LO's F-15 vs. the F-15 Streak Eagle

Andre “Raptor One” Joseph / Mav-jp:

 

Data was calculated from NASA sources and a public domain simulation program that makes use of the same NASA sources.

 

Anyway I remember Raptor One bragging about (kind of) confidential documents/charts/tables in his possession…:D

. . . . . if that NASA source and public domain simulation program turns out to be EngineSim, I'm going to have a good little chuckle :D

Well, it doesn't matter anyway. Even if the HFFM numbers are legit, they only apply to the F-16 - the fact that the F-15 uses a completely different, variable geometry intake changes the equation entirely.

Yep.

The static thrust value should be relatively close, though . . . . it's interesting to note that it's 5,000lb or so under the publicised value for the HFFMS as well :)

It should be possible for a forum member to recalculate the thrust figures, to be honest. We know most of the pertinent data (number of stages, compression ratio, pressure recovery at intake, con-di nozzle at the back) and can estimate efficiencies and temperature limits fairly well for that era.

I've got too much on my plate at the moment to deal with it - but someone really dedicated to the cause and with sufficient technical knowledge (failing that, technical ability and a decent textbook) could plough through it in Excel.

Then compare to Lomac and the HFFMS numbers.

Someone with a background in engineering floating around with some spare time to crunch the numbers?

edit - do I understand from reading through that the current engine lookups have been calculated from a peacetime-trim engine extimation rather than a chart available for a Vmax engine?

. . . . . if that NASA source and public domain simulation program turns out to be EngineSim, I'm going to have a good little chuckle :D

 

 

 

 

Yep.

 

The static thrust value should be relatively close, though . . . . it's interesting to note that it's 5,000lb or so under the publicised value for the HFFMS as well :)

 

 

It should be possible for a forum member to recalculate the thrust figures, to be honest. We know most of the pertinent data (number of stages, compression ratio, pressure recovery at intake, con-di nozzle at the back) and can estimate efficiencies and temperature limits fairly well for that era.

 

I've got too much on my plate at the moment to deal with it - but someone really dedicated to the cause and with sufficient technical knowledge (failing that, technical ability and a decent textbook) could plough through it in Excel.

 

Then compare to Lomac and the HFFMS numbers.

 

Someone with a background in engineering floating around with some spare time to crunch the numbers?

 

 

edit - do I understand from reading through that the current engine lookups have been calculated from a peacetime-trim engine extimation rather than a chart available for a Vmax engine?

I am pretty sure that the given NASA source was rather to save his ass (I would not be surprised if the guy works now for The Pratt & Whitney Company)

Look what I just found on rapidshare : http://rapidshare.com/files/28778357/HFFM-manual.pdf.html

The “HFFM-manual.pdf” …:D

Check out page 26, 29 and from page 40 to 47.

Check also out the standard atmosphere table used in falcon on page 13.

SK, for the God sake...

I made an error in my distance calculation. I edited my post - the distance discrepancy should have said 46%, not 228%.

if you are in doubt what gauge reading is wrong you can use F2 status bar speed to judge. Besides if M-meter at the board is to row you easyly can calculate your IAS vs altitude profile and then follow it....

As an engineer you can perform your own investigation.

After a compliment like that, how can I refuse? :)

Ok, I calculated the true airspeeds in km/h for each altitude level and flew the Dash-1 climb profile again. Here is what I found, flying in mil power more precisely this time:

Time to 40,000' in Lock On: 6:21 min

Time to 40,000' by Dash-1: 4:30 min

Average error: 41%

Distance to 40,000' in Lock On: 50.3 nm

Distance to 40,000' by Dash-1: 38 nm

Average error: 32%

However, Rhen's problem is even more severe than it appears above. This is because my "Average error" includes both the climb portion at low altitude, where the error is zero, AND the error at high altitude, which is larger than the above percentages would suggest. Consider:

The red dots represent my climb measurements in military power. Notice that for climbs to an altitude of 25,000 feet or less, the Lock On model is quite accurate - the red dots are touching the curves where they are supposed to be. The vast majority of the error (green bars) happens when the Lock On F-15C pilot tries to climb from 35,000 to 40,000 feet - in this region, the red dots are quite far from their correct locations. This means that most of the cumulative "41%/32%" time/distance error is being created in a quite narrow altitude band, where it is probably much larger than that.

(Note: the first chart is from the Dash-1 time-to-climb chart for military power, the second is from the distance-to-climb chart)

And after this great work is finished we'l be glad to know for example max L/D of F-15...

Can't we calculate this from the Streak Eagle data?

-SK

P.S. Someone please tell me if the following track plays correctly, because I might have accidentally recorded it on an incompatible beta version:

http://www.ecf.utoronto.ca/~pavacic/lomac/f15milcl.zip

SK, *only* measure distance from the point where you reach 350KCAS. Don't measure before. At least, I'm not sure if you were doing that.

SK, *only* measure distance from the point where you reach 350KCAS. Don't measure before. At least, I'm not sure if you were doing that.

I was doing that.

-SK

Ok. :) Didn't mean to doubt ya, it took me a couple times myself to figure out why I wasn't matching the -1 when I tried the max power stuff.

Thanks for your efforts :)

P.S. Someone please tell me if the following track plays correctly

It plays in 1.12a.

Here's maybe a silly question: Does the real F-15 HUD display KIAS or KCAS? If it is KIAS, why? Calibration/installation error should be known throught the envelope.

I made an error in my distance calculation. I edited my post - the distance discrepancy should have said 46%, not 228%.

 

 

 

After a compliment like that, how can I refuse? :)

 

Ok, I calculated the true airspeeds in km/h for each altitude level and flew the Dash-1 climb profile again. Here is what I found, flying in mil power more precisely this time:

 

Time to 40,000' in Lock On: 6:21 min

Time to 40,000' by Dash-1: 4:30 min

Average error: 41%

Distance to 40,000' in Lock On: 50.3 nm

Distance to 40,000' by Dash-1: 38 nm

Average error: 32%

 

However, Rhen's problem is even more severe than it appears above. This is because my "Average error" includes both the climb portion at low altitude, where the error is zero, AND the error at high altitude, which is larger than the above percentages would suggest. Consider:

 

 

The red dots represent my climb measurements in military power. Notice that for climbs to an altitude of 25,000 feet or less, the Lock On model is quite accurate - the red dots are touching the curves where they are supposed to be. The vast majority of the error (green bars) happens when the Lock On F-15C pilot tries to climb from 35,000 to 40,000 feet - in this region, the red dots are quite far from their correct locations. This means that most of the cumulative "41%/32%" time/distance error is being created in a quite narrow altitude band, where it is probably much larger than that.

 

(Note: the first chart is from the Dash-1 time-to-climb chart for military power, the second is from the distance-to-climb chart)

 

 

 

Can't we calculate this from the Streak Eagle data?

 

-SK

 

P.S. Someone please tell me if the following track plays correctly, because I might have accidentally recorded it on an incompatible beta version:

 

http://www.ecf.utoronto.ca/~pavacic/lomac/f15milcl.zip

Falcon in the old days suffered from the same problem (40k+bug?).

Falcon SuperPack2 (SP2) vs charts USAF

F15C (220 engine) 42,000 lbs fuel, clean. Time to climb profile full afterburner. The profile requires acceleration to 350 KCAS. At 350 KCAS, pull to a climb angle that allows you to maintain 350 KCAS. When you intercept Mach 0.95, maintain Mach 0.95 till the desired altitude.

SL - 10K F15C = 36 seconds, SP2 = 41

SL - 20K F15C = 1.0 minute, SP2 = 1.04 minutes

SL - 30K F15C = 1.24 minutes, SP2 = 1.33 minutes

SL - 40K F15C = 2.00 minutes, SP2 = 2.23 minutes

SL - 50K F-15C = 3.00 minutes, SP2 = 4.02 minutes

Overall, SP2 model did acceptably through 40K. After 40K SP2 time to climb fell off significantly.

3 possibilities:

A) Lockon/fc 1.12 and falcon are wrong at 40k+ and 50k+

B) Dash-1 charts are wrong

C) Both are wrong :D

2 SK

 

Please comment the following contradiction: as I posted in #175 the acceleration near climbing speed (curves slope) at 40000 in LO and Dash 1 is exactly the same.

You'll forgive me, but in post #175, it wasn't clear that you were talking about 40,000 feet at all.

At 10000 they are the same too.

I'm not sure why you keep repeating this, since nobody seems to disagree with it. If you mean to say that "same at 10,000" automatically proves "same at 40,000", then we disagree.

You should know that acceleration is the same thing as dHe/dt (altitude gain).

I'm afraid I don't understand this. I can gain altitude, even without accelerating. (?)

How could it be possible that time to climb that is a function of dHe/dt has such unbelievable error? Have you an idea?

When I followed your suggestion to calculate the TAS of Mach 0.9 at different altitudes, I was suddenly reminded that above 30,000 feet, there is an inversion - the TAS for a given Mach begins to decrease with altitude, and then to stabilize at around 35,000 feet. I began to wonder if maybe Lock On is not modelling the aircraft incorrectly, but rather - the air above 30,000 feet?

I didn't test the idea further.

Besides, the TIME TO CLIMB and DISTANCE TO CLIMB are both strictly coupled with a simple function of climbing TAS either in LO or in DASH 1.

 

How can it be that you obtain DIFFERENT errors in these cases?

If you mean my "Average error" - that is not a simple function, because the angle of the climb changes from the start to the finish of the climb. At low altitude, distance vs. time is small, while at high altitude, it's much larger. But the "averaging" only looks at the total time and the total distance. They are not proportional, so they have different percentage errors.

Anyway, the "average error" is not a very useful calculation. If you look at the green bars in the charts, the specific errors near 35,000 and 40,000 feet seem very similar for both time and distance, as one would expect.

-SK

Does the real F-15 HUD display KIAS or KCAS? If it is KIAS, why? Calibration/installation error should be known throught the envelope.

Anyone?

There's an old LO manual that says KCAS is displayed on the airspeed/mach meter and the 1.1 manual says it's KIAS. Is the error even modeled?

Bottom line: what airpeed indicators would a test pilot use to maintain the Dash-1 profile prior to the M0.9 intercept?

Both manuals are correct, this was changed with v1.1 because the real HUD displays KIAS as well. The error is modelled, AFAIK the readout in the status bar when in external view shows TAS.

Someone recently contributed this, but I don't have time to study it. If it's accurate, then maybe it can help with determination of L/D.

-SK

You can say it was me, SK.

It's from NASA Technical Paper 3414, December 1994

This one might be more helpful. :smilewink:

I don't see how LOMAC even gets close to the Dash-1 in performance. It's off in it's climb capability in MIL power. This is a little experiment I've been running comparing the -1 to LOMAC's F-15 in a 40,000Lb aircraft, clean (except the unremovable pylons), standard day.

Takeoff Procedure: YOU MUST USE FLAPS FOR TAKEOFF. Run Engines up to 80%, while holding brakes, release brakes, throttles to MIL, rotate @ 120KCAS to 10degrees, Gear/Flaps up when airborne, hold to capture 350KCAS in climb, maintain until M0.9, then climb @ M0.9.

I've also included statistical analysis for the numbers. The flight was flown 5 times, averaged, and compared to the Dash-1 numbers, to include deviation. If the Dash-1 number falls within the 95% Confidence interval, then there's only a 1 in 20 chance of the number being out of that range (p<0.05) which would be a statistically significant deviation from the Dash-1.

NOTES:

1. Time is from brake release to indicated altitude, Distance is from brake release, Fuel is from 350KCAS.

2. Fuel under Dash-1 is fuel from 350KCAS, fuel depicted below that includes, run-up, takeoff, and climb to 350KCAS, thus the apparent discrepancy.

3. If the Dash 1 number falls between the numbers, then there's no statistically significant difference between LOMAC and the Dash-1. If it doesn't... well then, obviously there's only a 1 in 20 chance that LOMAC is correct.

Finally, the real F-15 is capable of reaching 45,000ft at 40,000 Lbs. The LOMAC F-15 is incapable of reaching this altitude in MIL power at M0.9. It reaches it's combat ceiling at 43,600ft and it's absolute ceiling at 44,250ft.

Conclusion: To reject the null hypothesis that the LOMAC F-15 conforms to the F-15 Dash-1, The numbers derived from the Dash-1 should be between the numbers below the 95% Conf column. It does this at 25,000ft with respect to time to climb to this altitude. But as you can see, the LOMAC F-15 actually outperforms the Dash-1 below about 12,000ft then significantly slows/flattens it's climb profile. The slope rapidly decreases to nearly tangential at 45,000ft, which explains the large variations in leveloff time and distance.

With respect to distance, the LOMAC F-15 flys a flatter slope than the Dash-1 says the F-15 flys, above 12,000ft. The numbers begin to diverge significantly enough to notice, then become quite large in it's variance from that expected from the Dash-1. This also conforms to the large variance in time to climb to the higher altitudes.

With respect to fuel flow, the LOMAC F-15 appears to use less JP-8 than the real thing.

The final conclusion is that the LOMAC F-15 at 40,000Lbs in MIL power does NOT conform to the Dash-1. It significantly underperforms the Dash-1 at moderate to high altitudes, and outperforms the F-15 at low to medium altitude.

Finally, any errors in experimental technique are caused by two major areas. Pilot variations in manipulating the aircraft for the climb profile might cause small variations in time and distance to climb, expecially at the beginnning and end of the profile. Again, this is due to the underpowering of the LOMAC F-15 and the proximity of the combat, service, and absolute service ceiling of the LOMAC F-15 to 40,000 ft. Secondly, throttle setting could possibly be off and not fully in MIL power. Care was taken to maintain the closest throttle position without going into afterburner.

The other major source of error is in reading the tables in the Dash-1. This error is significant because some parameters must be approximated, such as drag corrections, and approximating positions on the chart near the low altitude portion of the charts.

Other sources of error are in assumptions about the atmospheric modeling of LOMAC. Is there an adiabatic lapse rate? Does it affect the engine thrust? How is engine thrust determined from altitude, speed, temperature?

This experiment will be repeated with a 30,000Lb aircraft, and a 50,000Lb aircraft to determine if the same results are obtained.

I wanted to include the track I flew, but it's too large since it ends at the absolute ceiling.

SwingKid,

I took a peek at your track and how you flew the comparison profile to 350KCAS. Some of what you're doing is a little off, which could affect the low altitude data you're trying to obtain - if a fair comparison with the Dash-1 is to be obtained.

I'm interested in low altitude performance too, but from my experience, the LOMAC F-15 doesn't seem underpowered below 10K, at least between 5-12K. Below that, there seems to be some problem, but it might be just the low altitude portion of the flight model, or the precision of the data below 5000ft in the charts, at least with respect to MIL. I haven't done a quantitative analysis of MAX power, but it doesn't "feel" right - at least for high AOA/slow speed/max G flight.

Rhen - the times and distances you cite from the "Dash-1" don't match my own, since they were measured from brake release. Didn't you suggest the Dash-1 data is only valid from the moment of reaching 350 kcas, starting the climb? Would it be difficult for you to revise your table, subtracting out the take-off run data? I think that it will still show, what you want it to show - it will just save us another 200 GGTharos posts if we all do what he says. :rolleyes: ;)

Seriously though, if you are just assuming 2 nm/60 seconds for acceleration to 350 KCAS according to the Dash-1 - I thought we had agreed this was inaccurate. I think that would explain why Lock On seems to "outperform" the Dash-1 at low alt. IMO, it's actually underperforming the real jet.

IMHO, takeoff flaps can only help properly if ground effect is modeled. It isn't in Lock On - not even for AFM. Therefore, I was unsure whether to include them would be better.

Also - can you explain again why the "Dash-1" fuel is different between your upper and lower charts? Is that a typo? Which are the correct figures to compare with the Lock On data?

-SK

2 Rhen

I performed the test at MIL to 40000 (39000 lb airplane) and my result was 6 min 20 s, i.e. 380 s from the brake release.

Can someone tell me what was Yo-yo's time to 30,000 feet in this track? I can't play it. :(

-SK

If you fly the appropriate profile, it takes about 28.8 - 31.4 sec to accelerate to 350KCAS in a MAX takeoff, and 58.5-59.3 sec in MIL, so the difference is negligible. However, I have no problems with that if you want to execute the timing from that point.

I didn't necessarily say it was inaccurate. What I said was that you should not focus on it since it represents a median value for aircraft of all weights, so the closer you get to a median weight, the more accurate that value will be.

Takeoff flaps are necessary because you become airborne at some point and have to retract the flaps, which might add to total drag at that point. Granted, this value might be negligable, but it's part of how the procedure is done. The Dash-1 tables are for takeoffs done using normal procedures. It allows you to become airborne at a slower speed, which will add to induced drag as well, since you're at a faster speed when using a flaps up takeoff.

Yo-Yo:

One of the inaccuracies about most western jet engines is that the RPM increases in afterburner, or needs to be 100% when this happens.

Nevertheless, for my tests, I pushed the throttle forward until it was as far as it would go without causing the afterburner to light. Remember, my data is for a 40,000lb aircraft (not 39,000) which might explain why your time to 40,000ft is so low. Other possibilities are that it has to do with your technique, or what airspeed you're using to climb. Are you using TAS instead of KCAS?

However, if I'm setting throttles too low, it still doesn't explain any discrepancy that's present in the data at and below 10,000ft.

Also - can you explain again why the "Dash-1" fuel is different between your upper and lower charts? Is that a typo? Which are the correct figures to compare with the Lock On data?

 

-SK

See Note 2