FSXA Too much drag/anemic engines(?)

[Heretic]

Update:

1) Pitch
I've adjusted the table for values I found somewhere in a book and I get 3° nose up atitude at 0.8 mach cruise. About normal-ish for a 727(?).
Also, according to Yves, only the slope in 404 matters anyway.

2) Drag
Hacked my calculated 430 and 154a tables into the air file and performance considerably improved. In fact, the 727 now flew a bit *too* well, so...

3) Thrust
...I've gone back to my very old net thrust table (1506; jx spec) with which the whole misery started. Performance got more into the ballpark.


Now for the fine tuning. Can I tweak drag with CD0 in 1101 and fuel flow with TSFC in the aircraft.cfg? Or is it better to rerun the numbers for the tables with, say, more accurate ram drag calculation?


All in all, I'm pretty happy with the progress so far.

 

[Heretic]

A bit of tweaking (adjusting 154a and TSFC in the aircraft.cfg) later...

Idle fuel flow within 40 lbs of the reference value (which is good, considering MSFS' TSFC behaviour), cruise fuel flow (only checked at FL310 and M0.78 and M0.8) is within 10 lbs of the reference value or sometimes even right on the money.
Spool time is okay, EPR is okay, N1-N2 is okay.

The 727-200 can now be flow by charts.


Thanks for everything, guys!


Will adapt table 430 to the new values in 154a and then post the aircraft.cfg and .air file here.
And then tackle the JT8D-15 and maybe the 727-100. Or the JT8D-200 from the MD-80...


- Edit: Here it is...

https://www.mediafire.com/?cxjh4iyc9r5gjfx

 

[Heretic]

Oh dear friggin' lord...

AAM did not fully import 1506 (because it does not seem to be able to handle more than six columns), which means that FSX' "thrust vs. mach" calculation stopped dead cold at mach 0.5. Cue too much thrust at 0.78 and me happily ramping up drag in turn.
And I've only noticed it as I was about to redo the exercise for the JT8D-15.
I'm back to square one. But first, I'll go cry in the corner over there.

 

[jx_]

Can I tweak drag with CD0 in 1101 and fuel flow with TSFC in the aircraft.cfg? Or is it better to rerun the numbers for the tables with, say, more accurate ram drag calculation?

Can you rephrase the question? I don't understand. What are result are you looking for by tweaking drag and TSFC and how are they related?


Sorry about your technical hiccup as well!

 

[Heretic]

I've used 1101(CD0) and/or 154a to artifically increase drag so that I need more thrust and thus get higher EPR output (as these rely on thrust output). TSFC was used to tweak fuel consumption at altitude.
(If you still don't understand it, never mind. It's not relevant anymore.)

A big part of fixing the problem was deriving a new method of calculating EPR. The old logic played reasonably well with the "wrong" 1506 table, but as the "new" table outputs less thrust at altitude, EPR values at cruise and altitude run in way too low.
What I did now was basically setting up a 1507-like table in the gauges to correct thrust output for ram drag and then correct the result for IAP to receive essentially a sea level thrust output for the EPR gauges at all speeds and altitudes. A huge disadvantage is that I have to tweak these gauges whenever I make changes to 1506, but on the other hand a net thrust 1506 table and said gauge logic offer a more accurate solution than using gross thrust in table 1506 and airflow factors in table 1507.

 

[Heretic]

Update:
My last solution featuring a table 1507-alike method to obtain accurate EPR values led nowhere. I was either running in too high or too low and never as much as I've wanted.
I found a less than elegant solution for the problem though. Out of desperation, I've mapped the EPR display to N1 (non-corrected) and believe it or not, I got better results than with thrust.
Tweaking table 1506 a bit für ram drag sealed the deal for the cruise part of my test scenario. A short test on the ground also yielded correct N1 values and a spoken "You S.O.B...".

I just hope that I didn't accidentially hit a "sweet spot" like I did with the "wrong" table.

 

[jx_]

I see two problems.

First, you MUST get drag correct first, then thrust, then EPR....in that order. Otherwise you will chase your own tail forever. Input the drag polar to the airfile, then using either a calculator, a spreadsheet, or flying in FS, record the drag value to maintain speeds that you have charted. This is thrust required of the real aircraft.

For example, our charted performance on the cruise charts indicates at FL290 M0.80, the real engine output is 85%n1. But when we do our calculation of thrust required and cross check that value to 1506, we might find our 1506 thrust is in the wrong place. Important: Correct thrust required and correct n1 before comparing to 1506.


Second, EPR is rear Delta2 / frontal Delta2.. This is Delta2 inside the inlet and turbine exhaust. You are thinking too much about the external ambient atmosphere and trying to find a fit from the wrong parameters. It doesn't work that way. Using your EPR output at off temperature or off speed will induce an error. For EPR you either need mass flow and accurate geometry, or you need the actual pressure/velocity readings charted to some corrected parameter. I'm going to save you some trouble here.


1. Make an accurate SSL EPR table and design an output system to send this to FS through your gauge. This can be tied to CN1 (TRIGGER VALUE). This should be tested successfully in the sim at sea level ISA with the brakes locked before continuing to step two.


2. Chart EPR from the real 727 by CN1 by Mach. So for example, if at FL290 and M0.80, the engine output is 85%n1 and 1.35 EPR; you would correct the RPM to CN1 then record 1.35 EPR next to that value. Viewing this in charted form, you will probably notice the EPR decreases with mach, but may become steeper with RPM. This is the curve you need to fit to. Use CN1 to produce an SSL EPR value, then modify that value 'live in the gauge' by Mach using a formula you create.


SSL base EPR * ( MACH CURVE FORMULA ) = AMBIENT EPR

CN1 changes with temperature, so your ambient atmosphere is already taken into account. No need to worry about altitude or temperature. Just use EPR from CN1 by MACH and you'll be fine. You're EPR gauge will react to temperature altitude and temperature velocity.

After you're done you can add features that check the thrust output and reduce EPR anytime thrust is less than expected (failures while in flight, etc.)

 

[Heretic]

Thanks for still sticking with me. I'm not sure what this thread has degenerated into, but it might just be quite painful to read for others by now.

First, you MUST get drag correct first, then thrust, then EPR....in that order. Otherwise you will chase your own tail forever. Input the drag polar to the airfile, then using either a calculator, a spreadsheet, or flying in FS, record the drag value to maintain speeds that you have charted. This is thrust required of the real aircraft.

I've already done that with basically your method (CD-CDp), so drag is not the point anymore.

Second, EPR is rear Delta2 / frontal Delta2.. This is Delta2 inside the inlet and turbine exhaust. You are thinking too much about the external ambient atmosphere and trying to find a fit from the wrong parameters. It doesn't work that way. Using your EPR output at off temperature or off speed will induce an error. For EPR you either need mass flow and accurate geometry, or you need the actual pressure/velocity readings charted to some corrected parameter. I'm going to save you some trouble here.

1. Make an accurate SSL EPR table and design an output system to send this to FS through your gauge. This can be tied to CN1 (TRIGGER VALUE). This should be tested successfully in the sim at sea level ISA with the brakes locked before continuing to step two.

2. Chart EPR from the real 727 by CN1 by Mach. So for example, if at FL290 and M0.80, the engine output is 85%n1 and 1.35 EPR; you would correct the RPM to CN1 then record 1.35 EPR next to that value. Viewing this in charted form, you will probably notice the EPR decreases with mach, but may become steeper with RPM. This is the curve you need to fit to. Use CN1 to produce an SSL EPR value, then modify that value 'live in the gauge' by Mach using a formula you create.

SSL base EPR * ( MACH CURVE FORMULA ) = AMBIENT EPR

CN1 changes with temperature, so your ambient atmosphere is already taken into account. No need to worry about altitude or temperature. Just use EPR from CN1 by MACH and you'll be fine. You're EPR gauge will react to temperature altitude and temperature velocity.

Assuming that the given real values for N1 at ISA SL (=CN1) and EPR are to be trusted, I can simply let FSX handle EPR decrease with altitude and speed, right?
This is achieved by means of [SL ISA N1]*sqrt(theta2)...at least that's what Roy states in his paper.


In other news, I think I've cracked the "Thrust vs. EPR" method. ISA SL ram drag is calculated in the EPR gauge with mass flow, W/g, etc...to match the values calculated in the Excel/Calc spreadsheet during net thrust calculation. It is then summed up with [Current Thrust]/delta2 to obtain sea level gross thrust. To simulate altitude and speed dependant EPR decrease, the resulting thrust is then multiplied with sqrt(theta2) before being used in a "Gross thrust vs EPR"-style table.
This does, however, not yield the same results as the "N1-EPR" method above.

The question is, however, which method is, by principle, the more suitable one. I can imagine that, in the real engine, EPR is more closely linked to thrust because a pressure increase by means of injecting more fuel is quicker to achieve than by overcoming the first stage fan's inertia and air resistance. But feel free to prove me wrong.

I know that I need to tweak the input data for both methods in any case. The cruise charts I have are for the -15 series engine and I'm doing the -17A which already behaves a wee bit different at static SL. Should have done the -15 from the start... *Cough*

After you're done you can add features that check the thrust output and reduce EPR anytime thrust is less than expected (failures while in flight, etc.)

...bleed air...

 

[jx_]

the problem with EPR is it is based on exit velocity versus inlet velocity. As you can imagine, the slower the inlet speed, the higher the inlet pressure and vice versa for the exit. Yes it is directly related to thrust, but to find EPR from thrust you need to find the velocity (where the probe is). Otherwise, you need detailed data on the engine compressors and turbines.

The simple way around is to use a table look up that is similar to table 1506. This is essentially EPR at all speeds at sea level. By doing it that way, and using CN1 versus Mach as the trigger... the EPR will be very accurate at all speeds and altitudes.

Assuming that the given real values for N1 at ISA SL (=CN1) and EPR are to be trusted, I can simply let FSX handle EPR decrease with altitude and speed, right?
This is achieved by means of [SL ISA N1]*sqrt(theta2)...at least that's what Roy states in his paper.

No. Roy was talking about setting up a table. We're talking about aligning real data to our table. So the reverse applies. Real N1 at Mach 0.XX and altitude XX,XXX / sqrt(theta2) = CN1

So you take the real N1/EPR charts, convert N1 to CN1.
Plot the EPR by Mach.
Convert that into a table that looks like table 1506 or a formula that will output the correct curve fit.
Then use FSX CN1 to drive the input of the gauge.

Theta2 will handle EPR decrease with altitude and speed. You will see the correct EPR at the correct N1.

 

[Heretic]

Okay, I've converted my engine to a JT8D-15 and tweaked it for fairly accurate N1 and N2 values at cruise.

Theta2 will handle EPR decrease with altitude and speed. You will see the correct EPR at the correct N1.

According to my charts, I don't.
I've got one for N1 and N2 over TAT and I've tweaked the engines according to this table. Nothing all too accurate, but I hit the numbers within 0.5 to 1% CN1 at cruise.
EPR is CN1 driven with a curve putting out correct values at SL. But according to my "EPR by mach and gross weight" chart, I'm way off during cruise. Dividing by sqrt(Theta2) should, but does not fix this.
It's a true headscratcher. Faulty source data?


- Edit:
I've fixed it with a corretion factor calculated from theta and theta2. Not the most elegant solution, but as long as it gets stuff done...

All that's left now (bar any further life-treatening bugs) is fuel flow.

 

[jx_]

you're trying to correct Sea level EPR. Don't. Set up EPR vs MACH at sea level. The Sea Level CN1 that drives EPR should be set in stone for any given Mach. Now simply drive a gauge that reads CN1 from the sim then outputs the EPR correlated to current mach. This is EPR at Mach driven by CN1...similar to how table 1506 is thrust at mach driven by CN1.

Dividing by sqrt(Theta2) should, but does not fix this.

No. Dividing by sqrt theta2 is used to FIND sea level static RPM values from non sea level static RPM values. Taking sea level data and dividing it by sqrt theta2 won't render anything useful.

For example: N1 at FLX and Mach X / sqrt theta2 = CN1

 

[Heretic]

you're trying to correct Sea level EPR. Don't. Set up EPR vs MACH at sea level. The Sea Level CN1 that drives EPR should be set in stone for any given Mach. Now simply drive a gauge that reads CN1 from the sim then outputs the EPR correlated to current mach. This is EPR at Mach driven by CN1...similar to how table 1506 is thrust at mach driven by CN1.

I don't have enough source information to do that.


Besides, I already get the numbers I want at altitude and mach...

 

[jx_]

okay onto fuel flow then.

 

[Heretic]

My sources differ regarding fuel flow, so I've just set it for a middle value so far.

 

[Heretic]

Done a flight with the new airfile. It was quite pleasant. Good numbers during cruise, climb was okay (I have no idea how well a real 727 climbs), but I had a really hard time slowing the thing down during descent.
Had I kept it at 1500 fpm longer, the DC-8 wouldn't have been the only civvie plane to break the sound barrier.

C_D is out of the question(?), so I guess I have to revisit 404.

Are tables 410 to 413 important?
I know they contribute to C_L, but I do not have data for these. Remove? Set to one or zero?

Also, what does 1101 - "Fuselage AoA at min induced drag" do (practically)?
The 737 has this set at -3.something (degrees?) because it cruises nose down?




- Edit:

"A Case study in aircraft design: the Boeing 727"

If anyone finds this online or in a shop, leave everything be and TELL ME WHERE.
I'd sell out humanity for this book!

 

[jx_]

Real jets are very slick. if you are using the drag polar and have properly separated CDp and CDi into the tables then your drag is right.

table 410/413 are stability derivatives; CL_de is elevator download. To produce a nose up moment ahead of the CG, there must be a tail down applied aft of the CG; and since the wing is aft of the CG, the entire tail download is loaded onto the wing. This basically means as the elevator is deflected, the wing's AOA increases to maintain Lift = weight. If you don't have data on this it is best to use small negative values and test the clean lift off speed. If your thrust and lift are correct, you will be able to find CL_de without needing the data.

AOA at min induced drag is the bottom of the CL_LINEAR line; but it is no longer read (afaik). FS9+ reads table 404 directly and uses AoA @ CL = 0

 

[Heretic]

I've sorted out my drag problem. Apparently, I've had two y-values at x=0...
Used the opportunity to redo 404 based on the one from the Vistaliners 727 as it looked quite sound.
It works so far. I can finally slow down again, lol.


Also did something I should have done long ago: Putting the "N1,N2,EPR for EPR vs TAT" table from the performance handbook into spreadsheets, chopping up the data and using it to build table 1502 and EPR and EGT gauges from it. It's way more detailed than what I worked with before and I could get rid of any correction factors for EPR. It's just a plain old table driven by CN1 now.
Stubbornness is an awful thing...*ahem*



- Edit:
Just happened to watch a home video made from a 727's cabin. Apparently, spoiler use is quite common in that bird.
I'm a bit relieved.

 

[Heretic]

Took the latest revision for a spin on sunday and everything went quite well. With some 150000 lbs gross, I still got - at moderate EPR settings - some 500 to 600 fpm out of it near FL320. Not bad.
Initial climbout after t/o looked quite a bit as in the videos I've seen. A 727 apparently isn't the best climber unless you give it some speed.
And on landing, I experienced the best quirk of 'em all. A 727 needs to be flown onto the runway due to the huge amount of drag from the flaps. If you close the throttle too early, you'll get a few passengers with back problems. Now guess what I did...

Anyway, most credit for the flight behaviour has to go to Vistaliners and HJG as their table 404 and flap drag and lift coefficients are driving this model.


Let's see if I manage to find the time to model the other engine variants now.

 

[jx_]

nice. I have a friend who once flew the 727 for Delta. He used to tell me stories about the unofficial SOP of extended from slats to full flaps at 200 feet and flying it on. Glad it worked out for you.

 

[Heretic]

Yup. It's been a long, long ride...