Engine Performance Tuning Tutorial

[Ivan]

This tutorial is about tuning Engine Performance, but the end result is to make our Flight Simulator aircraft perform as the actual aircraft did. There are factors in the simulator that we cannot control and there are also non-linear factors and interactions between parts of the aircraft that are not captured by a purely mathematical model.

So far we have been discussing Engine Output in Horsepower and I believe we have proven that we can tune that as precisely as we have patience for. One factor that we have not taken into account at all is the Engine's Exhaust Thrust which is shown at the bottom of the Jumo 213A graph.

This is usually on the order of several hundred pounds of extra thrust added to that of the propeller and is non trivial.

As an illustration, we can take the case of the Japanese Type Zero Fighter. The A6M3 Model 32 and A6M5 Model 52 have almost exactly the same airframe but for wing tips and minor details. The Engine is the same. The Model 32 uses an exhaust collector manifold while the Model 52 uses ejector stacks for exhaust thrust. The maximum speed of the Model 32 is about 335 mph. The maximum speed of the Model 52 is 350 mph or around 15 mph higher.

I don't know of a "correct" way to simulate exhaust thrust which has its greatest effect at high speeds.
The two ways I can see to do this are:
1. Decrease Airframe Drag which would have its greatest effect at high speeds.
2. Increase Engine Power.

In this case I chose to increase Engine Power.

[Ivan]

Guessing the amount of Engine Power required wasn't difficult. I simply used the MW50 War Emergency Power and throttled back until I was getting a bit over 426 mph. Perhaps my joystick (Sidewider Precision Pro) was acting strangely, but the actual power output seemed to jump around a bit. A touch over 1500 HP seemed to work.

As a guess based on the previous tests, I used 3.24 for a new Supercharger Boost Gain:
1511 HP gives 428 mph.

There are three issues that this test brings up:

1. I usually try to get 2-3 mph better speed than the documented tests show. The reason for this is because my testing is done with an autopilot while the original test would be done by a pilot with a kneeboard and only holding altitude and direction by hand or adjusting trim. If we Simulator pilots tried the same thing, we would probably also lose a bit of performance because we can't hold as steady as the autopilot can. There also tends to be some variation between aircraft on a production line. We want to give our pilots the best of the lot.

2. In reality, MW50 would not have done a thing to increase power at this altitude. This is a problem with the simulator and we can't fix it.
MW50 with the Focke Wulf and Higher Octane Fuel with the Spitfire increase the ALLOWABLE Boost of the Engine. They allow the Engine to use excess supercharger capacity where the throttle limits are because of temperature or detonation. At altitudes where the Engine is already using all of the supercharger's capacity, they obviously can't increase the supercharger's pumping ability.
GM1 or Nitrous Oxide injection on the other hand is basically a consumable supercharger. It works at any altitude where the supercharger cannot supply enough boost.

3. Engine WEP and Combat settings are mostly procedural "Paper" restrictions. Except for cases in which there is a limited supply of anti-detonant (Water Injection, Methanol-Water Injection), there isn't anything mechanical other than a shorter operating life to the next overhaul or temperature limits that prevent continuous full power operation. Running out of Water Injection while running manifold pressures that require it will almost certainly kill your engine, but that infamous 5 minutes 10 seconds of WEP in the Mustang before the Engine is crippled is a myth.
The maintenance people probably don't want to be replacing spark plugs after every flight or pull an engine after three flights for an overhaul, but bringing back a tired engine beats not bringing the engine back at all.

....

[Ivan]

Now that we have proper engine power at Critical Altitude and Sea Level, the next step is to see how the entire curve looks.
I generally test Engine Power at 2500 foot intervals. It is hard to test at Sea Level so the "Sea Level" test is actually at 500 feet.
Getting these numbers took about 5 minutes.

Getting the graph into a Spreadsheet took a couple hours.

The Red Line represents the power output we are trying to achieve.
Notice the big dip in the Red Line?
That is the switchover between the Low gear and High gear of the blower.
The Blue Line is what we are actually achieving.

The big difference here is that CFS doesn't handle multi-speed superchargers.
We are always tweaking a single speed supercharger.
If the numbers match up with an actual multi-speed supercharger at altitude, they will be pretty far off at some intermediate altitude.

We are around 300 HP or 20% too high at 15000 feet.
Any CFS Aircraft with a multiple speed supercharger will have this issue. It is just a matter of degree.
For CFS, the maximum level speed of an aircraft is generally quite predictable: it will be either at the altitude the engine generates maximum power or at the next higher altitude.

For this FW 190D, we get the following:
00500 feet ==> 1721 HP - 364 mph
15000 feet ==> 1971 HP - 437 mph
17500 feet ==> 1823 HP - 437 mph
20000 feet ==> 1632 HP - 432 mph
22500 feet ==> 1454 HP - 426 mph

Note that we are only a little over 1% off at Sea Level and nearly dead on at critical altitude, but around 10 mph off at medium altitudes.

A quick Service Ceiling test was also done in the following manner:
1. Put the aircraft in the simulator at 36500 feet with an expectation that the actual Service Ceiling will be around 37400 feet.
2. Throttle back to see what the minimum flying speed will be before the aircraft stalls.
3. Use full throttle to put the aircraft back to about 10 - 15 mph over that stall speed.
4. Set the autopilot to maintain a 500 feet per minute climb.

The low speed eliminates the possibility of a zoom climb.
Gradually the climb rate should drop off.
Note that altitude at which the aircraft can maintain at least a 100 feet per minute climb.

I conduct this test with a full internal ammunition load and with at least 50% fuel because weight greatly influences the result.
The results from this test were the following:
Service Ceiling: 37100 feet
Engine Power: 626 HP
Fuel Remaining: 120 gallons (Full Load is 150 gallons)
True Air Speed: 218 mph

I don't know that this is a terribly rigorous way of doing things, but the results are pretty repeatable and it seems to work well enough for me.

....

[Ivan]

Now that we have gotten this far, believe it or not, we are about to start over again.
The difference is that this time we have a VERY good idea of the values we need to achieve the intended flight performance.
The current results are not bad, but are way too high at medium altitudes such as 12500 feet to 17500 feet.
Since we are making specified power at 500 feet and still have a few mph too much speed, we can probably afford tune the power below specifications and still have reasonable performance. This should also work to reduce the power at medium altitudes.

There are three Power @ Altitude combinations we are trying to match
At 500 feet, we are currently achieving 364 mph on 1721 HP.
Because speed is proportional to the Cube Root of the power difference, we should be able to achive 360 mph on 1665 HP.

But wait!
Kurt Tank's test run was at 300 meters altitude which is 985 feet.
At that altitude, we are currently going 365 mph on 1730 HP.
If the speed increase 365 / 360 is proportional to the Cube root of 1730 / (New Power), we can predict we should need 1660 HP at 985 feet.

At 500 feet, we are looking for 1660 HP to 1665 HP.
(Yeah, I know the range starts a few HP lower, but I like nice round numbers.)

At 21650 feet, we are still looking for 1511 HP.
(Nothing has changed there.)

We are hitting the Service Ceiling at 37100 feet with around 626 HP.
We want the Service Ceiling at 37500 feet where we are currently making 607 HP, so....
At 37500 feet, we are looking for 626 HP.

With this information, re-adjusting the necessary parameters is easy....

....

[Ivan]

Adjusting the Torque down a bit (0.555 to 0.54) gets us
1661 HP @ 500 feet with a speed of 359 mph

Adjusting the Supercharger Boost Gain up a bit (3.24 to 3.33) gets us
1511 HP @ 21650 feet for a speed of 428 mph

Horsepower was still
607 HP @ 37500 feet which is a bit below the 626 HP that we need.

To address this, we need to reduce the Friction Loss and also reduce the Torque multiplier.
A single adjustment is unlikely to get the values we want:
1660 HP @ 500 feet
AND
626 HP @ 37500 feet

so we see where the values are off, adjust and re-test.
If the Power at Low altitude is too high, we adjust the Torque down.
If the Power at High altitude is too low, we adjust the Friction down.
Eventually the incremental changes get smaller and we get the results we want.
It took me about 15 adjust and test cycles.
The end result was the following:
1663 HP @ 500 feet
1516 HP @ 21650 feet
627 HP @ 37500 feet

A few more cycles and more decimal places could get us closer, but these values are close enough for me.
Besides, these kinds of repetitive edit and test cycles are BORING!

I still have not re tried the speed and service ceiling tests, but a couple HP should not make a significant difference.

Attached are the Table and Graph of updated Engine Power output.

- Ivan.

[Ivan]

Re Testing the flight performance should have been easy.
It was.... Up to a point.

The Speed Run at 500 feet altitude gave 359 mph
The Speed Run at 21650 feet gave 427 mph with 1516 HP.
The Highest Speed was 437 mph at 17500 feet

So far so good....

The Service Ceiling test was impossible to reproduce reliably:
One run would give 36900 feet. The next run would give around 38000 feet.
The Engine Power was as expected, but the power required and speed were not consistent.
When the Service Ceiling was too low, I even tried (successfully) to boost the engine power from the predicted 626 HP to 660 HP, but each attempt would end differently and unpredicably.

It finally occurred to me that the wacky looking Coefficient of Lift Graph might be to blame.
See the attached screenshots.
I believe that the original P51D CL graph is such that the Autopilot can not reliably detect the Stall.

I edited the graph a bit at the higher Angles of Attack which should not affect the speed tests.

[Ivan]

The end result of the edit to the CL Graph was that the Service Ceiling test worked out pretty much as expected.
The airspeed is a touch higher, the Power required is a touch lower and the Service Ceiling is less than a hundred feet higher than predicted, but everything is in the "acceptable" range.

I expect that a test by a human pilot instead of autopilot would Stall the aircraft a couple hundred feet lower because this aeroplane is VERY unstable with so little power at this altitude range.

- Ivan.