PDA

View Full Version : Engine power output and adjusting it.



anthony31
June 4th, 2010, 15:48
I am starting on the FDE for a new aircraft and I have entered all the aircraft specs into the aircraft.cfg.

The engine is a Rotax which produces 100hp at 5800 rpm (i have the specs sheets for the engine output vs rpm) with a fixed prop.

The problem is that when I enter the "piston_engine" data the best I can get is 75hp at 6000rpm. I have tried adjusting every parameter in the piston_engine data including power_scalar but the only one that seems to change this fundamental fact of 75hp at 6000rpm is engine displacement. Everything else, even the power_scalar, only seems to allow the engine to rev higher and produce more power that way rather than produce more power at 6000rpm.

Messing about (using aam) with tables 508 and 509 in the air file also seem to adjust only how high (or low) the engine will rev and no matter what I adjust I always end up with 75hp at 6000rpm.

The only thing that seems to work is adjusting table 512, prop power required coefficient. By adjusting the curve upwards I am able to produce more power.

The question is, am I going about this the right way? Is using table 512 to get the required power output from the engine the right thing to do?

Thanks in advance for any help.

Brett_Henderson
June 4th, 2010, 16:38
Try using the standard air-file (un-modifed)... and use realistic numbers in the cfg file.. and THEN allow for the gear-reduction (I believe it's in the Propeller paragraph) combined with adjusting the fixed-pitch.

fliger747
June 5th, 2010, 12:59
Compression ratio will also adjust the delivered HP. I suggest looking at the engine parameters with AFSD which will show both the output and friction HP values. It may be that you can reduce the friction values either in the cfg or air files and change the max delivered output.

Worth a try: T

Shane Olguin
June 5th, 2010, 15:20
You can get more power by reducing the friction at the rated RPM. Go to table 509. The last value on the table will be your max RPM. X is the RPM and Y is the amount of friction applied. Gradually reduce the fricton value until you get the sort of horsepower you want at your rated RPM.

Lemme know if that helps.

Brett_Henderson
June 5th, 2010, 18:21
You really don't want to change much in the air file. If you start "cheating" the physics, ala artificially low friction.. you might as well just crank up the power scalars (prop and/or engine) in the cfg file.

The delivered power in the MSFS algorithm (assuming realistic parameters), is a function of manifold pressure and PROP rpm. It gets tricky with these Rotax type engines, because they have gear reduction between the engine and the prop. A Rotax red-lines at well over 5000rpm, where a Lycoming/Continental redlines at ~2800rpm.

A normal GA prop is in the 6' diameter range, so you can't have the prop going much over 2800rpm, else the blade-tips go supersonic, and the prop becaomes very inneficient. ... hence the gear reduction for engines that run well over 2800rpm.

So what you have to do, is plug the gear-reduction into the cfg file, and carefully experiment with blade pitch.

Also.. most of the data you'd change in the air file, is over-ridden by the cfg file, anyway.

fliger747
June 5th, 2010, 23:55
There ae friction/RPM eficency and torque curves in the .air file (501-508-509) which can be tweaked to adjust the power deliverd vrs RPM or MP to match known values. However you might also consider the mechanical efficency and friction scalars in the .cfg file:

max_rpm_mechanical_efficiency_scalar = 1
idle_rpm_mechanical_efficiency_scalar = 1
max_rpm_friction_scalar = 1
idle_rpm_friction_scalar = 1

Hard to guess at the exact issues, but looking at the friction and output values for the engine, as well as the delivered thrust and prop efficency values using a utility such as AFSD. With a fixed prop the prop tables may have a significant effect on the loading of the engine.

Cheers: T

Ivan
June 7th, 2010, 12:53
I believe Shane Olguin has a pretty good method with editing record 509. Besides the friction table, there is also an efficiency table (508) that can be adjusted. I believe the reduction gear ratio is specified in record 510 and also in record 505.

I don't know if the Rotax is supercharged, but you can adjust how quickly the power falls off at altitude by balancing the values in 508 and 509. There are many pairs of values that produce the same power at sea level and up to the critical altitude, but these pairs don't produce the same power past the critical altitude on a supercharged engine.

BTW, Record 512 needs to be reasonable for your engine / propeller combination. Sparks can easily explain it better than I can, but I see this table as describing the relationship of how the engine power / reduction matches the propeller that is installed. It doesn't really describe either the engine or the propeller but is more a description of the match between the two.

Hope that helps.
- Ivan.

fliger747
June 8th, 2010, 00:27
One thing to note in the editing of the engine RPM Friction/efficency tables is the lower end values will effect the startup torque required. To high a value and you might have to tweak the starter torque up, too low a value the prop may continue to spin on shutdown.

Also the Rotax, being a normaly aspitated engine with a fixed pitch prop... will have it's critical altitude at sea level. The prop tables (which are not as obvious) will have a big effect on the RPM's developed at various throttle settings and on the thrust delivered.

Have fun.... T

Ivan
June 9th, 2010, 03:32
Hi Brett_Henderson,
Do you happen to know the actual formula used by MSFS to calculate engine power? It would save me a lot of experimentation if I could use a formula to plot the power curves. I would imagine it has more to do with the Engine RPM rather than the Propeller RPM.

Hi Fliger747,
There are enough values in the tables (thought barely so) to alter the high rpm effects without really affecting the low RPM values. I plot the values on a spreadsheet when I rework these tables to make sure the curves look reasonable. The funny thing is that often if I try to match the output at lower RPM and manifold pressures as from a Specific Engine Flight Chart, often at full manifold pressure, the HP will peak at a lower RPM than maximum. I have run into this situation a bunch of times: most recently with the B-25C and A6M2 I am working on.

- Ivan.

Brett_Henderson
June 9th, 2010, 05:36
Hi Brett_Henderson,
Do you happen to know the actual formula used by MSFS to calculate engine power? It would save me a lot of experimentation if I could use a formula to plot the power curves. I would imagine it has more to do with the Engine RPM rather than the Propeller RPM.




No, alas I do not.. but it has to be some combo of MP and RPM..

On that note.. engine RPM and prop RPM are the same thing. The prop and engine are rigidly, mechanically linked.. even if by gear reduction like a Rotax or a big radial. As engine RPM changes, prop RPM changes by the same percentage.. always.

Like I've tried to point out.. if you're going to allow yourself to manipulate things like frictions in the air file, you might as well just use the power and thrust scalars.

I'd work with the gear-reduction, prop-MOI, and blade pitch in the cfg, as a first step.. and THEN "slightly" tweak the air-file. Starting in the air file is like pretending you have magical motor oil that can let you defy friction..

Ivan
June 9th, 2010, 09:17
Hi Brett_Henderson,

.... or big inlines.....

I play mostly with CFS1, so there isn't the CFG Parameters to manipulate. Messing with the AIR file is the ONLY way with the earlier sims. There are a lot more factors than just Manifold Pressure and RPM such as Compression, Displacement, Efficiency (508), Friction Loss (509), Supercharger values, and probably a bunch more that I either don't know about or can't recall at the moment.

I chose to distinguish between Engine RPM and Propeller RPM even though they are related because there are lots of more or less identical engines that differ in reduction gear ratios and propeller sizes. When you plug the RPM intio your equation, it will be the engine RPM and not the propeller RPM.

If you are starting with a given engine, you should use the REAL reduction gear and propeller MOI and blade pitch rather than tweaking those values to get a particular effect. I believe that most the factors you mentioned can limit RPM and thus horsepower generated at a particular speed, but don't really affect HP otherwise. I don't really see how prop MOI actually affects HP at all though it would affect how fast the prop spins up.

As I see it, the efficiency and friction tables are the correct means of adjustment because they are a summary of the cumulative effects of engine features which are not otherwise described. Perhaps your engine has cams, headers or a tuned runner intake that produces a torque peak at a particular RPM. I don't see how you can duplicate that effect without tweaking tables 508 and 509.

- Ivan.

fliger747
June 10th, 2010, 03:03
In many real aircraft engines max HP may not be developed at max RPM. There were many difference between the R2800 B and C engines. One significant one was that they were able to increase the max RPM from 2700 to 2800. At max RPM's the amount of HP lost through slinging oil about the crankcase was something on the order of 400 HP. by carefully redesigning the oil scavenging system they were able to reduce much of this loss and operate effeciently at both higher RPM and also deliver more shaft HP. I am currently working on the R2800 22W and 34W engines for the various SOH Tigercats.

It would appear that FS derives HP by taking the displacement, Compression ratio and friction values vrs RPM. For instance delivered HP can be adjusted by varying the compression ratio. Usually starting with the book figures for any given engine will give up pretty close values. I don't know anything at all about the ROTAX engines. Gas powered Aircraft engines tend to have fairly low compression ratios vrs automotive applications. Part of the reason for this has to do with the desireability of relaiability at fairly high power values, maintained for quite some time. Auto engines are by comparison usually run at low % of max power for most of their duty cycle.

Generally not too much varation from standard book engine values are required to at least touch the performance curves at several points. That you are dealing with a fixed prop makes things more complex rather than simpler. The prop loadings continuously and variably effect engine loading and RPM, dependng on altitude, temperature, airspeed and throttle setting. Way to many things changing each other!!

Good Luck: T

Brett_Henderson
June 10th, 2010, 04:53
Way to many things changing each other!!


That's the problem we deal with.. Not that that itself is a bad thing.. we'd want all the variables to be in play, and interact with each other.. but the MSFS algorithms aren't that accurate to begin with.

When you start adjusting things like internal engine friction, and compression ratios; you're messing with the foundation of an already suspect building. I long ago gave up on thinking that careful research and accurate numbers will result in accurate results... and so did MSFS.. that's why there's an entire paragraph in the cfg file named [Flight_Tuning].

I did not know that CFS has no cfg file.. so you're forced to tweak within the air-file.. and I'll defer to those with CFS tweaking experience. But you're still dealing within the limitations of the MSFS "engine" in general... So you have to decide what your goal is...

Do you want the engine itself (ala testing with a guage to read % of power), to have a realistic power curve ? Or do you want the model in flight to have realistic performance ? In MSFS (as opposed to CFS), the further away you get from a 180hp, light single; the more impossible those goals become... and the more cheating is required. I'd imagine it's similar in CFS.. and with no cfg file to play with.. you really do end up opening cans of worms while chasing your own tail.

What I've found in MSFS, is to set the internal stuff at real numbers.. and then work from the outside in. When you work from the inside out, it's like trying to change the appearance of a house via the foundation.

If there are adjustments (cheats) for just power and thrust.. work with those before trying to get the power accurate from inside the engine... and then in an attempt to get actual in-flight performance accurate, work with parasitic/induced drag.. play with gear reduction ratios and prop pitches, rather than frictions and compressions.. then, eventually you'll have to custom calibrate guages and gauge bitmaps for in-cockpit realism.

No matter how you cut it.. you're gonna cheat somewhere.. I just think it's easier working from the outside, in.

Milton Shupe
June 10th, 2010, 15:16
Just another thought here ...

Can you develop the air file in FS9 or 8 using Jerry's Airwrench (free version for single engine).

Then use the output air file in CFS, or the tables that are compatible.

fliger747
June 10th, 2010, 18:13
Indeed you might give Airwrench a try as the free default version is for fixed pitch props.

Perhaps not so useful an approach for the project under consideration... but the approach I use is to get the HP and Thrust right and then adjust the drag to match the known performance points (as best as possible). For turbocharged engines this works fairly well, but supercharged engines with an accessory stage (perhaps with high and low speeds) MS cannot make the zig zags without some sophisticated gauge work.

Airwrench will change EVERYTHING in your files, If this isn't totally satisfactory you can take the adjusted pieces (ie. engine etc) and transfer them into you main air/cfg files.

Good luck! T

Ivan
June 11th, 2010, 13:34
Hi Fliger747,
Do you have actual data for HP versus RPM for the R-2800 engines? I am curious as to where the peak HP is reached if it isn't at max RPM. I know this happens a lot with Automobile engines. It is also a good exercise to see if I can edit friction and efficiency to reproduce a given graph. At this point most of the data I have is from SEFCs from various manuals.

Hi Brett_Henderson,
I am not really sure what you mean about the MSFS algorithms being inaccurate. Other than being unable to represent multi-speed superchargers and their shift points, I don't see a big problem. I figure that as long as the power output at a given manifold pressure, RPM and altitude can be approximated, there really isn't an issue.

The oil slinging issue that Fliger747 described is basically an increase in friction at a certain RPM. If we reproduce that by increasing the friction in our flightsim "engine", how is that cheating? On the next model engine, if we installed a windage tray and reduced this power loss, what is wrong with adjusting the friction table to take this into account?

In real engines, there often is a major difference in efficiency at various speeds. With a cam with lots of overlap, you may have very poor efficiency and torque at low speeds and pretty good power and torque at high speeds. You can even exceed a 100% volumetric efficiency in certain RPM ranges if the harmonics in the intake and exhaust are just right. What is wrong with simulating these features with the efficiency table?

- Ivan.

Brett_Henderson
June 11th, 2010, 13:51
If the algorithms were accurate, we wouldn't be having ths discussion :wiggle:

Think about it .. The designer provides displacement, number of cylinders, and compression.. and THEN is asked to supply a HP figure .. :confused:

Bore and stroke (biggest factors for a power curve) aren't even used... :isadizzy:

Like the flight model that won't stall accurately.. the engine model is very crude.. To think that all these tables have predictable, realistic effect when you start tinkering with them, is naive. The crude engine model works best if left alone, and the tinkering is best done "outside" of it..

If you really want to play with torque and power curves,,, you're best off designing gauges that read environment variables, and send performance paramters back.

fliger747
June 11th, 2010, 17:31
Ivan:

Of course the various R2800 models, especially the B and C models have a lot of difference in HP vrs RPM/MP etc. Each instalation varied in intake system and super/turbocharging. I am on the road (VHHH) at the moment and have no references available. I did see some engine data on the net recently for the F4U installation. I know we also had this info available for the A26 project. For the 22 and 34W engines used in the F7F I hope to have the info available soon, presuming the ordered pilots handbook arrives.

Bore/stroke are included to some extent in the calculations as displacement and compreesion ratios are considered. Most classic air cooled piston aircraft engines have similar bore/stroke ratios, close enough to allow a basic calculation that can be tweaked.

A certain degree of precision is possible, but only a certain degree. It is possible that the variation in performace between in service aircraft may be as great as our ability to approach the "book figures" America's Hundred Thousand is an interesting work in that where available both Factory and service test data are compared. The difference can be significant!

Cheers: T

Brett_Henderson
June 11th, 2010, 19:50
Bore/stroke are included to some extent in the calculations as displacement and compreesion ratios are considered. Most classic air cooled piston aircraft engines have similar bore/stroke ratios, close enough to allow a basic calculation that can be tweaked.


Agreed.. However.. displacement is the geometric volume of an imaginary cylinder; defined by bore/stroke. And compression-ratio is the ratio between the volumes "above" the piston; at the top and bottom of the stroke. It's determined by the piston size/shape... ie. you can raise compression-ratio by swapping in "high compression pistons", without changing displacement.

If we assume that the MSFS "engine" uses a generic bore/stroke so that displacement can define the bore/stroke, then we're admitting THE most important set of variable for a specific power-curve, aren't variables at all.

If you try to shape a curve "internally", you're assuming that the MSFS "engine" actually (and acccurately) takes those type of variables into consideration.. It doesn't. Sure, things will change as you play with that stuff, but not realistically. The model isn't that sophisticated.

If your end goal is realistic, in-flight perfromance.. you're better off leaving the guts alone, and working with prop-pitch, gear-reduction, and the brute force, power/thrust scalars. All you'll really "see", or even be able to document; are airspeeds per power-settings. You can sit down and make a chart, documenting airspeed by power-setting, but a rate of change in airpseed over a range of power settings (the curve) would be next to impossible to document.. It's great mental gymnastics to attempt, but not really achievable. It's all kinda like groping in the dark. I've found that you do most of it form "outside" the engine, you aren't chasing as many tails.

Now again.. it might be different for CFS, but I got a feeling they're the same "engines".. and these are just my experiences.. not meant to be argumentative.. I'll be looking forward to the results :salute:

anthony31
June 12th, 2010, 03:09
Thank you gentlemen for your input on this subject. You have certainly given me some good places to start looking.

I think I need to do some more tinkering to find out what does what especially in the .cfg file.

I have filled out the .cfg file with all the data for the engine and prop (compression ratio, capacity, gear reduction, prop size (this can have a big effect on the engine!) ) etc etc.

The Rotax is a 82.6 cu in 4 cylinder / 4 stroke engine producing 100hp at 5800 rpm and about 90hp at 5200 rpm (cruise speed), compression ratio 10.5 and an inbuilt gear reduction of 2.43.

As it isn't turbocharged am I correct in assuming that the manifold pressure values are not used?

stansdds
June 12th, 2010, 03:28
If it has neither a turbocharger nor a supercharger, then the manifold pressure would actually be a vacuum, just like most gasoline powered passenger car engines. Turbochargers and superchargers both force air into the manifold, changing manifold pressure condition from a vacuum (negative pressure) to a positive pressure.

Brett_Henderson
June 12th, 2010, 04:45
If it has neither a turbocharger nor a supercharger, then the manifold pressure would actually be a vacuum, just like most gasoline powered passenger car engines. Turbochargers and superchargers both force air into the manifold, changing manifold pressure condition from a vacuum (negative pressure) to a positive pressure.

It's never really a vacuum (zero pressure), but it's always less than atmospheric (sans turbo/super charging), except when the engine isn't running. If your MP pressure gauge were more accurate, you could use it to set your altimeter, before starting the engine (adjusting for field elevation)... because it's really just a barometer for the inside of the intake manifold :jump:

Brett_Henderson
June 12th, 2010, 04:54
Side note on the Rotax.. My very first freeware model was a Europa-XS (FS2002). I gave up on trying to make it work.. and just put a 110 HP Lycoming in there.. And oddly enough, some years later, our club bought a Liberty-XL2 (essentially an factory built Europa), with an IO-240 in it.


Edit: so.. you could just us a standard engine (down-size and de-rate the engine out of the default C172), and then customize the tachometer to display the Rotax-like RPMs..

Ivan
June 12th, 2010, 14:53
I will try to build the Rotax engine as an exercise, but I need a couple more details:
What is the cruise and maximum manifold pressures?
What is the prop diameter and reduction gear ratio?
What is the pitch of the prop and is there a measured static thrust at a given power setting?
Is there a torque or HP to RPM graph that I can try to match?

- Ivan.

Brett_Henderson
June 12th, 2010, 15:10
I will try to build the Rotax engine as an exercise, but I need a couple more details:
What is the cruise and maximum manifold pressures?
What is the prop diameter and reduction gear ratio?
What is the pitch of the prop and is there a measured static thrust at a given power setting?
Is there a torque or HP to RPM graph that I can try to match?

- Ivan.

Manifold pressure is weather and altitude dependent.. I.E.. if you're at sea-level, and the atmospheric pressure is 29.92, your theoretical MP limit is 29.92 (with a theoretically perfect induction system). As you climb the available MP decreases. As for what YOU set MP at.. it's a non-issue with a fixed-pitch prop. The throttle is your only means for adjusting RPM, so you never get the most out of the available, atmospheric pressure. With a constant-speed prop you'd have the throttle wide-open at cruise altitudes, and would pull the RPMs back with the prop-control. You lose about an inch per 1000 feet, so at 8,000, a wide-open throttle would only get you ~22 inches of MP.

Just use published numbers for the prop diameter (double-checking so that a prop-strike doesn't happen with hard braking).

The gear-reduction is published too, but trust me, you'll be tinkering with that.

Same for prop-pitch..

Trying to match a realistic power curve (via a chart) is admirable, but it'll be problematic when you try to get realistic performance out of an engine that lives up to that graph (if you could even document that it does)(see the testing post itn the other thrust thread)

Ivan
June 12th, 2010, 22:01
Hello Brett_Henderson,

I presume the "Other Thrust Thread" you are refering to is the one about Turbofans? I had not been paying any attention to that thread because I have no great interest in jets at the moment. There is plenty of things to read about old propeller driven planes for someone like me who is not an aeronautical engineer by trade. Jets may come later, but I believe they are just a temporary fad.

I am still messing around with aircraft from the WW2 era. I am a fan of big old piston engines. I have heard of the Rotax but know almost nothing about it. It's a bit too modern for me. I was hoping that someone here could provide some details because there is bound to be a difference betwen the version I find with a Google search and the one of interest here. Since this IS an exercise that isn't of great interest to me, I also don't want to spend a lot of time trying to research the topic. Thanks for the reminder about naturally aspirated engines. Last one I worked on for a plane was the rotary for a Fokker Eindecker that I will probably release eventually.

Although I agree with you that the engine modelling in MS FS or MS CFS is far from perfect (multispeed superchargers being the greatest example), I believe there is enough room to come fairly close to documented horsepower and thrust values for a given engine. My biggest difficulty is in finding the information on the original engine.

Here is an example of what can be accomplished with engine tuning for a Japanese A6M2 Zero:
(2550 RPM)
Altitude Power Speed Manifold Pressure
500 915 295 37.8
2500 931 304 37.8
5000 952 310 37.8
7500 973 317 37.8
10000 995 324 37.8
12500 1001 336 37.3
15000 906 330 33.7
17500 820 326 30.5
20000 735 321 27.5
22500 656 315 24.7
25000 583 310 22.2
27500 517 304 19.9
30000 452 295 17.7
32500 392 277 15.8
35000 339 247 14.1

Service Ceiling (100 fpm) is 36,100 feet.
Initial Climb Rate is around 3400 fpm
Best Climb Rate is 3580 fpm
Because it was flown with an autopilot during most of the testing, I expect that under manual control, its service ceiling would be a couple thousand feet lower.

The engine power is slightly low at 20,000 feet and presumably at higher altitudes but that was done as I described earlier to lower the service ceiling.

Performance and Engine power versus altitude are pretty much where I want them to be. They don't match any particular flight test report of an actual aircraft but aren't too far off any of the reports either. The engine power isn't too far off manufacturer's specs at lower altitudes but are noticeably off by 20,000 feet.

- Ivan.

Ivan
June 12th, 2010, 22:06
BTW, here is a screenshot of the Eindecker.

- Ivan.

Brett_Henderson
June 13th, 2010, 04:03
Neat stuff ..

The high altitude numbers do look a bit odd. The airpseeds look close, but not for the amount of MP in use.


Even with the handicap of no, multi-speed super-chargers.. you DO have some very handy (outside the engine) parameters to play with.. that could help "realize" that performance table. Namely; the critical-altitude, and max/min blade pitch... combined with, wing-efficiency/prop-thrust/induced-drag/parasitic-drag.

Like mentioned in the other thread.. first get it to where takeoff airspeed and lift are reached after using up a realistic length of runway (per the load).. Then work on cruise performance at altitude.. go back and forth until those extremes are close enough to satisfy you.. then work on climb-performance, obviously returning to the original tests as you progress to make sure they're still close.

A model might very well have a realistic engine performance curve.. but if it leaps off the runway too easily, or flies with jets at 35,000msl (or uses too much runway, and doesn't climb well), the in-game experience suffers.

Ivan
June 13th, 2010, 06:26
Hi Brett_Henderson,

Why do the high altitude numbers look odd? This is a supercharged 1700 cubic inch 14 cylinder engine.

In CFS1, there aren't the CFG parameters to play with. So far, from what I can tell in reading the Microsoft SDK just about all the variables we have been tweaking line up with the variables the SDK describes.

The variables you suggest tweaking are well documented for the real aircraft and should not be tweaked in my opinion. Why should we change the propeller pitch range for example when the actual numbers are documented in USN and TAIC reports? I bellieve we should stay as close to real numbers as possible.

As I see it, when you start changing other numbers to compensate for inaccurate engine output, other less than obvious performance parameters get screwed up. Engine output is relatively easy to adjust and doesn't depend all that much on other airframe variables, so I believe it should be the FIRST thing to be tuned. I don't believe your aversion to virtual engine tuning makes sense.

A FW 190D I worked on for a friend a couple years ago had engine output that was way too low. To compensate for this, the induced and parasitic drag were much too low. This plane would glide just about forever with the engine off which isn't very realistic.

Another series of aircraft I tested a few years back had engine power that was about 50% higher than it should have been. The aircraft flew well enough with power on, but were way too easy to land because throttling back the engine was about equivalent to deploying air brakes (which did not exist).

- Ivan.

Brett_Henderson
June 13th, 2010, 07:39
Why do the high altitude numbers look odd? This is a supercharged 1700 cubic inch 14 cylinder engine.


Odd in that you can get 247kias (over 400tas @ 35000msl) with only 14" of manifold pressure.


As I see it, when you start changing other numbers to compensate for inaccurate engine output, other less than obvious performance parameters get screwed up. Engine output is relatively easy to adjust and doesn't depend all that much on other airframe variables, so I believe it should be the FIRST thing to be tuned. I don't believe your aversion to virtual engine tuning makes sense.



I don't believe my method makes sense either.. lol But the simplicity of the MSFS model makes all of this an experiment, at best.

One way or another, if you're concerned about accurate in-flight , you're gonna have to start tweaking for it. Knowing that your engine itself is accurate, but the model won't perform realistically, defeats the whole purpose. Standing on a stack of data that proves the engine power curve is good means nothing if the airplane can't get airborne, AT the proper airspeed, using up the proper amount of runway.. or worse.. leaps off the ground well before it should at something other than V3.. or climbs ridiculously fast (or slow), or is capable of jet-like airspeed at cruise (or is too slow)and if your engine gauges aren't returning accurate data for what your asking the airplane to do, it gets even messier.

I've found, that if you don't play around "inside" of the engine, the other stuff is more achievable... especially gauge readings (ala MP).

Sure, you want to start with something close, "inside" the engine, but if you venture into air-file tables in search of an accurate engine, you're making the actual performance envelope harder to achieve.

I wish you didn't have to play with stuff like wing-effiency, prop-pitch, drags, etc.. when trying to mate takeoff performance to cruise performance (and everything in between), but you really have no choice. It would be wonderful if you could compartmentalize it all.. make a good engine.. make a geometrically/aerodynamically accurate air-frame and just "bolt" them together, and get predictable results.. but that aint gonna happen. I advise you to keep the engine "stock" air-file wise, until you get the performance right.. then (if you really wanna re-open the tweaking cycle), try to "realize" the engine in its own right. Just know that you'll have quite a task getting the takeoff/climb/cruise numbers back to where you want them, not to mention major skewing in gauge readings...

I'm just sharing hard-earned experrience, and truly do respect those willing to strive for realism.. ...and I'm realizing that these topics can get argumentative.. so I'll step aside unless asked for input.

fliger747
June 13th, 2010, 08:38
Some interesting things occur in altitude performance of supercharged engines. Sometimes the data does not lead one to the proper conclusions. The way FS is set up, the critical altitudes for ADI (Water Meth Systems) and Mil power are the same. In practical applications, the lower MP of the Mil power can occur at a higher altitude than the superchargers ability to provide the maximum manifold pressure acceptable for ADI use. The overall true airspeeds may not actually be that much different, but occur at different altitudes.

In FS maximum available MP and the power delivered will not change with altitude up to the critical altitude. In real engines, especially ones with several supercharger stages and gearing, the maximum power available at the high blower will be less than low blower due to the very high power consumption of the blower, which is subtracted from what the engine can deliver to the prop.

For high altitude operation of fast aircraft, mach drag must be taken into consideration. For WII aircraft this may show up as the plane being faster than book values at max speed at altitude.

Lots of stuff going on all at once!

Cheers: T

Ivan
June 13th, 2010, 17:49
Hi Brett_Henderson,

I guess the reason that the numbers don't look right is because what you are reading as Knots IAS is really MPH TAS.

Pity to sign off. This discussion was getting pretty interesting considering the radically different approaches we seem to have to basically the same problem.

- Ivan.

Brett_Henderson
June 13th, 2010, 18:12
That's kinda like asking for input.. :salute:

OK.. if it's 247 TRUE mph, then it's 215KTAS, wich means the indicated airspeed wouldn't be much more than 110 knots.. I'm no Zero expert, but that can't be right... though it probably is right for 14" of MP.

I wonder what Vs is at 35,000msl.. ? I'm thinking it should have long since stalled..

(I'll do some Zero research)..

Edit: That does indeed look pretty good, especially within the MSFS model. I'm not sure qbout the 14" of MP, but at this altitude extreme.. it's realistic enough for simming. Now, if your takeoff numbers are good, and the runway required for takeoff/landing is close... and it doesn't climb like a jet with after-burners (lol).. and your MP/RPM readings satisfy the discriminating sim-pilot..you've nailed the primary flight model. :applause:

fliger747
June 14th, 2010, 06:35
At their service ceiling most plane are pretty wimpy handlers and need a very light touch. About the only use for really high altitude is the high fast corridor which might keep the competition form ever quite getting up to you, as with a recon bird, or as a perch with which to dive on prey.

The Power off IAS speed at which the Zero would stall should remain about independent of altitude though the reduced thrust would also increase the power on stall speed. 110 knots (or maybe a bit slower) IAS might be somewhere around the min thrust speed for this plane, which is the speed you would creep toward at high altitude to get any extra thrust that might allow for some more climb.

The rate of climb figures are the hardest to come up with in your own test and probably the most unreliable real test data out there. If the thrust and drag (and weights) are correct the ROC should fall into line.

Cheers: T

Brett_Henderson
June 14th, 2010, 09:45
Yeah.. rate-of-climb numbers are tough to nail in the sim.

The next step for this flight model, is to see how it takes off. Weren't they often flown from carriers ? So.. load it up with fuel and payloads it would see in carrier service.. and see if it can get up and flying from a carrier-equivelant length of runway.. then check the glide-ratio AT published best-glide.. and then see if it can be controllably landed at a carrier approach speed/configuration.

Ivan
June 14th, 2010, 13:55
The A6M Zero is a very light aircraft (about 5500 pounds fully loaded) with a fairly large wing: 39 foot wingspan and 240 square feet of area. There is nothing particularly weird about the airfoil. It has a couple degrees of washout. Basically its reputation for agility is because of the very low wing loading.

The stall speed I am getting with this aircraft is 68 mph clean at an altitude of 500 feet which is in the ballpark. This is by reducing power and speed and maintaining altitude until the aircraft stalls. If I throttle back to idle, the aircraft slows down so fast that I can't always note the speed at stall.

Longitudinal stability is VERY poor near the service ceiling. On manual controls instead of autopilot, I am lucky to come within 2000 feet of what the autopilot can do. The test for service and absolute ceiling (about 800 feet higher) was done with about 75% fuel. With only 50% fuel, it will go about 1000 feet higher. Service ceiling should only be about 33,000 feet.

Documented climb rates are all over the place for this plane: anywhere from just under 3000 fpm to 4500 fpm depending on the source.

Takeoff performance and low speed acceleration were rather mediocre until I tuned the power curves to increase the HP below max rpm. It still peaks at 2550 rpm, but HP doesn't drop off very fast as rpm is reduced. I start off with the stock P-51D flight model but it seems like the propeller tables don't really fit this plane. I have been experimenting with swapping in the stock Hurricane propeller tables.

With a plane like this, I don't know of a source for Glide Ratios or even cruise throttle settings. The max throttle settings are a compromise. Research documents an "Overboost" setting, but the sim doesn't really support increased MP AND RPM with WEP.
Climb tests are a pain to conduct and I don't know that I really got those right, but they are my best effort.

I am not so sure ROC really falls into place so easily because it depends on a propeller efficiency at an advance ratio that is very far from where max speed is achieved. Way back with the 714th was around, there was a lot of discussion about tuning propeller efficiencies to adjust climb rates without affecting max speed. I also have a Spitfire Mk.IX in the works that is using P-51D propeller tables that accelerates way too well and climbs at 4900 fpm while the real ROC should be about 4100 fpm. The engine between the two planes is almost exactly the same but it seems that in real life, the Spitfire had a much worse propeller.

- Ivan.

Brett_Henderson
June 14th, 2010, 17:46
When I was poking around, I remember seeing something like 70knots as a clean stall speed. That seems low for an airplane of that size and power, but would fit the low wing-loading that you mention. Your 68mph translates to 59knots and is too far off (16%) for an accurate model, but well within tweaking, and certainy admirable accomplishment going at it from your methods.

Edit: Make that 70mph .. so your Vs is dead-on.. ignore the 16% references.. but keep in mind checking the key airpeeds (I've really got to stop thinking exclusively in knots.. *sigh* )

When best-glide is unpublished, something between Vx and Vy works, or if those are unknown, something near Vref is close.. and if that's not known, just go with "1.3 X Vs", as Vref.

The actual glide-ratio itself isn't as important, as getting a decent glide ratio AT best-glide airspeed. Not so much for flying an engine-out ducumentation, but as a target for putting the whole flight-profile together... especially aprroaches..

This is where we see a small problem cascade into several erroneous performance figures. If Vs is off by 16%, you really can't come up with a useable Vref, and the Vx/Vy stuff is probably off by more than 16%.... and as you try to correct all of this while working under an already highly tweaked engine, power curve; your adjustments are less predicable for performance at high power settings... For example; the first thing I'd try to do about the stall-speed, is to increase wing efficiency (and counter it with drag adjustments) , and that will no doubt wreak havoc on your cruise speeds, and high-altitude performance...

Your airspeed/power/altitude numbers are too good for me be critical, other than to re-mention that for the whole envelope, it works out better when you go after the "inside engine" stuff, after you get the in-flight stuff right, using the brute force power/thrust scalars. But of course, that's just my preference..

Is this a CFS model ? Meaning that I'll never get to play with it in FSX if you ever upload it ?

On the theory of using as realistic numbers as is possible (and how the MSFS model makes that nearly impossible) .. When I set to make an air-file/cfg set for a model... I get all of the coordinates right out of the modeling software (Gmax)... Obviuosly that works well for contact-points, and gear-compression animation.. but for wing/stabilizer/control-surface locations, not so much. When I built my Cessna 310; accurate location, area, and deflections for the elevators made it virtually un-flyable. The compromises there are ridiculous, but the important goal (ultimately) is for accurate, in-flight performance. My V-tail Bonanza was the ultimate in compromise, as I had to "simulate" ruddervators within a flight model that only reckognizes rudders and elvators. The animation though, worked out quite well.. how they respond to both pitch and yaw inputs from the pilot..

Ivan
June 14th, 2010, 18:57
Hi Brett_Henderson,
I thought the stall speed was pretty close also so when you mentioned 70 knots, I had to go check:

From: Information Intelligence Summary No. 85 (December 1942)
(Probably a summary of flight testing of Tadeyoshi Koga's A6M2 recovered from Akutan)
Stall Speeds were the following
Gear Up, Power On ==> 64 knots
Gear Up, Power Off ==> 68 knots
Gear Down, Power On ==> 53 knots
Gear Down, Power Off ==> 60 knots

Climb Rate at 30,000 feet was 850 fpm
Climb Rate at Sea Level is stated as 2750 fpm
Climb Rate at 15,000 feet is stated as 2380 fpm
My plane climbs at 3580 fpm at 12,000 feet.
My plane also achieves best climb at 170 mph which is about 30 mph higher than it should be.

Ceiling is stated as 38,500 feet but I believe this is way too high. Other sources state 10,000 Meters or 32,800 feet.

BTW, If you want to check if this plane for CFS will be useable for any more modern sim, I have a A6M5 (and a bunch of other stuff) hosted here. It is a very old visual model and flight model, but should let you know really quickly if the newer plane built with the same methods will work.

I still don't think I understand how getting the basic flight performance numbers correct with the scalars outside the AIR file and THEN tuning the engine parameters works out in the long run. Once you tune the engine power, you would cause all kinds of changes to the basic flight performance so you would need to adjust the scalars yet again?!?

- Ivan.

Brett_Henderson
June 15th, 2010, 03:52
I thought the stall speed was pretty close also so when you mentioned 70 knots, I had to go check:


It is dead on.. I caught my mistake and edited the post (in red)... I'll start working in MPH, instead of knots, so that we're on the same page :wiggle:

As for the other numbers.. Having a Vy that's 30mph off means that approach speeds will be way off too. Three numbers (Vy / Vref / Best-glide) should be pretty close as initial references when making a flight model. The amount of power from the engine (by power-curve or throttle position) will effect the ROC, but should NOT effect Vy. However, the MSFS model WILL alter Vy by power. If you adjust the power curve to match a ROC (level flight is also a ROC that exactly counters gravity), you'll be chasing Vy all over the place. This points out the first, major compromise. We have to assume that any time you'd want Vy, you'd be at high (if not full) power.. and on an approach (Vref) you'd want low (if not idle) power.

So... adjusting the power curve for airspeed per altitude first, means throwing important V-speeds out the window (ala a Vy that's 30mph off).. and if you DO go back and chase V-speed accuracy by airframe tweaks (wing-efficiency, drag, etc), your power curve (airspeed per altitude), and ROC (including level-flight) goes whacky.

In summary... if you're trying for in-flight accuracy, you'll find that what the MSFS model already does to power/altitude, makes a realistic flight model more achievable. Obviously you'll want to go back and forth (power-curve / airframe) for finishing touches, but starting off with an air-file-ized power-curve (as oppose to just brute-force power scalars and accepting the built-in power-curve), makes it way more complicated than it needs to be. Chances are that when all is said and done; the air-file tweaks will be all but negated.

fliger747
June 15th, 2010, 06:56
It is salient that the best rate of climb figure is so far off on the fast side. That may speak to the thrust being off quite a bit at lower speeds. This would also show up in poor takeoff performance. The zero at a typical weight should get off in 500 ft or so in zero wind. Even the monster F7F I am working on at the moment has a zero wind takeoff in the order of 850 ft.

The other possibility has to do with the lift drag curve being off by a considerable ammount. If this were the case one would have quite a sense of being in the area of reverse command at any sort of lower airspeed.

I don't know if you have used Airwrench at all. One of the features that Jerry has in there are very useful curves from the calculated data. I haven't used it in quite a while, but Jerry is constantly updating and improving it's capabilites. Certainly a good place to look to see if something is wierd and from whence it might issue.

Just a guess, but a plane such as the Zero shuld have a L/D ratio somewhere around 12:1. AFSD generates a stream of real time data on engines, thrust, drag, AOA, LD ratio etc. Using the FS playback feature (not available in CF2?) one can evaluate the goings on quite well, with a lot of test flying and hard work.

Good Luck, sounds like a great project. If all elese fails, steal a prop table from a similar plane of HP, Speed and performance, such as F4F.

Cheers: T

Ivan
June 16th, 2010, 14:15
Hi Brett_Henderson, fliger747

I already saw your correction regarding 70 mph. I was just trying to post the actual data I had which BTW was from IIS 82 VIA the book "Zero" by Robert Mikesh. Seems like Mr. Mikesh converted the numbers from the original MPH to Knots. If I have the opportunity, I will have to ask why he did that.

The actual numbers from IIS 85 were the following:
Gear + Flaps Up, Power On ==> 74 mph
Gear + Flaps Up, Power Off ==> 78 mph
Gear + Flaps Down, Power On ==> 61 mph
Gear + Flaps Down, Power Off ==> 69 mph

Level Speeds:
Sea Level ==> 270 mph
5000 feet ==> 287 mph
10000 feet ==> 305 mph
16000 feet ==> 326 mph
20000 feet ==> 321.5 mph
25000 feet ==> 315 mph
30000 feet ==> 306 mph

I have also attached an extract from a report from Wright Field regarding A6M2's performance.
I'll be glad to send what I have from IIS 85 if you email me at Ivan1GFP@yahoo.com.

One of the things to keep in mind is that this aircraft was hardly in perfect condition. The operators also had no flight manuals. I am fairly convinced that this aircraft wasn't quite up to the mark of a A6M2 in good condition. I believe there is actually quite a lot of inaccuracies in recorded Japanese WW2 aircraft performance due to various combinations of poor condition, bad test protocols and bad records keeping.

The test reports state that the gross weight of the aircraft was 5555 pounds. The typical quoted figure for loaded weight of the A6M2 is 5313 pounds. Even the A6M2-N floatplane was only 5423 pounds. Perhaps this accounts for some of the reduction in climb performance.

Brett_Henderson: I may try to adjust Vy when I get a chance by adjusting the propeller efficiencies.

Fliger747: I AM working on finding a better propeller. That is why I am messing with a stock Hurricane propeller. I only steal from the stock aircraft so that no one can make the claim that I stole something from their plane.

BTW, the actual consideration of propeller selection is Cp, the Coefficient of Power which is a combination of power, reduction gear, prop diameter, engine RPM, etc. I worked out a spreadsheet after reading a document from Sparks' site and doing a bit of poking around on the Internet for some additional information. I needed the spreadsheet because the numbers are very non-intuitive.

- Ivan.

P.S. My F4F-3 Wildcat is actually using a P-51D propeller. ;-)

Brett_Henderson
June 16th, 2010, 16:10
I'd like to see that data.. This is a good discussion, and I think we're all learning something.. I'll PM my email address.. :salute:

Mean-time, you're going to discover the frailty and flaws of the flight-model... and see where trying to build the flight-model in a realistic manner (ala aiming for an accurate engine power curve), just doesn't work.

The key focal point for in-flight realism, is to get Vy / Vref set, and to get them to be aprox 130% of Vs... and then build from that foundation for an airplane that will start flying AT the proper airspeed, and AT the proper length of runway,,,, and then maintains an accurate max speed at altitude, AT a reasonable power setting. That's a very tall order in itself.. If you manage to also get a realistic climb performance profile through the service-ceiling, you've done very well.

Here's the method that you've either discovered is problematic, or are just unknowingly going to attempt:


Brett_Henderson: I may try to adjust Vy when I get a chance by adjusting the propeller efficiencies.


Vy is not changeable by propeller performance, nor power (either by adjusted curve or actual throttle setting). A V-speed (any V-speed), is essentially an AoA. It's a function of the wings and airframe. Vy at 50% power is the same as Vy for 100% power. Yes, those two power settings will yield two different RATES of climb, but at either of those power settings, the best-rate-of-climb happens at the same airspeed (Vy). Not to mention that if you try to adjust V-speeds by thrust (engine power or prop performance), you'll throw the other parts of the performance envelope out the window. You have to adjust for V-speeds by airframe numbers.

Now the evil-twist from MSFS. You actually can change Vy by thrust (albeit by less of a factor than airframe tweaking.. as it should be), which is utterly unrealistic. Which is why I mentioned in an earlier post, that when designing a flight-model, you have to assume that a sim-pilot will be pitching for Vy at high power settings (and pitching for Vref at low power settings). It's just something we live with.

There's always more than one way to skin a cat.. I'm not saying it's impossible to build an accurate model around an accurate engine, but it's pretty near impossible. I've done countless flight models (even a few big jets). I started out determined to do it just like intuition would lead you.. I.E.. setting up absolutely, dead-on geometry, weights, MOIs, and control-surface deflection for the airframe... and using ultra-realistic engine specs. The initial testing reveals that only for a light-single (C172-ish), will you be anywhere close to something that can be flown realistically. I did a Saab 340 a couple years back.. and I literally gave up (as bad as the MSFS model is.. it's out of this world bad for turbo-props), and just tried to make it something that was pleasant, and a bit challenging to fly, with accurate takeoff performance, and accurate cruise performance.

If I were to build a flight model for an A6M; I'd use the P-51 air-file (and not touch it).. and then build a cfg file with realistic airframe/engine numbers. The construction would then entail tweaking airframe numbers, and power-scalars to get as close as possible for the entire flight envelope. If there was some "bending" of the engine power curve needed at that point, or some prop-performance tweaking (you'll be hard-pressed to do a better job of tweaking the prop by something other than min/max pitch in the cfg.. especially while trying to keep manifold-pressure and RPM number relationships accurate).. I'd open the air-file. Doing so before getting everything else right, makes the whole process infinitley more difficult. And 99 times out of 100, the "bent" engine won't be all that realistic anyway.

When a sim-pilot pushes the throttle forward for takeoff.. it's of little significance that the modeler can claim that his engine could be put on a virtual dynamometer, in an pressure controlable wind tunnel, and produce a data-sheet that rivaled real-world numbers (or that the model demonstrates accurate, full-throttle airpseeds at a given altitude), if it leaps off the runway and climbs like a jet (or cant' get airborne at the proper airspeed, using up realistic ruway length, and takes forever to get to altitude), or a realistic/stable approach can't be set up at/near Vref.

ANYway.. one last reminder. When you work on Vy.. don't do it from the engine/prop, and do your testing at takeoff power first. Vy should be chased by wing-efficiency an drag .. normally inverting the drag numbers (increasing induced means decreasing parasitic, etc). If Vy is unknown, just use 130% Vs. If you get it to where the vertical-speed is highest at Vy, you're off to a good start (you'll tune for the actual vertical speed by engine/prop power/thrust scalars.. or if you're ambitious, this is where you'd mess with the engine power curve in the air-file .. hoping that you keep the altitude effect on the curve to stay reasonably predicatable, and not make chasing takeoff-perfromance vs cruise-performance a tail-chasing nightmare.. (why I leave the air-file alone until the very end.. and tweak for in-sim results, not realistic dyno readings)..

:kilroy:

Ivan
June 16th, 2010, 18:19
Hi Brett_Henderson,

File sent.

I believe that perhaps the formulas you state may apply to modern General Aviation aircraft but don't necessarily apply all that well to WW2 fighters. The A6M2 Zero we have been discussing is an example where Vs * 1.30 << Vy. Vy according to IIS 85 is 105 knots or 120 mph. Vs (Clean) = 74 mph, so the Vy using this formula should be 96 mph. The Spitfire Mk.IX and F6F Hellcat are two others that don't come close to this ratio either.

I don't believe I agree with you that Vy is a constant AoA no matter what the power is. Vx I believe won't vary as much as Vy, but isn't constant either. The reasoning is thus: Drag (Parasitic + Induced) is a U shaped curve plotted against velocity. At the low speed end, it rises because Induced Drag rises even though Parasitic drag drops. At the high speed end, it rises because although Induced drag is dropping, Parasitic drag rises as to V^2. The thrust curve may be either an inclined straight line for a Jet or an inverted U for a propeller. There should be two points where the two curves cross. The low end is Vs (Really Vmin). The high end is Vmax.

At points in between, the thrust curve will be above the drag curve. The speed at which the greatest difference is found will be Vy.

So far, I am sure you are in agreement. Imagine now that we adjust the thrust curve a bit. Reduce the low end and decrease slope of the curve (because of a high minimum pitch setting) and increase the peak value because of higher horsepoer (and thrust). Now the greatest difference in thrust will be at a much higher speed.

In case you are thinking this has no place in the real world, consider that a Spitfire Mk.I or Mk.II had 1030 hp. The Spitfire Mk.IX in its final versions had over 2000 hp. There was some additional weight but the form of the airframe didn't change much. Consider the Me 109F to Me 109K transformation. Again, in form, the airframe didn't change much, but power was much increased.

I am guessing that Vx should occur at the AoA of CLmax but I am not sure. If this is the case, then consider what differences would be if the aircraft had a LOT of power as versus not enough power to match drag at CLmax. I believe an example of not enough power would be a powered Sailplane.

BTW, FWIW, all of the flight models I have put together have been based on the Stock CFS1 P-51D. They all were extensively modified though. Now as far as not touching the actual AIR file and doing all changes in the CFG file, note that there isn't one with Combat Flight Simulator. Also the Propeller tables have a couple hiccups (discontinuities) in the P-51D.

I don't believe the version of AIR file that is used in CFS1 has an option for Turboprops at all. I believe that perhaps you can simulate their effect by adjusting the propeller tables though....
;-)

Also, did you notice that my A6M2 power curve falls off with altitude more quickly than the one from IIS 85? Perhaps that explains the difference in service ceilings.

What is your protocol for testing climb rate? I know the FAA methodology which obviously works, but I don't have the patience for it.

- Ivan.

Brett_Henderson
June 16th, 2010, 19:07
I don't believe I agree with you that Vy is a constant AoA no matter what the power is. Vx I believe won't vary as much as Vy, but isn't constant either. The reasoning is thus: Drag (Parasitic + Induced) is a U shaped curve plotted against velocity. At the low speed end, it rises because Induced Drag rises even though Parasitic drag drops. At the high speed end, it rises because although Induced drag is dropping, Parasitic drag rises as to V^2. The thrust curve may be either an inclined straight line for a Jet or an inverted U for a propeller. There should be two points where the two curves cross. The low end is Vs (Really Vmin). The high end is Vmax.

At points in between, the thrust curve will be above the drag curve. The speed at which the greatest difference is found will be Vy.


Of course the drag curves by airspeed.. that's why Vy is a constant airspeed. You pitch so that the AoA (airspeed) is constant. Different power settings will result in different angles of incidence, because you chang the pitch to counter the power change and maintain a Vy AoA (airspeed). The wings don't know what the angle of incidence might be.. all that they see is the angle in which they're passing through the air.

See attached image: The available and required power lines are curved as a result of drag at increasing airspeed. Changing power would result in moving the entire purple line up or down. The point at which the greatest distance between the two power curves stays the same.. it's Vy.

Ivan
June 16th, 2010, 20:34
Hi Brett_Henderson,

What happens when the power curve changes to the one shown in Red?
Seems to me that Vy would be higher.

Also, changing power settings I believe would do more than just move the entire power available curve up and down. It would change the shape of it as well.

- Ivan.

fliger747
June 16th, 2010, 23:49
To give an example, as one approaches the absolute ceiling (for aircraft in the very low sub sonic range) power decreases, finally to a point where excess power exists only in the bottom of the "bucket". As you can see from the above examples the point of excess power (think thrust) has to move a bit to the left.

For early jets for which pure thrust is related to airspeed, the shift and effects are considerable more pronounced!

Cheers: T

Brett_Henderson
June 17th, 2010, 03:02
What happens when the power curve changes to the one shown in Red?
Seems to me that Vy would be higher.


The power curves in that image are bent due to drag changing by airspeed. That red curve would be for a completely different airframe, with different drag characteristics. The available power curve does not represent a change in the power being generated.. it represents how much of that power is useable due to the effect of drag as airspeed changes.




Also, changing power settings I believe would do more than just move the entire power available curve up and down. It would change the shape of it as well.


I'm sure it would change a little bit, but not enough to matter for a flight-sim model or even in a real airplane. That curve represents how drag-by-airspeed will "bend" a fixed amount of power.

Vy is Vy, regardless of power (where you set the throttle, or how you tune the engine). More power will result in a higher angle of incidence, hence higher rate of climb, but the highest rate of climb for any power setting is the same airspeed... Vy.

Think about a sustained climb. The power literally DOES change with altitude. When I climb into a real airplane, I pitch for Vy as soon as a climb is established.. and continue pitching for Vy through the climb. As power lessens due to altitude, the pitch angle needed to maintain Vy becomes lower.. the angle of incidence lowers.. the rate of climb lowers.. but the BEST rate of climb always happens at Vy.. and it stays the same.

Brett_Henderson
June 17th, 2010, 04:59
Try looking at it from this angle:

If you walk out onto the ramp to your little Piper Warrior, and cut the plug-wires to one of the cylinders, you've got a completely different engine, with a completely different power curve.. right ?

HOWever (if you're goofy enough to fly this airplane), as you roll down the runway, the airspeed at which you can rotate hasn't change at all.. nor has the airspeed where you'll get the best rate of climb.

It will take longer to reach rotation airspeed (and use more runway), and your actual vertical-speed will be less, but the vertical speed will be at its highest, at the same airspeed it would be at its highest, if you were using all four cylinders.

Ivan
June 17th, 2010, 12:33
Hi Fliger747,
Seems like we are in agreement here.

Hello Brett_Henderson,
The Green curve represents Airframe Parasitic + Induced Drag.
The Purple curve represents Thrust output from your engine, not drag.
The Red curve isn't a new airframe, it is just an engine with a different power curve.

This MUST be the case because where the two curves cross represent points where the thrust and drag are equal. The two points are minimum and maximum speed.

Regarding your example with the Piper Warrier:
Your Power Off stall speed is unchanged.
Your Power ON stall speed is higher and closer to your power off stall.

(I believe fliger747 was also trying to state this.)
Your Vy speed is now totally dependent on your new power curve. There are a couple possibilities:
1. You have plenty of power but your Vy speed goes down because there isn't enough power at higher airspeeds for the max surplus power to be at the same airspeed as before.
2. There is so little power that you can't even REACH your former Vy speed, so you now have a much lower Vy speed. (More specific example of Case 1.)
3. There is so little power that you can't get airborne in which case Vy is Zero.

- Ivan.

Brett_Henderson
June 17th, 2010, 15:33
I think I see how you're getting confused...

(Fliger747, you're welcome to chime in any time)


The Purple curve represents Thrust output from your engine, not drag.


No... The thrust (power) is not a variable on that graph. The purple line does not represent a changing power.. not by power-setting, nor RPM, nor altitude, nor anything. The purple line represents how much of a CONSTANT (non-changing) amount of power is left available as airspeed (drag) changes. If an engine's ouput were a variable, it would have to be on a 'Z' axis. The graph we're seeing would be a cross-section of that 3D graph.. Any cross-section would be representing power as a constant.. different cross-sections would have the entire purple line at different levels.


The Red curve isn't a new airframe, it is just an engine with a different power curve.



Again.. the engine's power curve is not in play in that graph. Anywhere along the purple line, as it goes up or down.. the engine power doesn't change.. the modulation in that line represents the power remaning after airpeed-drag takes its toll. .. i.e.. 'Available Power'


This MUST be the case because where the two curves cross represent points where the thrust and drag are equal. The two points are minimum and maximum speed.


You're agreeing with me here. Yes, they are max/min airspeed for a CONSTANT power. Think about it... if the engine's power output is a variable on that graph, the max airspeed would vary too. Max airspeed is a set point in that graph, because engine output is a constant. Think about my example of how the purple line would be at different levels on differnt cross-sections. A cross-section with the purple line elevated would indeed extend the max-airspeed intersection out to a higher airspeed. More power, higher max airspeed.


Regarding your example with the Piper Warrier:
Your Power Off stall speed is unchanged.
Your Power ON stall speed is higher and closer to your power off stall.


Yes.. the stall speeds differ by power (the power-on speed would be lower, not higher), because at the very edge where a wing stops flying, airflow generated by the prop becomes proportionally significant enought to alter the AoA, and at a high pitch angle, thrust itself becomes a vector that counters gravity. That's why I didn't mention stall speeds.. it's a different discussion. And again, it's moot. Different engine power does not come into play on our Vy graph. In the Warrior scenario, Vy does not change by power. Vertical speed will change AT Vy, but not Vy itself.

To stay one step ahead of you.. I'll ask (and answer) why the thrust vector doesn't change VY, especially on high power-to-weight airplanes. If you're using thrust as lift, you're venturing into Vx territory... a better climb angle.


(I believe fliger747 was also trying to state this.)
Your Vy speed is now totally dependent on your new power curve. There are a couple possibilities:
1. You have plenty of power but your Vy speed goes down because there isn't enough power at higher airspeeds for the max surplus power to be at the same airspeed as before.
2. There is so little power that you can't even REACH your former Vy speed, so you now have a much lower Vy speed. (More specific example of Case 1.)
3. There is so little power that you can't get airborne in which case Vy is Zero.


(1 & 3): Again.. Vy does not change. The rate-of-climb is dependent on available power, and will go down for the cut-wire Warrior, but Vy reamains the same. The cut-wire scenario would be no different that taking off with the throttle set lower than takeoff power, on an engine using all four cylinders (you still pitch for the same Vy climbing out at the reduced power setting).. or even like taking off at a high-density-altitude airport. Vy is still Vy.

(2): The lessened power will give me a lower VERTICAL speed at Vy, but it does not change Vy

Ivan
June 17th, 2010, 18:43
http://www.genebenson.com/Training%20Pvt%20-%20Comm/drag_clip/drag_wrapper.htm
http://www.slkelectronics.com/help/Eca00020.htm


Stall speed with power off is higher than stall speed with power ON.
With a smaller amount of power available, the NEW power ON stall speed will be closer to stall speed with power OFF and therefore a higher not lower speed. There are other factors such as increasing airflow over parts of the wing and tail, but the most simple factor is that the thrust line is inclined upward. Thus if the AoA is 15 degrees (stall in level flight), there is a force of Thrust * sin (15 degrees) offsetting weight.

I still believe you are interpreting the graphs incorrectly. Note that the parenthesis shows Vy at the maximum of the DIFFERENCE between thrust and drag. If the Purple line represented thrust minus drag, then the maximum excess thrust would be at the PEAK of the Purple line. As you can see in the graph, the Vy velocity is NOT at the peak of the purple line.

If the Purple line represents thrust minus drag, then why is the intersection of the two lines significant? It is significant because the Purple line represents thrust and when it meets the drag curve, the aircraft has reached its maximum level speed.

Perhaps you are getting the posted graphs confused with the graph at the bottom of this page?

http://www.allstar.fiu.edu/aero/BA-Form&gra.htm


- Ivan.

Brett_Henderson
June 17th, 2010, 19:33
You explained properly (as did I), that the thrust vector effects power-on stall speed.. but you've still got it backwards. The lifting thrust would allow you to fly SLOWER before a stall sets in...

The other charts you've indicated introduce another variable (altitude)... we have to come to grips with this graph first.. LOL : )

Now.. if the purple line is the engine's power curve, and engine power is a variable.. you'll have to explain to me why the max airspeed is a fixed point (or more importantly how it would occur at less than max thrust). I won't mind at all being proved wrong here, because that means I've learned something..

ANYway.. I can't keep repeating myself, so I'll wait for your interpretaion of why max airspeed is a fixed point. (in the morning) .. :sleep:


Edit: Did you edit the line in your last post about power-on airspeed.. or am I just that tired ?

Brett_Henderson
June 18th, 2010, 03:19
OK.. I got a good night's sleep.

We're in agreement that the amount of power effects the power-on stall speed, and why.. so we're done with that topic.

Now.. If the purple line represents an engine's power curve; by what influence is it curved ?

Typically, a piston engine power curve is graphed across RPM (throttle is the influence).

For an airplane (with a fixed throttle setting), the curve can be graphed across altitude (atmospheric/manifold-pressure is the influence).

No matter what the influence might be; if the line represents the engine's power, then that graph suggests that the airplane will continue gaining airspeed toward max-airspeed well past max power WHILE drag continues increasing. That aint possible.

Further.. if the purple line DID represent an engine power curve, the max airspeed would have to be a curved line too. More power (by RPM, altitude, or any influence), would result in a higher, max-airspeed.

If, as I'm suggesting, the power is a constant, and the curve represents the power that is 'available' for climbing as airpeed increases (the purpose of the graph is to show where/why VY occurs)... a fixed max-airspeed makes sense.

Also, consider this. If the amount of actual power (not available power) DID influence Vy.. Vy would have to be a curve too.. but in this graph, it's a fixed point on the purple curve.

To step away from graphs and theory.. allow me to use real-world data. If I were to climb into a 300HP C206, with no passengers or payload, and a light fuel load; I'm going to takeoff and climb with manifold pressure well under maximum (to save engine wear and fuel). The airspeed at which I have enough lift for rotation is the same as it would be at full power. The airspeed at which I'll get the best rate-of-climb is the same as it would be at full power. Full power will get me to those speeds more quickly (and generate a higher vertical speed at Vy), but those airspeeds stay the same.

Brett_Henderson
June 18th, 2010, 04:08
I neglected to respond specifically to a couple of yourt points.. I'll do it now:


If the Purple line represents thrust minus drag, then why is the intersection of the two lines significant? It is significant because the Purple line represents thrust and when it meets the drag curve, the aircraft has reached its maximum level speed.



It's not just thrust minus drag; it's how much thrust is available for climbing (or accelerating) as airpseed changes.

Remember, this is a graph to define Vy. By definition power has to be a constant (i.e.. climb power). With power set to a fixed amount, how does a pilot control airspeed ? By pitch, of course.

So the airspeed changes on that graph, as we pitch for different airspeeds.

Pitching for max level airspeed (zero rate of climb), the amount of power 'available' for climbing decreases as airpseed increases toward the point where 'available' power meets 'required' power.. a FIXED max airspeed, where there's no power 'available' for climbing (with more 'actual' power, max airspeed would be higher).

Conversely, pitching for a lower (climbing) airspeed moves us leftward on that graph, and the power that is 'available' for climbing increases. The point where power 'available' for climbing has the greatest seperation from the power 'required' for climbing, is obviously where the rate-of climb will be the highest.

fliger747
June 18th, 2010, 07:00
To add a little confusion.... the drag vrs airspeed curve is only applicable for a given weight.... An interesting fallout of this is that the maximum distance an aircraft will glide without power does not vary with weight, just the optimum speed varies. So the drag curve moves to the right at higher weights.

A very good explination of all of these factors is contained in "Aerodynamics For Naval Aviators"... essentially aerodynamics for the rest of us. Available in most pilot shops as a reprint.

The reason that one cannot reach the ultimate ceiling in stable flight at stall is one enters the area of reverse command where drag increases with decreasing speed. Any manuver will increase drag beyond the power available and speed will further decrease and drag will further increase. Level flight cannot be maintained and descent will occur, one way or another.

Cheers: T

Ivan
June 21st, 2010, 17:14
Hello Again Brett_Henderson,
Let's see if I can address each point in turn:



Now.. If the purple line represents an engine's power curve; by what influence is it curved ?


The Y-axis in this graph is force (either thrust or drag)
The X-axis in this graph is velocity.
The Thrust of a propeller changes because the combination of propeller pitch angle and advance ratio change as the aircraft changes speed. The efficiency at low speeds is fairly poor if the propeller cannot reach the fine pitch that would be optimal. The efficiency at high speeds is poor because the propeller blade at coarse pitch is spinning the air stream probably more than it is moving air backward. Also, as the propeller blade tips get closer to the speed of sound, efficiency drops.



No matter what the influence might be; if the line represents the engine's power, then that graph suggests that the airplane will continue gaining airspeed toward max-airspeed well past max power WHILE drag continues increasing. That aint possible.


Between the minimum speed and maximum speed, there is more thrust than drag which is why the aircraft can accelerate. If thrust is higher than drag, the aircraft can accelerate (or climb) and the AMOUNT of excess thrust determines how quickly this can be done. As the aircraft continues to accelerate toward max speed, the thrust drops and drag increases until they are equal at maximum speed.



Further.. if the purple line DID represent an engine power curve, the max airspeed would have to be a curved line too. More power (by RPM, altitude, or any influence), would result in a higher, max-airspeed.


The maximum speed is a point along the X-axis at which point thrust and drag are equal. We are not plotting max speed against altitude thus there is no curve here. Using THIS particular graph, if the throttle setting is changed or the altitude is changed, there would be another power curve entirely shaped differently than the Purple line.

- Ivan.

Ivan
June 21st, 2010, 17:19
It's not just thrust minus drag; it's how much thrust is available for climbing (or accelerating) as airpseed changes.


Total thrust minus total drag IS the amount of thrust available for climbing (or accelerating).



Remember, this is a graph to define Vy. By definition power has to be a constant (i.e.. climb power). With power set to a fixed amount, how does a pilot control airspeed ? By pitch, of course.


The throttle setting may be a constant, but the amount of thrust changes with airspeed. Even though the engine's horsepower may remain constant, the thrust changes depending on propeller pitch and advance ratio.... which determines efficiency.

- Ivan.

Brett_Henderson
June 21st, 2010, 18:24
The X axis is velocity.. we agree..

The Y axis is AVAILABLE thrust. It is consumed as airspeed increases, so less is available for climbing..OR the inverse is true too. As you use the thrust for climbing, airspeed goes down. Perfectly logical, and displayed perfectly on the graph.

You are correct in that propeller properties play a roll in how it's consumed as airspeed changes, just like airframe drag (primarlily induce as we change pitch) plays a roll... those are what curve the constant power into the AVAILABLE thrust curve.

Go back and watch the slide-show link you provided. It states clearly that power is a constant in this graph. We can assume that it's climb power (most likely full power) since it is designed to show the give-n-take between level-flight max-speed, and the climbing speeds, controled by pitch; specifically Vy. Consequently, they are both fixed points.

A graph that would show actual engine power curves would have to have something like manifold-pressure, RPM, or altitude as one of the axes.

Your link also mentions that it's for a fixed-pitch prop. But that's really of no consequence in our discussion. A constant-speed prop would probably flatten out the purple line a bit, as that prop's goal is to apply all power via a prop that maintains a constant RPM. How much the varying blade-pitch would curve the line, depends on a few more variables. But again, of no concern when you're simply trying to graph the compromises between level flight and climbing, for a PREDETERMINED amount of power ;)

I'm confident this will click for you eventually.. If not, take the meat of this thread and submit it to others for discussion. I'll be happy to explain myself if confusion continues.. I've been discussing graphs like this many times over my 30+ years of piloting.

Brett_Henderson
June 21st, 2010, 18:46
Just for reference..let's introduce the constant-speed prop, because it indeed gives us the ability vary the whole power-source.


Back to my C206... without getting silly, we could set prop-rpms, and manifold pressure to any number of combinations.. but the airspeed where it rotates stays the same, and Vy stays the same.


EDIT: But the rate-of-climb, and max-speed WILL vary..

Brett_Henderson
June 22nd, 2010, 04:21
OK.. let's try this from another angle..

This graph we've been working with, is a generic power-curve. It's purpose, as mentioned, is to show the compromises involved between level-flight, and climbing. Or it allows you to visualize being in front of, or behind, the power curve.

Every point on the airspeed axis is a pitch-controlled airspeed, and is a steady-state. There is no 'delta time'... There is no acceleration happening.

I've modified the image with relative airspeed lines that for our purposes will be: Vy(60)Cruise-climb(100)Max-airspeed(120)

At any one of those lines, we're locked into a thrust, pitch, RoC, and atmospheric-pressure. In other words, at 60, we aren't accelerating on our way to 100. And at 100 we aren't accelerating on our way to 120. Any point along the airpseed axis represents an equilibrium.

There is no axis/variable (i.e. time) that would allow for representing an actual thrust curve. The only way for thrust to be represented on this graph, is as a potential, or 'available' amount of an assumed constant., at different states of equilibrium.

If we WANTED a graph were were time was a variable. We could manipulate one where airspeed and time can co-exist on the X axis. If that assumption is applied to the current graph, we've got the physical impossibility of an airplane that continues accelerating while thrust decreases, and drag increases.

fliger747
June 22nd, 2010, 09:20
An interesting variation on this graph could be displayed in three dimensions, with the Z variable giving a depth and form. Any sorts of variable that affects the shape of the curve, weight, altitude, temperature etc could be added, giving a much better idea of how the two dimensions we are looking at here are just a slice of a particular specific condition. Then time could be added and we are venturing from vectors to tensors.....

The main and simple thing to be derived from study of such curves are where and how optimal operating airspeeds are obtained for various regiemes of flight. The second thing to be understood, a bit more complex, is how a change in conditions changes the shape of this curve.

FS generally does an OK job of simulating these curves.

Cheers: T

Brett_Henderson
June 22nd, 2010, 09:52
An interesting variation on this graph could be displayed in three dimensions, with the Z variable giving a depth and form. Any sorts of variable that affects the shape of the curve, weight, altitude, temperature etc could be added, giving a much better idea of how the two dimensions we are looking at here are just a slice of a particular specific condition. Then time could be added and we are venturing from vectors to tensors.....

The main and simple thing to be derived from study of such curves are where and how optimal operating airspeeds are obtained for various regiemes of flight. The second thing to be understood, a bit more complex, is how a change in conditions changes the shape of this curve.

FS generally does an OK job of simulating these curves.

Cheers: T

Agreed.. and if go way back to where this thread got started.. it was about modeling an ENGINE/PROP performance curve first, regardless of how that rendered V-speed stuff while in-flight.

FSX handles the "3rd dimensions" accurately enough to make chasing after them (and skewing V-speed stuff) a veritable excersize in futility (been there many times). It even does a relatively good job of modeling the differences between fixed-pitch, and constant-speed, props.

We all (me too), got a little turned around while trying to super-impose an actual thrust-curve, onto a graph that only allows for situational, fixed, relative thrust.

Actual thrust can be graphed in many ways... ala; at X newtons, Y airspeed can be achieved... all the way out to an engine/prop's max airspeed (where max airspeed could not be where there is less than max-thrust).. but this has to be at level flight (or a constant RoC, and to stay on 2 dimensions, we'd have to pretend altitude was a constant).

Ivan
June 24th, 2010, 12:48
The Y axis is AVAILABLE thrust. It is consumed as airspeed increases, so less is available for climbing..OR the inverse is true too. As you use the thrust for climbing, airspeed goes down. Perfectly logical, and displayed perfectly on the graph.

If the Y axis were available thrust, the shape of the purple graph would be different: At some speed above Zero, the available (excess) thrust would drop to zero, because below this speed, the drag would exceed thrust (behind the power curve). The shape of the curve does not appear that it has this characteristic. Also, at maximum speed, available (or excess) thrust would also be zero because all thrust is used to overcome drag. Again, the purple curve does not show this.

What it does show is an intersection between thrust and drag at maximum speed and a termination before zero speed. It would make sense if the left endpoint were the stall speed of the aircraft.

Also, please note that the label "Maximum Excess Power" is shown as the DIFFERENCE between the purple curve and green curve which is labeled as "Power Required". If the purple curve represented available power as thrust minus drag, then Vy would be at the highest point of the purple curve and it is not.



A graph that would show actual engine power curves would have to have something like manifold-pressure, RPM, or altitude as one of the axes.

As I understand it, the actual engine POWER is not being graphed here. It is the engine thrust that is being generated. The manifold-pressure, RPM, and altitude are constants as far as this graph is concerned. If any of those factors change, you would get a new curve for engine thrust.

- Ivan.

Brett_Henderson
June 24th, 2010, 14:13
You're still lost to the idea that this generic graph cannot display thrust the way you're implying..

If your red-curve were applicable (it's like an outside variable as un-applicable as TAS would be on a graph set up for IAS).. then my C206 would have different Vr and Vy for different power settings.

And even if this graph did allow for thrust as you want to represent it, your curve would be an engine so different as to make it a new airframe (as I mentioned earlier in this thread).

We're gonna have to just respectfully agreee, to disagree :wavey:

Ivan
June 25th, 2010, 10:38
Hello Brett_Henderson,

I can see your argument, but just don't agree with you. Respectfully agreeing to disagree sounds like a reasonable outcome.

Thanks.
:salute:

Hello Anthony31 (The Fellow with the ROTAX who started this discussion)

Need a couple details about your installation:
What is the propeller diameter?
What is the expected maximum level speed?
The propeller is a fixed pitch, but is it optimised for climb or level speed?

- Ivan.

Ivan
June 27th, 2010, 13:40
Hello Folks,

I spent a couple hours last night unbolting the rotary engine from my Fokker E.III Eindecker and bolting in a Rotax 912 ULS - D.C.D.I engine. Not everything fit all that well. Had to do some minor modifications to the airframe to make the thing fly straight. Also had to add some trim tabs so that the autopilot would work, but here is the result with a constant speed propeller.

From the Rotax site, here is how I read their graph:
RPM Power
5800 98
5500 95
5000 90
4500 80
4000 67
3500 55
3000 45

Here is output from the Rotax mounted in a Fokker Eindecker.
Propeller diameter is 63 inches
Idle speed is 594 RPM
Tested at 500 feet at a speed of around 83 mph TAS:

RPM HP (Actual RPM)
5800 100 (5799)
5500 95 (5518)
5000 87 (4987)
4500 78 (4487)
4000 71 (4018)
3500 61 (3487)
3000 52 (3018)

It isn't an exact match, but isn't all that far off either. I didn't try to tune the propeller tables, but scrounged Table 512 from the stock CFS P-47D Thunderbolt because its power coefficient was closer than the one from the stock CFS P-51D. I also adjusted the pitch limits so that the engine could develop full RPM at a reasonable airspeed.

Anthony31, if you PM me, I can email you the resulting AIR file.

Regards.
- Ivan.