speed envelope and that is best climb speed. I like to think of it as the most efficient conversion of thrust to lift. It is similar to how best glide speed is estimated except in the latter case thrust is non-existant.

Constant thrust, if that is indeed how it is modelled, may still induce the large deviations at low speeds such as incorrect Vbg (best glide speed) or Vbc (best climb speed) because to keep the easily measured performance parameters like top speed and climb rates in check, small fudges could have been made to the drag coefficients or aircraft weights. Coefficients of drag in particular could drastically alter the values of Vbg and Vbc.

On a side note, how do you figure Oleg arrives at these coefficients in the first place? I doubt they are readily available for all 300 or so FB aircraft. So I'm guessing at least part of it involves an estimation based on performance numbers. Would be interesting to see the methodology used.

NonWonderDog
02-17-2006, 05:56 PM
Bah, prop efficiency is off by some degree in every commercial sim. It can only ever really be "close enough" when you're modelling more than one plane -- especially when you probably don't even have precise prop shapes or the software to analyze them for any plane.

The approximation probably used is that prop efficiency is more or less constant (at a constant RPM) in a plane equipped with a CSP over a normal range of speed. At very low and very high speed, prop efficiency drops off sharply -- thrust becomes more or less constant instead. The exact point at which this happens is approximated... for every plane in the sim. It's just the way it has to be done.

The variation of propeller efficiency with CSP RPM, air density, Mach number, and whatever else is rather complex, too. I don't even know what the precise relation should be; I have no idea how well it's done in the sim.

WWMaxGunz
02-17-2006, 07:00 PM
If you know the mass, rolling friction and horsepower of a plane then you should be able
to get an idea of lowspeed prop efficiency just by the accel starting the takeoff roll.
It's not that low in small plane I've been in or we'd never have hit takeoff speed on
the runways we did!

As to what is modelled, can I assume that one piece of code speaks for all planes that
have their own sets of parameters? Not as in all are CSP or VSP or fixed but that once
the simulated mechanism has determined pitch then one routine uses the values of the
prop blades, the rpms, the forward speed, the air density and hopefully angle of the
prop disk to the air (we do have p-factor and note, angle of the disk not blade) to get
a value for thrust. As long as none of the factors changes, all or parts of the calcs
can be saved until a change occurs -- you only have to recalculate on a change -- or
the whole works can be tabled just as entire FM's have been tabled on very limited PC's
compared today. HOW is not the point as I sure don't know, WHAT is the point and from
some of the many effects we get I feel that most or all I've listed above do count.

Consider the 109's on manual pitch, what you can do with them that causes change. All
the other props and modes easily boil down to ways of setting blade pitch, a status kept
in the model that the thrust routine reads and uses along with individual parameters.

With all that, WHY would thrust be constant?

GR142-Pipper
02-17-2006, 10:03 PM
Originally posted by Bearcat99:
I would imagine that would be a matter of pitch and engine RPM. Just what do you mean by prop thrust... If the pitch is more co**** then the thrust will be greater... LOL.. the filter took out co****... probably because it thought i ws saying ****.. LMAO... this cracks me up.... You're right, that is funny.

GR142-Pipper

Jetbuff
02-18-2006, 01:15 AM
OK Neal, I just did something similar. Sort of a rough guide if you will.

I chose the P11 for this test because its fixed prop means that its efficiency band is very narrow and will fall off quickly very early; if any plane could show whether or not prop-efficiency governed thrust, this was it. That's why they developed variable pitch and constant speed props in the first place, right? http://forums.ubi.com/images/smilies/16x16_smiley-wink.gif

Hypothesis: If thrust is not modelled as a constant and falls off as it's supposed to, I would expect a full throttle dive to accelerate to terminal dive speed faster than a throttle closed dive. The terminal dive speed itself though, if it can be achieved, should not vary significantly between the two runs since that speed is highly likely to be well outside the efficiency range of this fixed prop.

Methodology: I dove the P11 twice from 12000m on the Crimea map, 100% fuel, to a point 3000m out. (~14? dive angle) First run was at 110% throttle, the second at 0% throttle. I had to turn invulnerability on and engine overheat off though to allow the test to proceed. Repeated thrice and best 2 runs selected based on control deviations during the dive.

Results: The full throttle dive yielded 640kph IAS at impact. Of note is that it did appear to be at or close to terminal velocity at this point based on poor acceleration over the last few thousand meters, but I'll have to look closer at the UDP output later to verify this.

The closed throttle dive in contrast only got to about 590kph IAS at around 4000m and then actually regressed to about 490kph IAS at impact.

That's a whopping 150kph IAS difference! Frankly, I never expected anywhere near this margin. However, I also found the quite abrupt regression in speed below 4000m just as surprising. I had intentionally used IAS because I felt it (at least partially) accounted for air density variation, much as drag does, unlike TAS. So, I had expected to reach a terminal velocity (possibly lower than for the first run) and stay there. Indeed, I paid close attention to my reruns of this dive just to be sure that I wasn't seeing things. Always the same thing.

Conclusion: Unless the extra drag from the windmilling prop on the closed throttle dive can account for the sizable difference in speeds achieved, thrust must have continued to be effective even after the maximal operating parameters of the prop were exceeded.

Caveats: Turning off engine overheat as well as the high Mach numbers achieved (definitely ran into compression on the power on test) could have seriously interfered with the validity of this rough test.

It is still not conclusive of course, but definitely warrants a closer look imo.

Comments?

NonWonderDog
02-18-2006, 02:53 AM
Originally posted by WWMaxGunz:
If you know the mass, rolling friction and horsepower of a plane then you should be able
to get an idea of lowspeed prop efficiency just by the accel starting the takeoff roll.
It's not that low in small plane I've been in or we'd never have hit takeoff speed on
the runways we did!

With all that, WHY would thrust be constant?

Actually, prop efficiency is nearly zero at zero speed. You won't really get any prop efficiency data at all from standing start accelleration. Prop efficiency should increase (almost) linearly to some maximum value during the takeoff run, then drop very slightly from that value as speed increases.

During the range in which prop efficiency is increasing, thrust is almost constant. This is odd, I'll admit. It's not exactly true, either... and I don't know if there is a physical explanation for constant thrust. As far as I know it's just what you get from the mathematics when you assume a linearly increasing prop efficiency.

Remember, though, prop efficiency itself is an approximation. It's just a lot easier to talk about advance ratio and power loss than to try to analyse a rotating airfoil. So there's never really constant thrust... it's just the approximation you have to use for takeoff if you want to use the idea of prop efficiency.

If you assume constant power propelling the plane at all speeds, you end up with a plane with infinite accelleration at zero speed. That's clearly wrong.


Unless the extra drag from the windmilling prop on the closed throttle dive can account for the sizable difference in speeds achieved,

Most likely. A windmilling prop is a HUGE source of drag. A transonic prop should also be a huge source of drag, but whatcha gonna do. Really, though, just the fact that non-csp equipped planes *exist* means that thrust can't be constant.

Holtzauge
02-18-2006, 04:38 AM
Originally posted by Jetbuff:
I wouldn't go so far as to make such a definitive conclusion just yet Holtzauge. For one thing, for most props, particularly CSP-governor equipped ones, the speeds at which prop efficiency falls off are really low. Much lower than best climb speed.

Best climb speed has more to do with the total drag figures, as a sum of induced and parasitic drag. Whereas the former increases with angle of attack (and hence lower speeds), the latter increases with the square of velocity. Since both curves are exponential, there is a midpoint where the sum of both is lower than anywhere else on the speed envelope and that is best climb speed. I like to think of it as the most efficient conversion of thrust to lift. It is similar to how best glide speed is estimated except in the latter case thrust is non-existant.

Constant thrust, if that is indeed how it is modelled, may still induce the large deviations at low speeds such as incorrect Vbg (best glide speed) or Vbc (best climb speed) because to keep the easily measured performance parameters like top speed and climb rates in check, small fudges could have been made to the drag coefficients or aircraft weights. Coefficients of drag in particular could drastically alter the values of Vbg and Vbc.

On a side note, how do you figure Oleg arrives at these coefficients in the first place? I doubt they are readily available for all 300 or so FB aircraft. So I'm guessing at least part of it involves an estimation based on performance numbers. Would be interesting to see the methodology used.

You are correct on the point that the prop efficiency does not fall off so quickly at lower speeds IF it was possible to lower the blade angle sufficiently at lower speeds. However, my point was that the prop dia was limied on most WW2 fighters due to the constant ground clearence but as the power increased the disc loading went up. A higher disc loading than optimal means that the prop blades have to operate at a higher AOA than optimal at low speed and hence at a detrimental section lift/drag ratio giving a lower efficiency. However, as speed build up the advance ratio moves into a regime where the prop can operate at a more optimal AOA and therefore the efficiency goes up again.

But whatever the cause, we are IMHO still seeing to good climb performance in the low speed regime.

WWMaxGunz
02-18-2006, 06:49 AM
Fixed props and as far as I know all props have a twist built into them that more than
accounts for differing radial velocity of points of the blade.

To put it more simply, twist in fixed props ensures that only part of the blade has an
effective AOA to produce most efficient thrust for any one rpm and forward speed. The
blades are made to give a range of speeds, one for best top speed would not have efficient
thrust or acceleration at low speed. Even the 2 and 3 speed props were like this, the
extra speeds just cut more AOA and allowed faster flight.

IIRC even CSP's have the twist. How could you fly if the prop mech failed otherwise?
it might not be so much but hey fixed prop planes that I know of didn't have much speed
range at all.

The modelled VSP and esp CSP props we have do seem to be very efficient and have very
wide pitch ranges though I wouldn't say 0 (flat) to 90 (feathered).

If efficiency is really near zero then how do you get a multi-ton fighter to scramble?
I've got a film with Spitfires pulling out nicely from stop. I've been in more than one
small plane that had to be braked hard to stay in place while pre-takeoff engine tests
(the runups) were done.

NonWonderDog
02-18-2006, 07:03 AM
Propeller efficiency is defined as the fraction of power that is converted from engine torque to thrust. Thrust is equal to engine power multiplied by propeller efficiency divided by velocity.

T = eta*P/v

At zero velocity and constant power, thrust would increase to infinity. That's impossible. To make this all make sense, efficiency also has to go to zero at zero velocity. Zero divided by zero is still not defined, so there is some maximum thrust that has to be measured and applied at zero velocity.

Since prop efficency increases (almost) linearly up to maximum at very low speed, thrust is (almost) equal to this constant maximum thrust at very low speed.

This is all really annoyingly approximate. There are much more accurate ways to describe everything, but they're not nearly as easy.

WWMaxGunz
02-18-2006, 07:18 AM
Okay yes I can see the math. However I think that in the face of observation it is the
formula that loses. http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif

I'd say efficiency is nowhere near peak at all but the efficiency I think of is torque
to output.

I have a thought here, you check me. A spinning prop even when the plane is stopped will
still generate wind will it not? So the velocity of the prop is to that air, not velocity
of the plane, hey?

NonWonderDog
02-18-2006, 07:27 AM
Yeah, a spinning prop still makes thrust at zero speed, however that speed is measured. It makes about the same amount of thrust from zero speed up to the speed at which full fine pitch is set for, is the strange thing.

The point I was trying to get across is that all the formulas associated with propeller efficiency simply do not work at very low speeds. You have to either assume constant thrust or do some heavy-duty aero analysis on the prop. You can neither measure nor apply anything related to efficiency from a standing start.

Jetbuff
02-18-2006, 12:12 PM
How about at high speeds? My understanding is that once the inflow's speed exceeds max rpm/co****st pitch ability to generate an angle of attack, thrust would fall to 0 and possibly even go negative for negative angles of attack.

I'll repeat my rough experiment above with a feathering prop to eliminate/reduce the drag difference. If anything, I would expect the feathered prop/closed throttle run to end up being faster in this scenario. Would that be a safe assumption?

NonWonderDog
02-18-2006, 05:43 PM
Eventually the blades will stall completely, and efficiency will drop sharply. That's at *REALLY* high speed, though. REALLY high speed, and low RPM. You'll get a runaway prop first, which will both counter the effect and kill thrust through transonic tip drag. I think I might have mispoken before and inadvertently said thrust was constant at very high speed; it should really drop to zero very quickly. This is all academic, though, maximum speed is in almost all cases limited by the drag on the plane, not the maximum advance ratio of the propeller.

I don't know about your test. If the prop is spinning at 10,000 rpm, it might be capable of producing thrust in the sim (efficiency depends on advance ratio, not airspeed directly). What you *should* run into is an incredible amount of drag from your prop going transonic, but I don't think that's modelled. Unless you turn the difficulty down, the engine will explode first. It seems that you should still get drag from the air spinning the prop up to 10,000 rpm, which would negate any possibility of producing thrust, but I don't really know.


Really, though, supersonic propellers are beyond my area of expertise. Take this all with a grain of salt. I'm not entirely sure how propeller efficiency is measured at near-Mach speed.

WWMaxGunz
02-18-2006, 11:15 PM
Originally posted by Jetbuff:
How about at high speeds? My understanding is that once the inflow's speed exceeds max rpm/co****st pitch ability to generate an angle of attack, thrust would fall to 0 and possibly even go negative for negative angles of attack.

I'll repeat my rough experiment above with a feathering prop to eliminate/reduce the drag difference. If anything, I would expect the feathered prop/closed throttle run to end up being faster in this scenario. Would that be a safe assumption?

Dive a CSP plane engine idle once with 100% pitch and once with 0% pitch.
0% pitch will co****n the prop as totally as possible.

Try the same with power but go from 70-80% at the start of the dive to 0% by the time you
reach 700kph which happens pretty quick at say 100% power. See how low the revs ever get.

WWMaxGunz
02-18-2006, 11:29 PM
NWD there's supposed to be something in there for mach effects as once we did get a short
answer in a dive speed and props discussion that compression effects are included in the
FM back then.

A few books I've read that went into dives and mach speed state that at some high speed
the prop disk is regarded as a solid disk of material the diameter of the prop for purposes
of drag calculations. It's one of the big reasons why no prop planes get to mach 1. The
.89 Spitfire dive report I had seen had the Spit not in a records competition but had been
at some near insane alt and stalled then lost it for a short while and got stuck in a dive
until he got low enough to slow down. There was a meteorology station or radio station and
some telemetry that allowed for later calculation of speed, alt and mach. Really high up
there's less drag, that's where the good early mach attempts occurred.

Wasn't there an Me-163 flight that hit 1000kph just before losing control? Was that Heinie
Ditmar or Hannah Reich that died in that one? Or another famous German test pilot?
There were many trials of 163 to hit or exceed 1000kph, the thing went into a violent nose
tuck and tumble right about there. I've seen the film more than a couple times, got it in
a box of tapes somewhere. No prop disk for brake, next problem is shadowed elevators and
center fo lift moving rearward.

Jetbuff
02-18-2006, 11:38 PM
Neal, varying rpm directly affects thrust. That is known. The question is, does X rpm give Y thrust at all speeds? It shouldn't.

WWMaxGunz
02-19-2006, 12:04 AM
Take a 109F-2 and run in manual prop pitch. Get rpm and speed steady at say 80% pitch.
Now change pitch to 60% and work the rpm to the same with throttle. What is your speed?
It is good to have a plane with such control. http://forums.ubi.com/groupee_common/emoticons/icon_cool.gif

OTOH I've noted since FB that I can fly say a P-51 at 95% power and achieve higher speed
at 2km alt with less than 100% pitch (really rpm). I've watched the nose go up a bit just
hitting the -5% pitch button after my speed was stable and plane trimmed at 100%.

Somewhere about 450-500kph at 90-100% power in the -5 I really need to be at under 100% rpm. I do practice at less than WEP most of the time.

Jetbuff
02-19-2006, 01:01 AM
I'll give it a go. Meanwhile, here's the graph from that rather surprising P11 performance. I've also redone it with other planes that have prop feathering available and the speed achieved with power on was significantly higher than that with the throttle closed and the prop feathered.

Anyway, here's the rather quirky graph for the P11:
http://img96.imageshack.us/img96/3465/p11dive8dv.th.gif (http://img96.imageshack.us/my.php?image=p11dive8dv.gif)

Things of note:

1. The throttle closed acceleration is actually higher at first?!

2. There is a very weird phenomenon occuring at the 4000m mark. I would have expected it to be compression, but I remember the plane did not visibly do anything other than decelerate. Furthermore, why does this only happen in the throttle zeroed run?

3. The gap between the two runs like I said is huge and when you consider the minimal impact of the supposed extra drag of the windmilling prop up to about 4000m, it makes even less sense.

4. I didn't notice it specifically, but once graphed it is obvious that even the second run suffered a decrease in speed. However, the pulsating rythm is consistent with the wild yawing and pitching motions encountered in the compressibility range.

NonWonderDog
02-19-2006, 01:27 AM
That's a bit odd looking. Air density increases at low altitude, of course, and drag force is proportional to air density -- but that doesn't entirely explain why it slows the way it does.

I'm not sure what it would look like, but I'd guess that the graph would make MUCH more sense converted to TAS.

I have no idea why the prop would ever give negative thrust, though. That makes no sense at all.

How'd you get her up to 40,000 feet, anyway? http://forums.ubi.com/images/smilies/53.gif

Jetbuff
02-19-2006, 01:59 AM
NWD, I got it up there with the FMB. I also did a TAS conversion - one that I'm not 100% happy with though - and it changed little apart from emphasizing the regression. I'll post it if I can get a better conversion algorithm. The one I'm using seems far too simplified.

WWMaxGunz
02-19-2006, 11:40 AM
Originally posted by NonWonderDog:
Yeah, a spinning prop still makes thrust at zero speed, however that speed is measured. It makes about the same amount of thrust from zero speed up to the speed at which full fine pitch is set for, is the strange thing.

The point I was trying to get across is that all the formulas associated with propeller efficiency simply do not work at very low speeds. You have to either assume constant thrust or do some heavy-duty aero analysis on the prop. You can neither measure nor apply anything related to efficiency from a standing start.

Makes sense to me when I consider that v is the velocity of the incoming air which very soon
for a plane sitting still but engine going is nothing like zero. Consider the power of the
propwash, the velocity of propwash behind a braked plane doing the engine runups before taxi.
You cut the rudder hard and loosen the brakes, does the plane turn? Oh yes. Does the wash
hit the same speed as 'the speed at which full fine pitch is set for'? That's very testable
for anyone with access to even a small plane and an anemometer (sp? wind guage).

If so then the only weak point of the theory is the refusal to consider the prop bootstrapping
itself by directing airflow through it regardless of relative velocity to the ground.

WWMaxGunz
02-19-2006, 12:03 PM
Jetbuff I see you've graphed speed vs alt and wonder about the curves.
What you get if you graph speed against time instead?

Ugly_Kid
02-19-2006, 12:54 PM
I don't understand why one even asks this question? Is there any particular hint it was modeled as a constant?

Jet has a constant thrust - so 262 should have a constant thrust for example.

Prop aircraft has a thrust which decreases with speed - this regardless of prop efficiency. Prop efficiency comes on the top of this.

Now these two factors are one reason why early jet pilots regarded acceleration of 262 as lousy. In a prop plane you have very high thrust in low speed so it accelerates very fast initially - towards the top speed the acceleration then decreases rapidly. Jet is slow to accelerate but the acceleration drops less rapidly towards the top speed...

Now since the game models two parameters climb rate and top speed somewhat correctly I don't see how the thrust could be constant? If it was and one would reach correct top speed the climb would be too slow.

Jetbuff
02-19-2006, 01:36 PM
Anyone have a reliable IAS to TAS conversion formula? The one I used was off the net and I find it does not match the in-game TAS dial. Alas, I don't think we can actually extract TAS out of the DeviceLink.

Here's a graph against time though Neal, not sure what it adds though:
http://img52.imageshack.us/img52/5355/p11dive1ta.th.gif (http://img52.imageshack.us/my.php?image=p11dive1ta.gif)

Jetbuff
02-19-2006, 01:46 PM
Originally posted by NonWonderDog:
That's a bit odd looking. Air density increases at low altitude, of course, and drag force is proportional to air density -- but that doesn't entirely explain why it slows the way it does.
Precisely. It makes no sense.

I'm not sure what it would look like, but I'd guess that the graph would make MUCH more sense converted to TAS.
Like I said, I need a handy Excel equation to do the conversion reliably. The one I used is:

TAS = IAS*((Alt*3.2808399)/50000+1)

I have no means of verifying why this equation is what it is even since apparently many values were already substituted.

I have no idea why the prop would ever give negative thrust, though. That makes no sense at all.
I should have been more precise. Once the prop can no longer 'keep up' with the inflow, it is simply contributing to drag, no?

Jetbuff
02-19-2006, 01:56 PM
Originally posted by Ugly_Kid:
I don't understand why one even asks this question? Is there any particular hint it was modeled as a constant?
It is a hypothesis (a wild one I'll admit) that may help explain why dive acceleration is more heavily weighted in favour of the higher thrust-to-weight aircraft. In addition, it may contribute to some wonky slow speed behaviour - fudges in drag coefficients altering Vbg and Vbc or aircraft hanging on their props at ultra low speeds. We observe a discrepancy and propose some possible explanations. Then we try to disprove our hypothesis. At least it's better than the 'weight is not modelled' argument which has been disproved.

Jet has a constant thrust - so 262 should have a constant thrust for example.
I thought jets also have an efficiency range, but being better at speeds?

Prop aircraft has a thrust which decreases with speed - this regardless of prop efficiency. Prop efficiency comes on the top of this.
I didn't know that. Could you elaborate?

Now since the game models two parameters climb rate and top speed somewhat correctly I don't see how the thrust could be constant? If it was and one would reach correct top speed the climb would be too slow.
Well "somewhat" is a very subjective term. I also find it interesting that the discrepancy seems larger with some higher thrust aircraft. I often refer to the latter as clown-wagons because of not only their optimistic rates of climb, but the speeds at which these rates are achieved. If my hypothesis is correct, this would fit within it: bigger fudges to account for constant thrust from bigger engines. If it isn't, (and I'm totally open to that possibility) then we need to come up with another hypothesis.

Ugly_Kid
02-19-2006, 03:23 PM
Originally posted by Jetbuff:

I didn't know that. Could you elaborate?



Thrust force for a prop plane is given by formula:
T=nyy*P/V

where:
nyy is prop efficiency (usually around 0.85-0.9 on a good prop and is pretty constant over a large speed range with a CSP)

P is power

V is the airspeed

So you see there is linear decrease with speed for a CSP already from this relation. Changes in efficiency and drop in efficiency at high speed where prop tips may go supersonic causes a further decrease.

Note the thrust for jets is given directly as thrust force and there is not the previous relation with airspeed.

Jetbuff
02-19-2006, 04:51 PM
I see now. Thanks for the explanation. So given constant power, thrust is lower at low speeds AND low prop efficiency. Let me rephrase my supposition then: perhaps nyy is not being calculated but is assumed to be a constant, continuing to provide thrust at speeds where nyy would aproach 0. (based on prop advance ratio and blad tips going into compression)

WWMaxGunz
02-19-2006, 06:54 PM
Jetbuff, in that formula Thrust = prop_eff * power / speed, the lower speed is the higher
the thrust given power is held. At zero speed though, thrust becomes infinite/undefined.

You IAS/TAS formula above, the constants look very much like for alt in feet.

EDIT: you can try TAS(kph) = ((IAS/3.6)x(1+Alt/15000))x3.6

I kind of wish you had graphed speed vs time as speed as a function of time but dunno if
it'd make any difference that somehow the zero power plane started faster and got down
soon after the full power plane. It's a screwy mystery. It only takes a couple seconds
extra getting one headed down as opposed to the other to do that though and well within
human error I can tell ya from having processed a few dozen divespeed tracks, my own and
others sent in. The planes with the most torque and power start out 'kicking' as it were
by compare in missions and few players want to go through a long 'getting stable and even'
process for 4 to 12 planes over and over. I was trying to figure out how to set up an
even-matching starts algorithm when life intruded, it's not an easy thing and no matter
what the persons viewing the graph or data would have to understand more than 'that curve
is above that over one'. Things are usually that close.

How to make a big difference in dive one plane to the next: be the first rolled upside
down and using your lift to sling you downward. That force is for a short while capable
of overshadowing every other including gravity many times over. If you can pull up to
blackout or nearly, well apply that downward with gravity assisting. Even if the other
plane does manage such assist they will have travelled along across on the original
course 100's of meters before they could start which increases your seperation widely.
Tactics and execution can therefore outstrip any difference in raw dive speed differences.
How many airwar films have a squad rolling over and peeling off and down one member after
another to go into a fight? Well, in BoB the 109's did nose straight down since the Brit
planes engines would choke if they tried and they'd take many seconds to roll over. Again
tactics, once started diving and before high speed it's easy to roll unloaded.

Jetbuff
02-19-2006, 10:07 PM
Actually, the plane started nose down. I set it to 0 speed pointed at wp2 on the deck and almost directly below wp1.

NonWonderDog
02-19-2006, 11:33 PM
Originally posted by Jetbuff:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">I have no idea why the prop would ever give negative thrust, though. That makes no sense at all.
I should have been more precise. Once the prop can no longer 'keep up' with the inflow, it is simply contributing to drag, no? </div></BLOCKQUOTE>

I meant at the beginning. Adding power slowed the plane down at 40000-20000 feet, somehow.

It's propably something to do with poor altitude modelling in a plane never supposed to go that high.

The-Pizza-Man
02-20-2006, 01:07 AM
Originally posted by Ugly_Kid:
I don't understand why one even asks this question? Is there any particular hint it was modeled as a constant?

Jet has a constant thrust - so 262 should have a constant thrust for example.

Prop aircraft has a thrust which decreases with speed - this regardless of prop efficiency. Prop efficiency comes on the top of this.

Now these two factors are one reason why early jet pilots regarded acceleration of 262 as lousy. In a prop plane you have very high thrust in low speed so it accelerates very fast initially - towards the top speed the acceleration then decreases rapidly. Jet is slow to accelerate but the acceleration drops less rapidly towards the top speed...

Jets don't have a constant thrust. They work on exactly the same principle as a propeller. They loose thrust at high speed but they remain effective longer than propellers because the exit velocity of the exhaust is much higher than the the velocity of the air being pushed through a propeller.

Viper2005_
02-20-2006, 06:21 AM
That isn't strictly true. The jet has a diffuser in front of its intake, which means that a substantial proportion of the velocity of the intake air into a pressure rise.

This increases the thermal efficiency of the engine.

The pressure rise is converted into an increased exhaust gas velocity. This means that the turbomachinery doesn't know that it's flying, and prevents it from losing efficiency in the same way that an airscrew does.

In addition, because of the compression in the intake, the faster you go, the greater the mass flow the engine handles, and so the more fuel you can burn, which compensates for the reduction in thrust per unit mass flow as speed increases.

Jets tend to lose thrust slightly up to about 300 mph and then gain thrust from there on until they hit one limit or another (often intake efficiency or nozzle efficiency, duct pressure or one of various temperature limits).

Jetbuff
02-20-2006, 12:05 PM
Originally posted by NonWonderDog:
I meant at the beginning. Adding power slowed the plane down at 40000-20000 feet, somehow.

It's propably something to do with poor altitude modelling in a plane never supposed to go that high.
In that case I agree, it makes no sense. I'll apply Neal's TAS conversion and repost with TAS vs alt and TAS vs time as soon as I get the chance. First though, I'm thinking of redoing the tracks and averaging them out instead of choosing the 'best' run to reduce random error.

Jetbuff
02-20-2006, 12:11 PM
Originally posted by WWMaxGunz:
Jetbuff, in that formula Thrust = prop_eff * power / speed, the lower speed is the higher
the thrust given power is held. At zero speed though, thrust becomes infinite/undefined.
But prop efficiency also approaches zero at zero speed right? i.e. we end up with:

Thrust = 0 * infinity

But anything time 0 = 0, and anything times infinity = infinity. Egads! My brain is about to explode!

You IAS/TAS formula above, the constants look very much like for alt in feet.
It certainly looks like it (50000 feet I guess?) but the spreadsheet it was in used meters. Just goes to show never to trust anything off the internet without a litmus test.

EDIT: you can try TAS(kph) = ((IAS/3.6)x(1+Alt/15000))x3.6
Thanks, I'll give that a try.

I kind of wish you had graphed speed vs time as speed as a function of time but dunno if
it'd make any difference that somehow the zero power plane started faster and got down
soon after the full power plane.
Not sure what type of graph you are requesting here...?

Holtzauge
02-20-2006, 01:13 PM
Originally posted by The-Pizza-Man:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Ugly_Kid:
I don't understand why one even asks this question? Is there any particular hint it was modeled as a constant?

Jet has a constant thrust - so 262 should have a constant thrust for example.

Prop aircraft has a thrust which decreases with speed - this regardless of prop efficiency. Prop efficiency comes on the top of this.

Now these two factors are one reason why early jet pilots regarded acceleration of 262 as lousy. In a prop plane you have very high thrust in low speed so it accelerates very fast initially - towards the top speed the acceleration then decreases rapidly. Jet is slow to accelerate but the acceleration drops less rapidly towards the top speed...

Jets don't have a constant thrust. They work on exactly the same principle as a propeller. They loose thrust at high speed but they remain effective longer than propellers because the exit velocity of the exhaust is much higher than the the velocity of the air being pushed through a propeller. </div></BLOCKQUOTE>

True to a point. However, the thrust of a jet engine is dependant on the bypass ratio. More modern jet engines have a higher bypass ratio since this is good for fuel economy. However a straight jet with no bypass (like 50, 60's type jet's in Mirage III etc) can acually have increasing thrust with speed due to the ram.

Holtzauge
02-20-2006, 01:22 PM
About the 0/0 division issue.

I do belive that the charts showing efficiency on the y-axis and advance ratio on the x-axis are not meant to be used to determine static thrust but should be valid for climb and top speed though.

Sometimes, these graphs are complemented with a graphs showing the prop factor Ct on the y-axis instead for different prop blade angles. These may be used to determine static thrust (zero advance ratio).

Ct=Thrust/(airdensity x proprevs**2 x propdia**4)

Jetbuff
02-20-2006, 01:43 PM
I have seen those graphs Holtzauge, and they definitely seem more amenable to calculating thrust at zero speed.


While we're at it though, I've reread the entire thread and will try to summarize the highlights:
Thread summary to date:
<UL TYPE=SQUARE><LI> Thrust is probably not constant versus speed or rpms in-game, as it should be.
<LI> We do not know however if prop efficiency is factored into this variation. i.e., if thrust is not constant, does it conform to the equation:
Thrust = Prop_eff * Power / Velocity
or do we have something like this:
Thrust = Some_constant * Power / Velocity?
<LI> That P11 dive test sure is wonky! http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif[/list]

WWMaxGunz
02-20-2006, 01:50 PM
Originally posted by NonWonderDog:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Jetbuff:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">I have no idea why the prop would ever give negative thrust, though. That makes no sense at all.
I should have been more precise. Once the prop can no longer 'keep up' with the inflow, it is simply contributing to drag, no? </div></BLOCKQUOTE>

I meant at the beginning. Adding power slowed the plane down at 40000-20000 feet, somehow.

It's propably something to do with poor altitude modelling in a plane never supposed to go that high. </div></BLOCKQUOTE>

I have not seen the track but if the engine off run has a stopped prop that stays stopped
over most of the drop in question then it's going to have a much lower drag than the power
on prop. Once it starts turning though, it's going to have higher drag from more force
required to turn the thing. This is a possibility that occurs to me.

Jetbuff
02-20-2006, 01:52 PM
I'm afraid the prop was windmilling Neal iirc. Even if it wasn't, since I started at 0 airspeed, the 110% run should accelerate faster at first.

PS, your TAS equation is outputing some rather absurd values. Are you sure there's no typo?

PPS, how about using the d.alt/time to get TAS?

WWMaxGunz
02-20-2006, 02:09 PM
Originally posted by Jetbuff:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by WWMaxGunz:
Jetbuff, in that formula Thrust = prop_eff * power / speed, the lower speed is the higher
the thrust given power is held. At zero speed though, thrust becomes infinite/undefined.
But prop efficiency also approaches zero at zero speed right? i.e. we end up with:

Thrust = 0 * infinity

But anything time 0 = 0, and anything times infinity = infinity. Egads! My brain is about to explode! </div></BLOCKQUOTE>

I'm not sure why prop efficiency becomes zero at speed zero but it tells me the thrust
formula is either limited in domain or more complicated than it's spelled above which
certainly makes prop_eff look like a constant.

And from courses in calculus and even pre-calc I can tell ya that there's different
infinities and there's actual usable mathematics that resolve whole ranges of using them
to come out with finite values. I've forgotten the rules and tools but it is done.
It's kinda related to how we can take an infinite number of infinitely thin slices of
area under a curve and come up with the real value of the finite area under the curve.
Some of the first apps we learned were area of a circle or part of one and volume of
a sphere which for me was a triumph as for years I had to rely on a formula without
knowing why it was what it was.


<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">You IAS/TAS formula above, the constants look very much like for alt in feet.
It certainly looks like it (50000 feet I guess?) but the spreadsheet it was in used meters. Just goes to show never to trust anything off the internet without a litmus test.

EDIT: you can try TAS(kph) = ((IAS/3.6)x(1+Alt/15000))x3.6
Thanks, I'll give that a try.

I kind of wish you had graphed speed vs time as speed as a function of time but dunno if
it'd make any difference that somehow the zero power plane started faster and got down
soon after the full power plane.
Not sure what type of graph you are requesting here...? </div></BLOCKQUOTE>

I'd like to see time as "the X axis" (aka "the domain") and speed as a function of time
on "the Y axis" (aka "the range") rather than the other way around. It's a science thing.

Jetbuff
02-20-2006, 02:32 PM
Quick question, the Power in this equation: Thrust = prop_efff * P/V, what does it denote specifically? Engine power? If so, is it supposed to be constant for a given altitude, throttle and rpm?

WWMaxGunz
02-20-2006, 03:05 PM
Originally posted by Jetbuff:
I'm afraid the prop was windmilling Neal iirc. Even if it wasn't, since I started at 0 airspeed, the 110% run should accelerate faster at first.

PS, your TAS equation is outputing some rather absurd values. Are you sure there's no typo?

PPS, how about using the d.alt/time to get TAS?

Since you started at speed zero with engine off, wasn't the prop stopped then?
Oh H, it's a sim so we can't know from graphics and we can't know what is modelled how.
We can only guess.

I ran that equation (picked up from SimHQ as good but not perfect) on a spreadsheet
and came up with... well I won't type the whole thing and I can't paste in this window...

Speed in kph, alt in meters:
IAS 100, alts
--- 0000, 1000, 2000, 3000, 4000, 5000, 6000, 7000, 8000, 9000, 10000
TAS _100, _106, _113, _120, _126, _133, _140, _146, _153, _160, __166
with all TAS fractions rounded off, down, never up even the .67's.
IAS 300
TAS _300, _320, _340, _360, _380, _400, _420, _440, _460, _480, __500

I used the underlines to hopefully keep the TAS numbers lined up under the alts.

A quick look comparing my table to airwarfare.com->essential files->precision IAS/TAS
for bombers, at 300kph IAS and 'my' table is very slightly high, like 1 kph up to 4000m
= very close.
Hey, just grab that file and unzip. Okay there'll be some work linking to it.
The file is .xls, excell, but works fine on Solaris OpenOffice.
Isn't that Butch's site?

WWMaxGunz
02-20-2006, 03:12 PM
Originally posted by Jetbuff:
Quick question, the Power in this equation: Thrust = prop_efff * P/V, what does it denote specifically? Engine power? If so, is it supposed to be constant for a given altitude, throttle and rpm?

Pretty sure it's engine torque which is I think the most common measure of engine power.
Fun part would be getting all the units right. Most of what I get from it is that Thrust
decreases directly as Velocity increases. As for the rest, ask Holtzgage cause he is
beyond me here by far. I just diddle with the terms and numbers sometimes. http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif

The-Pizza-Man
02-20-2006, 04:33 PM
Originally posted by Viper2005_:
That isn't strictly true. The jet has a diffuser in front of its intake, which means that a substantial proportion of the velocity of the intake air into a pressure rise.

This increases the thermal efficiency of the engine.

The pressure rise is converted into an increased exhaust gas velocity. This means that the turbomachinery doesn't know that it's flying, and prevents it from losing efficiency in the same way that an airscrew does.

In addition, because of the compression in the intake, the faster you go, the greater the mass flow the engine handles, and so the more fuel you can burn, which compensates for the reduction in thrust per unit mass flow as speed increases.

Jets tend to lose thrust slightly up to about 300 mph and then gain thrust from there on until they hit one limit or another (often intake efficiency or nozzle efficiency, duct pressure or one of various temperature limits).

I didn't know that the exit velocity was able to increase that way, thanks.

Viper2005_
02-20-2006, 05:20 PM
In all fairness the answer I posted is somewhat glib (detail generally isn't appreciated on the 'net, and my free time for posting is limited!), but you will generally find that the conclusions hold true, at least in my limited experience. BTW, note that SFC is far from constant with Mach number, despite the protestations of the basic textbooks...

For a serious answer, I suggest

http://www.amazon.co.uk/exec/obidos/ASIN/0262111624/qid.../202-0644286-9453445 (http://www.amazon.co.uk/exec/obidos/ASIN/0262111624/qid=1140481065/sr=8-1/ref=sr_8_xs_ap_i1_xgl/202-0644286-9453445)
or
http://product.ebay.com/Aircraft-Engines-and-Gas-Turbin...csZ1388QQsoprZ119739 (http://product.ebay.com/Aircraft-Engines-and-Gas-Turbines_ISBN_0262111624_W0QQfromZR31QQfvcsZ1388QQ soprZ119739)

I have the 1977 edition; things have moved on since then, but the thermodynamics are still the same (I hope!).

NonWonderDog
02-20-2006, 08:04 PM
Originally posted by Jetbuff:
Quick question, the Power in this equation: Thrust = prop_efff * P/V, what does it denote specifically? Engine power? If so, is it supposed to be constant for a given altitude, throttle and rpm?
P = Engine power in W or ft^2*lb/s^2 (Or something like that)
V = Velocity in m/s or ft/s
Efficiency is dimensionless. Power is not constant, but rather varies with altitude with a formula I've forgotten at the moment. It can be almost constant in supercharged planes up to critical alt, though, depending on supercharger setup.

This breaks down at low speeds, as you noticed. At very low speed the rudimentary approximation is one of constant thrust. There are much more detailed ways of analyzing things, but I'm not very familiar with them.

The-Pizza-Man
02-20-2006, 08:39 PM
Originally posted by Viper2005_:
In all fairness the answer I posted is somewhat glib (detail generally isn't appreciated on the 'net, and my free time for posting is limited!), but you will generally find that the conclusions hold true, at least in my limited experience. BTW, note that SFC is far from constant with Mach number, despite the protestations of the basic textbooks...

For a serious answer, I suggest

http://www.amazon.co.uk/exec/obidos/ASIN/0262111624/qid.../202-0644286-9453445 (http://www.amazon.co.uk/exec/obidos/ASIN/0262111624/qid=1140481065/sr=8-1/ref=sr_8_xs_ap_i1_xgl/202-0644286-9453445)
or
http://product.ebay.com/Aircraft-Engines-and-Gas-Turbin...csZ1388QQsoprZ119739 (http://product.ebay.com/Aircraft-Engines-and-Gas-Turbines_ISBN_0262111624_W0QQfromZR31QQfvcsZ1388QQ soprZ119739)

I have the 1977 edition; things have moved on since then, but the thermodynamics are still the same (I hope!).

I think I'll probably learn about the intracacies of jet engines at some stage in the next 2 years, I'm studying for my BE(double major in Space and Mechanical engineering). So I'll wait until then to learn the finer details, don't want to spoil all the suprises http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif. The pressure rise didn't occur to me at all, but now that I've read your post it makes a lot of sense.

WWMaxGunz
02-21-2006, 02:34 AM
Don't they teach jet fighter pilots to dive before climbing if speed is not optimal?
Seem to remember an article on that at SimHQ years ago.

Jetbuff
02-21-2006, 03:51 AM
OK, since we're not getting an answer it seems, I started out with the P11 again. I keep coming back to this plane because its fixed angle propellor means that the effect of its prop efficiency on thrust will be much more evident if there is one. Here's what I'm planning to do and I'm hoping people will correct any mistaken assumptions or bad math if they see them.

Step 1: Determine empty mass of the aircraft.
Dive plane down to ~1000m +/-50m, level out as best as possible and track deceleration around a suitably high speed where induced drag is negligible. (~350kph which is higher than top level speed of the P11 seems appropriate but 400kph is doable too) Not sure if that's strictly necessary, but it seemed like a good idea at the time. http://forums.ubi.com/images/smilies/16x16_smiley-wink.gif

During the run, note rate of deceleration at closed throttle from UDPGraph (find two adjacent values spanning the selected reference point and do a dv/dt) and then plug into the following formula:

Drag = m * a

Naturally, this will need two different runs, with different fuel loads and I can solve the equations simultaneously to determine aircraft mass.

BTW, rough first draft came out surprisingly low, ~350kg below Empty P11 weight! (784kg vs actual weight of 1147kg) Of course it could just be unit issues or the shoddy job I did on the runs themselves. I was in a hurry to get some real values to work with, just to get a feel for it.

Step 2: Find Drag values at two reference points
With mass in hand, another high speed deceleration is performed with engine once again throttled down. Drag values are extrapolated from the m*a equation for two different speeds where we might expect that there would be a significant difference between prop efficiency values. (say 250kph and 500kph, the former being below top level speed is definite to have a non-zero prop-efficiency value and the latter may have one approaching zero)

Step 3: Solve for Thrust at the reference points
Aircraft is now decelerated again through the two drag reference points but this time throttle is full open during the deceleration. (or indeed acceleration at the lower reference point) The difference in deceleration is the result of thrust.

Thrust - Drag = m * a

And, since we have mass and drag for the reference points, we can find the value of thrust at both points.

Step 4: See if Thrust conforms to the prop_efficiency equation
Fortunately, not more tracks, just a little fuzzy math:

If Thrust = np * P/V
but we are unsure of either np or P we solve the equation for the two velocities to find the product of np*P. Now P should be constant at constant altitude. (hence my earlier question)

i.e. we are assuming the equation above should resolve to:

Thrust = Some_Constant/V if prop efficiency is not being accounted for. i.e. if Thrust1*V1 = Thrust2*V2, prop efficiency is being ignored.

Unit questions:
Speed: m/s. Not sure though whether using TAS for determining acceleration is the right thing, should I be using IAS instead? I'm pretty sure the former is the right way but I've been wrong before. I'm using MaxGunz' conversion algorithm which works quite well at approx. 1000m based on both in-game dial values and estimated values. (have yet to verify it against the IAS/TAS charts though)

Time: seconds. I'll probably use smaller increments (500ms or less in the data recording phase to get narrower speed ranges for more accurate point deceleration values.

Altitude: meters of course. The key point here though is what should be my allowable error level? During even the rush jobs I could get +/-5m during the deceleration but it's bloody hard to get it any lower than that and I really don't want to cosine resolve acceleration due to gravity to avoid complicating things and magnifying any errors.

Weight: Kilograms. That's how Devicelink's fuel is reported - and thank God for that, less worrying about the extra error of using a guestimate for fuel density.

Forces: Newtons I guess since we're using Kg and meters.

Energy: Joules.



Well that's it I guess. So, is it a load of bull? Oh and my metric predisposition doesn't count, I'm Canadian! http://forums.ubi.com/images/smilies/16x16_smiley-tongue.gif

Viper2005_
02-21-2006, 04:38 AM
The easiest way to work out the mass is to carry out constant IAS glide tests with different fuel loads.

i) pick a glide angle.

ii) fly at that angle and measure your IAS.

iii) fly at that angle again with a different fuel load.

This works because if you're gliding at the same angle then you must be attaining the same L/D which means (provided that you're not on the wrong side of the drag curve) that you must have the same Cl and Cd.

Therefore your IAS will vary as the square root of the mass ratio.

You know what the maximum fuel load of your aircraft is, and as avgas has a specific gravity of around 0.7 you can make a pretty accurate guess as to the mass change caused by the fuel.

Because the test is steady state you should be able to get pretty decent accuracy...

Jetbuff
02-21-2006, 04:50 AM
That's what I started out trying to do initially, (even recorded a couple of tracks) but I got lost somewhere along the line. A formula or two might help here as my head is swimming with sigmas, deltas and mu's at the moment! http://forums.ubi.com/images/smilies/16x16_smiley-wink.gif

Basically, once I find the angle (not easy btw from DeviceLink) and the IAS, how do I use them to estimate mass?

I started messing around with rate of PE loss as a measure of drag but wasn't 100% sure of the following two points:

(a) is per second energy loss really equal to drag force?

(b) wouldn't drag be slightly offset from the direction of travel?

PS: Fortunately, fuel weight is readily available in kg from devicelink. No guesswork required like figuring out what size tank you have, rate of consumption or verifying whether it's C3 or B2 or what-have-you. (different densities)

The-Pizza-Man
02-21-2006, 05:27 AM
Originally posted by WWMaxGunz:
Don't they teach jet fighter pilots to dive before climbing if speed is not optimal?
Seem to remember an article on that at SimHQ years ago.

I think how that works is that you unload to zero g so the wing doesn't have to produce any lift, so 0 degree AoA so no induced drag. I think they do it to get past the transonic drag hump. I might have read the same article, is it the one how he talks about applying the technique, which he learnt in modern combat sims, to WW2 ones?

Holtzauge
02-21-2006, 10:05 AM
Originally posted by WWMaxGunz:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Jetbuff:
Quick question, the Power in this equation: Thrust = prop_efff * P/V, what does it denote specifically? Engine power? If so, is it supposed to be constant for a given altitude, throttle and rpm?

Pretty sure it's engine torque which is I think the most common measure of engine power.
Fun part would be getting all the units right. Most of what I get from it is that Thrust
decreases directly as Velocity increases. As for the rest, ask Holtzgage cause he is
beyond me here by far. I just diddle with the terms and numbers sometimes. http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif </div></BLOCKQUOTE>

Being brought up with te metric system I have only mastered this way of doing my math. I must say I'm surprised anyone can do the math with BTU's, slugs/ft**3 and whatnot.... http://forums.ubi.com/images/smilies/53.gif

Seriously, I would do the calculation like this:

P = horsepower * 736 = Watts
Thurst = Newtons
Speed = meters/sec

Example: Assume P=1800 hp, assume prop efficieny 0.8, assume climb at 280 km/h =280/3.6=78 m/s

Thrust=(1800*736*0.8)/78=13588 Newtons or 1305 Kp or 3051 lb

So i reckon within reasonable deviations from this speed, the thrust should increase/decrease with the inverse of the TAS speed.

WWMaxGunz
02-21-2006, 11:11 AM
Originally posted by The-Pizza-Man:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by WWMaxGunz:
Don't they teach jet fighter pilots to dive before climbing if speed is not optimal?
Seem to remember an article on that at SimHQ years ago.

I think how that works is that you unload to zero g so the wing doesn't have to produce any lift, so 0 degree AoA so no induced drag. I think they do it to get past the transonic drag hump. I might have read the same article, is it the one how he talks about applying the technique, which he learnt in modern combat sims, to WW2 ones? </div></BLOCKQUOTE>

Yeah he did by using data from the WW2 sim mentioned there. In reality the data needed
either does not exist or does for maybe a couple of planes. It's very interesting but
really applies to jet power curves. WWSensei who used to fly F-16's straightened me out
on that one when I asked if it would be a good edge to know for the squad.

BM357_Sniper
02-21-2006, 11:18 AM
Originally posted by Jetbuff:
But doesn't prop efficiency does correlate with thrust?

Prop efficiency = Ct/Cp*V (or something like that)

i.e. prop efficiency determines the rate of conversion of available power to usable thrust. At low efficiency, much of the blades' work (aka engine power) is wasted, no? I only read the first few posts and saw that it got way off subject. So I'll just ask here. Did the question get answered? Can we use published manifold and rpm settings for cruise and climb? I'm assuming no since the only thing I've noticed changing prop pitch will do in game is help cool the engine a bit faster. Too bad really that its that way.

WWMaxGunz
02-21-2006, 11:26 AM
Originally posted by Holtzauge:
Being brought up with te metric system I have only mastered this way of doing my math. I must say I'm surprised anyone can do the math with BTU's, slugs/ft**3 and whatnot.... http://forums.ubi.com/images/smilies/53.gif

Seriously, I would do the calculation like this:

P = horsepower * 736 = Watts
Thurst = Newtons
Speed = meters/sec

Example: Assume P=1800 hp, assume prop efficieny 0.8, assume climb at 280 km/h =280/3.6=78 m/s

Thrust=(1800*736*0.8)/78=13588 Newtons or 1305 Kp or 3051 lb

So i reckon within reasonable deviations from this speed, the thrust should increase/decrease with the inverse of the TAS speed.

First I need to apologize for repeatedly spelling your forum name wrong.

Next, I was schooled in both systems at once. I paid attention to the units first and
then the conversions or just the data second and things generally worked themselves out
since I usually got good grades on those subjects.

Hey, I just look at divide by V and know that increase in V is decrease in result.
I haven't done the details of things like that since before 1990 to tell the truth.
I even have to stop and look up or ask before thinking about force vs work any more.
What the hey, I'm still alive, won't be forever and nearly wasn't a few years ago so
I'm not doing that badly. Some day my contribution will be helping the grass to grow.

Viper2005_
02-21-2006, 01:25 PM
Jetbuff, it all falls out of basic trig.

If you imagine level flight we can define lift a force which points up and equals weight, and drag as a force which points backwards and equals thrust.

Now take the thrust away.

To keep things simple let's start by pretending that the atmosphere is isotropic and that the ground is an infinite distance away, so that we can just experiment with a steady state for as long as it takes...

To maintain constant speed we have to re-align the drag force until it is cancelled out.

By convention, lift and drag are always measured relative to a fixed datum on the aircraft (generally you'd use the mean aerodynamic chord).

This means that variations in alpha are taken into account by changing the lift and drag numbers rather than by re-aligning the direction of the lift and drag vectors.

In other words, lift is defined as being the component of the aerodynamic force acting at 90º to the mean chord, and drag is the component of force acting along the mean chord in opposition to the direction of motion. This keeps things nice and simple.

Now, let's go back to our brand new glider.

In a steady state glide, we're holding a constant glide angle, and so because our speed in the direction of the aircraft's motion (forward and slightly down) isn't changing, it follows that the speed of the aircraft along the horizontal axis isn't changing.

We can then do some trig:

Lift*sin(theta) = Drag*cos(theta)

Where theta is the angle between the aircraft's flight path and the horizontal.

In addition we can of course say that:

Lift*cos(theta) + Drag*sin(theta) = Weight

I think that the first equation is simpler to work with. It reduces to...

(L/D)*(sin(theta)/cos(theta) = 1

ie:

L/D = 1/tan(theta)

Thus if you hold constant glide angle you're holding constant L/D.

If we constrain ourselves to flying only on the right side of the power curve (ie faster than L/D max) then we know that L/D vs alpha is a 1:1 function and so we can say that at any given glide angle there is only one possible Cl and Cd.

From the lift equation we have it that

L = Q*Cl*S

and

D = Q*Cd*S

So we can say that the increased force required to maintain 1 g flight when fuel is added must come from a Q increase. Since Q varies as v^2 we can expect v to vary as the square root of the mass ratio.

Now it's time to remove those starting assumptions.

The wonderful thing about airspeed indicators is that they don't measure v at all; they measure Q. Thus if we hold constant IAS we're holding constant Q. Therefore we don't care that the atmosphere isn't isotropic because it gets cancelled out by our test setup, provided that we quietly forget about Mach number effects (which are pretty insignificant in this context) and Reynolds number effects (which are also pretty insignificant in this context).

As for the ground, you can just hit refly http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif

Holtzauge
02-21-2006, 01:25 PM
Originally posted by WWMaxGunz:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Holtzauge:
Being brought up with te metric system I have only mastered this way of doing my math. I must say I'm surprised anyone can do the math with BTU's, slugs/ft**3 and whatnot.... http://forums.ubi.com/images/smilies/53.gif

Seriously, I would do the calculation like this:

P = horsepower * 736 = Watts
Thurst = Newtons
Speed = meters/sec

Example: Assume P=1800 hp, assume prop efficieny 0.8, assume climb at 280 km/h =280/3.6=78 m/s

Thrust=(1800*736*0.8)/78=13588 Newtons or 1305 Kp or 3051 lb

So i reckon within reasonable deviations from this speed, the thrust should increase/decrease with the inverse of the TAS speed.

First I need to apologize for repeatedly spelling your forum name wrong.

Next, I was schooled in both systems at once. I paid attention to the units first and
then the conversions or just the data second and things generally worked themselves out
since I usually got good grades on those subjects.

Hey, I just look at divide by V and know that increase in V is decrease in result.
I haven't done the details of things like that since before 1990 to tell the truth.
I even have to stop and look up or ask before thinking about force vs work any more.
What the hey, I'm still alive, won't be forever and nearly wasn't a few years ago so
I'm not doing that badly. Some day my contribution will be helping the grass to grow. </div></BLOCKQUOTE>

WWMaxGunz: No offense meant. I only try to lighten things up a bit sometimes. Mayby I have a lousy sense of humor...

About spelling the my name: Don't apologize, I'm the one that got it wrong in the first place: should be Holzauge I believe. However, not being bred on knackwurst as a kid I got it wrong myself and now I have to live with it.

BTW: Arn't the Brits and Yanks going metric inch by inch nowdays?

Viper2005_
02-21-2006, 01:35 PM
I'm a Brit. I use SI for most things, which is great; but sometimes it's convenient to use other systems, especially if you want to units of force and mass which are identical in magnitude in our terrestrial context, or if you're working with old stuff (like WWII fighters) in which case it's easier to just work in horsepower etc.

Jetbuff
02-21-2006, 02:28 PM
Originally posted by Viper2005_:
As for the ground, you can just hit refly http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif
And what about an exploding head? Any reflies for that one? http://forums.ubi.com/images/smilies/16x16_smiley-sad.gif

I'm probably beyond stupid here, but I still don't see how we could determine mass from those equations. The test may be far easier, but the math is so much more complicated.

OK, I dive twice at two different weights, m1 = empty-weight+50% fuel and m2 = empty-weight+100% fuel. I note the IAS, descent rates and and angle of glide in both tests giving v1, v2, d1, d2, t1, t2, theta1 and theta2. Now what?

Viper2005_
02-21-2006, 02:32 PM
No, you use the same angle of glide in each test. That's the whole point.

Then M1/M2 = (V1/V2)^0.5

M = Mass
V = IAS

Of course this assumes that g is constant (really I should talk in terms of weight instead of mass), but the error is likely to be smaller than the experimental error anyway even if Oleg has bothered to vary g with altitude.

NonWonderDog
02-21-2006, 05:12 PM
We Yanks are taught metric and imperial systems side by side, but the US industry uses the imperial system almost exclusively. Only scientific institutes use SI with any regularity; most engineering work is still done with imperial units. http://forums.ubi.com/images/smilies/53.gif

Jetbuff
02-21-2006, 09:23 PM
M1/M2 = (V1/V2)^0.5
M1 = m + 23.5 (10% fuel)
M2 = m + 235 (full fuel load)
V1 = 104m/s
V2 = 108m/s

m+23.5/m+235 = (104/108)^0.5
m+23.5 = 0.9813*(m + 235)
m+23.5 = 0.9813*m + 230.6
m = 207.1/0.0187
m = 11074kg? A bit much, no? Actually in the order of 10x too much! http://forums.ubi.com/groupee_common/emoticons/icon_eek.gif

Something's off I'm afraid.

Viper2005_
02-22-2006, 12:49 AM
I think I made a mistake;

The formula should be

M1/M2 = (V1/V2)^2

I just mixed up Q and V. In effect the ASI reads in Q^0.5 because it's calibrated in units of velocity. As we're interested in Q we need to square the airspeed ratio.

Plugging your figures in gives

M = 2674.123

Assuming that the fuel mass was measured in kg that gives a pretty reasonable mass for a WWII fighter (which one did you use?).

For those still working in old money, that would be 5895 lb.

Jetbuff
02-22-2006, 12:55 AM
I used the P11 and that's still more than twice its empty weight of 1147kg. Unless the guns and ammo weigh another tonne! http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif

I must have done something really wrong in the test runs is all I can think of at the moment. I'll repeat it when I get the chance and see if I end up with any better results. Perhaps try another aircraft too?

WWMaxGunz
02-22-2006, 04:05 AM
Originally posted by NonWonderDog:
We Yanks are taught metric and imperial systems side by side, but the US industry uses the imperial system almost exclusively. Only scientific institutes use SI with any regularity; most engineering work is still done with imperial units. http://forums.ubi.com/images/smilies/53.gif

Really I had worked to both systems, one or the other or in two places we spec'd in both
(I was a draftsman at one and an Engineering Aide that became Production Engineer at the other
but never for real bucks) between 1974 and 1981.

And then the country went Voodoo and Astrology by Royal Decree in the 80's. Science became
politicised, education went down the toilet and neither has recovered.

Extreme Sarcasm ON: What Global Warming? Imagination! A commie plot to destroy the US!
Those tree hugging liberals don't know anything right! So what the sea is a little warmer.
So what the ice caps are shrinking? Change means opportunities. Burn more coal!
ESM OFF:
Sorry folks but I learned science before it got twisted. Between the oceans and the icecaps
the signs are as clear as a busted bank account that mankind is in for some really bad times.

Viper2005_
02-22-2006, 05:00 AM
I suggest diving in wonderwoman view and using the artificial horizon provided.

Hold a 10º dive angle for 1 km of altitude and read off the IAS. You did use IAS right? Otherwise we'd need to correct for air density variations, and that would be nasty...

This has the advantage that you can trim it out.

I don't have device link;

i) where can I get it

ii) can you post the mass of the fuel carried by a selection of the aircraft in the planeset?

It seems to me that the P-51, Fw-190A9, Spitfire IX (+25) and P-47 would make for a good starting point.

It could well be that my sums are out, but I'm certain that the basic idea is correct...

Viper2005_
02-22-2006, 05:40 AM
I just carried out a very quick & dirty test using the P11c.

I started at 3 km, killed the engine at once and feathered the prop(!).

I turned so that I was pointing out to sea.

I flew on the crimea map and held a 10º dive angle in wonderwoman view.

I then let the thing glide down until impact.

Just before impact I paused the game and read off the airspeed from wonderwoman view.

I got about 320 km/h with 100% fuel, and about 300 km/h with 25% fuel.

Using your numbers, 100% fuel is 235 kg.

Therefore 25% fuel is 235/4.

I plugged the answers into the formula and got a mass of 1220 kg (rounded to the nearest kg).

That seems pretty reasonable to me.

I will carry out a more accurate test presently...

Jetbuff
02-22-2006, 02:06 PM
How did you figure out the angle? I used FMB at 3000m and a target 9000m ahead to aim for.

Also, I did not kill the engine and was not aware the prop could be feathered on this plane. (quite surprising actually)

Finally, note that your IAS dial goes in increments of 10kph. You need to use Devicelink for an accurate readout of your IAS.

Viper2005_
02-22-2006, 05:40 PM
The artificial horizon in wonder woman view has attitude marks on it; I just used them. The angle of attack is the same in each run and therefore cancels out; that's the beauty of this mathematical approach - you don't really need to know very much about what the aeroplane is doing at all!

Quite why the P11c has a feathering prop I don't know, but it does, and it's useful in this context because it allows you to just treat it as a glider...

As for IAS, wonderwoman view again has the answer. It's got a digital TAS gauge. Now it just so happens that this method doesn't care what units of speed you use, and therefore by extension it doesn't care if you use TAS provided that you take your TAS measurements at the same altitude on each run.

I took my measurements in the last second or so before impact.

If you want serious accuracy you could record a track and play it back in slow motion, allowing you to take data at an exact altitude (or indeed several if you felt that way inclined) to the nearest km/h TAS...

As such the only thing you really need device link for is to tell you the mass of a full fuel load.

Jetbuff
02-22-2006, 07:16 PM
The artificial horizon in wonder woman view has attitude marks on it; I just used them.
Serves me right for not using WW view more often. I never noticed that dial before. http://forums.ubi.com/images/smilies/16x16_smiley-happy.gif

Ok, the numbers I got for 100% fuel: 87m/s, for 0% fuel: 80m/s. I used devicelink because unlike the digital display it is more precise. (1kph as opposed to 10kph increments)

Plugging in the numbers:
M1 = m (0% fuel)
M2 = m + 235 (full fuel load)
V1 = 80m/s
V2 = 87m/s

m/m+235 = (80/87)²
m = 0.84555423437706434139252212973973*(m + 235)
m = 0.84555423437706434139252212973973*m + 198.70524507861012022724270048884
m = 198.70524507861012022724270048884/0.15444576562293565860747787026027
m = 1286.5697177074422583404619332761kg

MUCH better! (guns and ammo ~= 140kg + empty P11 weight of 1147kg)

Now, are there any other idiocies I might have committed in my subsequent steps? i.e. is using the difference in acceleration (or decceleration) at full power vs. engine off a valid means of calculating thrust? And once calculated does my final test make sense?

PS: I just realized that I didn't strictly need to know the value of m for the latter steps. I could just figure out if (np * P)/m was a constant versus a*V at two different velocities where we would otherwise expect np to vary. Still it's helpful to know that we can now test quite reliably for the mass of any aircraft in the sim.

DaimonSyrius
02-22-2006, 07:19 PM
Originally posted by Viper2005_:
I don't have device link;

i) where can I get it


Hi Viper,

http://avcsite.ru/il2dl/graph_en.htm

UDPgraph is the best option for producing graphs (and logging data) IMO.

All you need is to add to your config.ini:

[DeviceLink]
port=21100
IPS=127.0.0.1
host=127.0.0.1

That's for running UDPgraph on the same PC as IL2. If you can run it on a second PC over a LAN, just set the IP's accordingly (IPS: UDPgraph client; host: IL2). The web site provides basic set-up info.

Cheers,
S.

Jetbuff
02-22-2006, 07:22 PM
Oh, btw, while the option to feather is welcome, (albeit surprising) would it not upset things if I estimated engine-off decceleration with the prop feathered before comparing it to the power-on decceleration? I know it doesn't matter for the mass estimation so long as the prop is feathered/un-feathered in both runs.

i.e., wouldn't a feathered prop mean that the power-off test would have less drag acting on the plane versus the power-on test?

Jetbuff
02-22-2006, 07:35 PM
Sorry, yet another question. I was really glad that I seemed to have nailed the correct glide angles in the two runs. But for the sake of argument, let's say one was to dispute that the angle was exactly similar for both runs. To minimize the chances of that, let me pass the following two precautions by you and let me know if they make sense:

1. We only use a suitably large dive angle so that a small deviation, if it exists, is less significant. e.g. a 1? deviation from 10? glide angle is a lot more serious than 1? deviation at 30?. Of course, we can't use too large an angle or it becomes difficult to reach a static glide state in a suitable amount of time. Does that pass the reason test?

2. We could calculate TAS*dt to map the flightpath distance versus dalt/dt. This gives us the adjacent/hypotenuse (cosine) relationship for the glide angle. In fact, I think it may be possible to configure UDPGraph to display ˜ right then and there for more accuracy than the artificial horizon provides.

DaimonSyrius
02-22-2006, 07:52 PM
Originally posted by Jetbuff:
In fact, I think it may be possible to configure UDPGraph to display ˜ right then and there for more accuracy than the artificial horizon provides.
Hey Jetbuff http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

Yes it's possible, you can use all kinds of trigonometric functions, and even macros for more complex math/trig functions can be defined.

About the sense your first point makes... hmm I'm not sure if deviation/angle is really important when it comes to using the trigonometric functions of said angles... but then, don't mind me, my trigonometry is both dusty and rusty right off my head http://forums.ubi.com/groupee_common/emoticons/icon_smile.gif

Cheers,
S.

WWMaxGunz
02-22-2006, 08:36 PM
Originally posted by Viper2005_:
The artificial horizon in wonder woman view has attitude marks on it; I just used them. The angle of attack is the same in each run and therefore cancels out; that's the beauty of this mathematical approach - you don't really need to know very much about what the aeroplane is doing at all!

Quite why the P11c has a feathering prop I don't know, but it does, and it's useful in this context because it allows you to just treat it as a glider...

As for IAS, wonderwoman view again has the answer. It's got a digital TAS gauge. Now it just so happens that this method doesn't care what units of speed you use, and therefore by extension it doesn't care if you use TAS provided that you take your TAS measurements at the same altitude on each run.

I took my measurements in the last second or so before impact.

If you want serious accuracy you could record a track and play it back in slow motion, allowing you to take data at an exact altitude (or indeed several if you felt that way inclined) to the nearest km/h TAS...

As such the only thing you really need device link for is to tell you the mass of a full fuel load.

How steeply do you have to glide for pitch to be the same as path or the difference
between the two not matter for the speeds involved? Serious unloading?

EDIT: Oh hey Viper, would you like a copy of UDPSpeed? You can run it in playback and
it'll spit out a custom data log. You make successive runs and each one adds to the log.

Jetbuff
02-22-2006, 08:45 PM
Here's what I used:

asin(altimeter.dt/(@TAS*$DT))

Where @TAS is tas2(ispeed,altimeter)

I was quite surprised to see the degrees seem to range from 0 (level) to -0.90. (straight down) Interesting choice of units. (deg/100?)


As for angle variation, we need there to be zero (or as close to zero) difference between the full fuel run and no fuel run. Otherwise the glide slope is different and everything goes out the window. That's why I want to measure the glide angle more accurately.

Viper2005_
02-23-2006, 08:17 AM
i) the TAS gauge in ww view has a digital output which reads to an accuracy of 1 km/h - that's less than 0.3 m/s.

ii) the alpha will always cancel in these tests; that's the whole point*. The only caveat on that is that you need to give the aeroplane time to sort itself out in the glide at constant IAS which can take quite a while. Give it a couple of km and it'll be fine.

*I could probably thrash out a proof if you really wanted me to, but it should follow naturally from the fact that if we constrain ourselves to flying only on the right side of the power curve (ie faster than L/D max) then we know that L/D vs alpha is a 1:1 function and so we can say that at any given glide angle there is only one possible Cl and Cd.

If the aeroplane is flying too slowly, provided that you're on the correct side of the power curve it'll have less total drag despite the fact that it's having to pull more Cl to maintain 1 g flight. It will therefore accelerate until it reaches equilibrium when the L/D corresponds with the commanded glide angle.

As such you just need to put the gunsight on a fixed marker.

The "Heath Robinson" approach to doing this is to replicate German dive-bombing technology; just get an OHP transparency, attach it to your monitor and draw a line somewhere above the gunsight (if you want to be "scientific" about it, you could carry out a test glide and mark the line on the basis of that test). Then just put the horizon on your guideline for all of your tests and you're sorted.

Once you get good at it, you should be able to have the aeroplane trimmed out by the time you get down to 1.5 km or so. Then you can start recording and it should be a simple matter to get at the glide angle using trig by just using the map and altimeter.

All of these glide tests should really be carried out with a stopped prop to keep things simple.

You can stop the prop in most of the fighters in game using manual prop pitch; this can have a substantial impact upon glide performance.

Once the mass has been determined, the next stage is to glide at a variety of indicated airspeeds and measure sink rate in order to determine the power required to sustain flight.

Then the thrust horsepower may be calculated from climb tests at a variety of speeds.

This allows a complete picture of aircraft performance to be built up.

However it's very time consuming. Given that changes will doubtless be made in 4.04 it seems to me that we should use the time remaining to us in this patch to get the method nailed.

I would eventually hope to be able to publish a complete flight testing procedure to enable all of the aeroplanes in the game to be tested against real life data by anybody who feels that way inclined; I personally don't have the time to test fly more than a couple myself.

WWMaxGunz
02-23-2006, 10:03 AM
Can you find your unknowns by gliding different paths but always the same sink rate?
VSI is more knowable data in the sim and directly viewable and loggable via devicelink.

Jetbuff
02-23-2006, 03:20 PM
Actually Neal, you can determine glide angle quite easily like I said earlier. A simple function can make it visible via UDPGraph's display.

Viper, if you have not used UDPGraph you're missing out. With it, accuracy is assured. Not only can it pull values like IAS, altitude, g, etc. from the game, it can also pre-process them versus delta time, trig functions, whatever you need. Via this method there is no guessing as the UDPGraph bar with degrees, acceleration or any other calculated information is displaced right there while you are flying in real time!

Here's the readme:

Russian web page http://avcsite.ru/il2dl/graph_ru.htm
English web page http://avcsite.ru/il2dl/graph_en.htm

You can contact me:

e-mail: atchilikine@hotmail.com
ICQ: 107784586

Please feel free to send me any suggestions.

Regards,
Andrey Chilikin aka AVC aka GR_Patrick


UdpGraph 2.0.15 14-June-2005
* New delta parameters added: .ds delta per second, .dm delta per millisecond
----------------------------------------------------------------------------------
UdpGraph 2.0.14 13-June-2005
* Fixed request string for multiple eng[x].x parameters
* When adding new graph "visible" is set by default
----------------------------------------------------------------------------------
UdpGraph 2.0.13 13-June-2005
* Added checking for valid file name characters for log file. For example
for Bf109E/4.
* Fixed processing of Exit command using context menu;
--------------------------------------------------------------------------------
UdpGraph 2.0.12 29-Oct-2004
* UdpGraph now can distinguish game pause and start of new track
* $TODMSEC removed because there is no way to calculate it properly
--------------------------------------------------------------------------------
UdpGraph 2.0.11 26-Oct-2004
* Three new functions added to calculate a/c energy:
Ep(altitude) as altitude*9.81
Ek(tas) as (tas^2)/2
Esum(tas,altitude) as a summ of Ep and Ek
* Fixed double bracket for functions calls like (1+sind(90))
* Subversion info added on dialog header and main dialog page
* CUdpLog::ClearLog() function added to remove old log file at startup;
* Fixed calculation of initial Y window position.
* Fixed saving of nameless items to an ini file
* Fixed extention for multiple files - it was "cvs" not "csv"
* Read-only flag removed from log files
--------------------------------------------------------------------------------
21 Oct 2004
UdpGraph version 2.0 beta 10
What is new:

* cpu() function checks time interval internally and updates value only
every one second
* Window Settings dialog box fixed. Now it saves font name properly.
--------------------------------------------------------------------------------
20 Oct 2004
UdpGraph version 2.0 beta 9
What is new:

* Global variables automatically created for all graphs. To access
global value of graph use @ prefix. For example check UdpGraph.ini
file.

* Log file settings dialog box updated. New format (milliseconds) and
example of time stamp added.

* Global variable $TODMSEC added - it stores "Time of Day" in milliseconds.
Because IL-2 provides time on second basis $DT used to adjust TOD to
millisecond resolution

* Global variable $TRMSEC added - it stores number of milliseconds since
connection start.

* UdpGraph creates error log file using configuration file name and
replacing .ini with .log. Only parsing errors are logged.

* side(hypot,side) changed: returns 0.0 if hypot < 0.0 or side > hypot
else returns sqrt(hypot^2-side^2)

* All 'engine[x].xxx' DeviceLink keys renamed to 'eng[x].xxx'

* sind() and cosd() functions added to calculate sin and cos using degrees.

* cpu() function updated. Now it can be called with process name to get
CPU usage for a process. I.e. cpu() returns total CPU usage and
cpu(UdpGraph) returns CPU usage for UdpGraph. Note: it is recomended to
use cpu() function with polling intervals not less then 1 second, otherwise
it can return very inaccurate values.


16 Oct 2004
UdpGraph version 2.0 beta 8

Main features of version 2:

* UdpGraph configuration can be done using provided GUI.
To open dialog box click right button and select Settings.
Manual editing of UdpGraph.ini still possible but undesirable because file format might be changed in future versions.

When you start UdpGraph 2.0 first time it automatically converts UdpGraph.ini file into a new format.

* Context menu for list of graphs - you can add, delete change position of graphs at run time

* More settings for each graph - visible, logged, precision,

* List of functions supported by UdpGraph:
Trigonometric functions:
sin() cos() tan()
asin() acos() atan()

Hyperbolic functions
sinh() cosh() tanh()

Exponential and logarithmic functions
log() log10() exp()

Power functions
pow() sqrt()

Absolute value
fabs()

Right triangle functions
hypot(sideA,sideB) - calculates the hypotenuse
side(hypot,side) - calculates a side using hypotenuse and second side.
if hypot < 0.0 function returns 0.0
if hypot > 0.0 and side > hypot function returns hypot
if hypot > 0.0 and side < hypot function returns sqrt(hypot^2-side^2)

Special functions
cpu() - takes no parameters, returns current total CPU load
tas2(IAS, ALT)
IAS - indicated speed (km/h), ALT - altitude (m)
Function returns TAS(m/sec) = IAS/3.6*(1 + ALT/15000)

tas(IAS, ALT, pressure, temperature)
IAS - indicated speed (km/h), ALT -altitude (m),
air pressure at see level (mm of mercury), air temperature (C)
Function returns TAS (m/sec)

* Set of constants added to be used for calculations:
Definitions of useful mathematical constants
$E = 2.71828182845904523536 # e
$LOG2E = 1.44269504088896340736 # log2(e)
$LOG10E = 0.434294481903251827651 # log10(e)
$LN2 = 0.693147180559945309417 # ln(2)
$LN10 = 2.30258509299404568402 # ln(10)
$PI = 3.14159265358979323846 # pi
$PI_2 = 1.57079632679489661923 # pi/2
$PI_4 = 0.785398163397448309616 # pi/4
$PI_180 = 0.017453292519943295769 # pi/180
$1_PI = 0.318309886183790671538 # 1/pi
$2_PI = 0.636619772367581343076 # 2/pi
$2_SQRTPI = 1.12837916709551257390 # 2/sqrt(pi)
$SQRT2 = 1.41421356237309504880 # sqrt(2)
$SQRT1_2 = 0.707106781186547524401 # 1/sqrt(2)

Special global variables:

$DT - time elapsed between polls, seconds
$TOD - "time of day" from IL-2
$TODSEC - "time of day" in milliseconds
$CLOCK - "time of day" in HH.MMSS format
$TRSEC - time of track in seconds
$TRCLOCK - time of track in HH.MMSS format

New constants can be added manually to UdpGraph ini file into [globals] section.

* Global variables can be used across all functions. To declare global variable start its name with $
For example following set of functions will calculate offset of a/c from point of origin:

$tas=tas(ispeed,altimeter,745,25)
$vect=$tas*$DT
$vect=side($vect,altimeter.dt)
$dx=$dx+$vect*sin(azimuth*$PI_180)
$dy=$dy+$vect*cos(azimuth*$PI_180)

Note: if you declare global variable in [globals] section it will not be set to 0.0 between tracks. Global variables declared in function body set to 0.0 on end of track.

* dlkeys.ini file contains list of DeviceLink key that can be used with UdpGraph. If you like you can change names of any key or add new keys.

* To get delta of any DeviceLink key or global variable add .dt to the name of key or variable: altimeter.dt or $tas.dt

* Note: because of way how timers are used poll intervals very imprecise and in current version are multiple of ~109 milliseconds, i.e. if you select 500 msec poll interval in reality it might oscillate between 547 and 563 msec. There is a special global variable $DT which stores how much time (in seconds) elapsed between function calls.
It would be awesome if we could produce an easily followed structure for producing the necessary tracks and running them through UDPGraph to derive EM diagrams. So definitely, keep the ideas coming.

Viper2005_
02-23-2006, 03:54 PM
Ok I'll get UDP graph sorted out post haste (after I've taken 4.04 for a quick spin)...