Engines go mad after start

[trisager 16]

We have seen this a few times in the beta, but I personally have never been able to re-create it.

On my system it will happen on every takeoff with engine anti-ice enabled. Engines go to 100% and I just have to wait until they decide to settle down, 30 seconds or more in some cases.

Tom

[Aerosoft Team [Inactive Account] 51558]

On my system it will happen on every takeoff with engine anti-ice enabled. Engines go to 100% and I just have to wait until they decide to settle down, 30 seconds or more in some cases.

Tom

But as we do nothing more than control FSX functions, this should appear as well on other aircraft.

People sometimes got a incorrect idea on flight models. It's is not that the developer can make the aircraft fly. The developer sets the parameters and starts functions of the sim. We don't 'program' anti-ice, we switch it on, what happens after that is what FSX thinks should be done. In a turbine engine there should not be any link between de-ice functions and engines (piston engines have carburettors that need to be heated in icing conditions).

[trisager 16]

But as we do nothing more than control FSX functions, this should appear as well on other aircraft.

People sometimes got a incorrect idea on flight models. It's is not that the developer can make the aircraft fly. The developer sets the parameters and starts functions of the sim. We don't 'program' anti-ice, we switch it on, what happens after that is what FSX thinks should be done. In a turbine engine there should not be any link between de-ice functions and engines (piston engines have carburettors that need to be heated in icing conditions).

I believe that the engine anti-ice on the Cheyenne is some kind of inertial separation system that keeps ice from reaching the tender parts of the turbine? The observable effect in the Cheyenne is a drop in engine performance, but whether that is through direct manipulation or via FSX functions i wouldn't know. The Pilatus PC-12 from Flight1 has a similar system, but I have not seen this particular issue with that one.

At any rate it was mostly meant as a suggestion of a way to reproduce the issue if you want to.

Tom

[Ken Stallings 1]

Could you confirm this does not happen with the default turbine props in FSX?

I'm sorry, could you elaborate? What do you mean by "default turbine props?"

Cheers,

Ken

[JAL 0]

Hi!

After some testing it seems that the problem is essentially a Microsoft bug (in my humble opinion), which involves these variables:

Explanation of variables:

1. fuel flow gain: how fast the engine reacts to power inputs (variable modifiable in aircraft.cfg)

2. static thrust produced by the engine: The thrust produced from the engine by the exhaust gases as measured when aircraft is not moving. (variable modifiable in aircraft.cfg)

3. condition lever simulation: Similar function (but not the same) as mixture lever in a piston aircraft. In turboprops some aircraft have high and low idle positions (e.g., King Air) and some do not (e.g., Twin Otter).

Here's how the variables affect the engine and start-up:

1. If you increase fuel flow gain you get a surge at engine start (most noticeable in the torque and fuel flow gauges)and engines react more quickly to power inputs.

2. If you increase the static thrust produced by the engine there is a greater surge at start up.

3. If you push the condition lever to high idle during a manual start engines surge more at start than if you just put it to low idle.

Experiments with default King Air:

On the default King Air 350 fuel flow gain is set low: 0.011. The engine reacts pretty slowly to power inputs, which I have a feeling is unrealistic for this aircraft. The start is fine though. This is also helped by the fact that the King Air has a low idle condition lever position.

If I increase fuel flow gain in the aircraft cfg in the King Air and do a manual start where I push the condition levers to high idle I get the same engine surge that we now get with the Twin Otter.

Now to the Twin Otter:

1. Fuel flow gain

Steve has confirmed that engines do react fairly quickly to power inputs in the Twin Otter (although there is certainly some delay- such that a turbine pilot has to anticipate power changes ahead of time, whereas a piston pilot does not). The question here is just how high should we set this variable. Currently it is at 0.08. I've also tested out a setting of 0.06, which actually seems pretty good to me. Our Twin Otter pilots on the board may want to try changing this setting (maybe try 0.04?) and recommend what they feel is the most realistic. Let me know if you need help in editing the aircraft.cfg. As Steve pointed out the biggest problem of having the engines react too slow is on flare. You put the power lever to idle and it takes the engine forever to spool down causing more float.

2. Static (exhaust) tthrust

Currently static thrust is set at 88 (lbs) which comes very close to the real life value of 90. The Twin Otter 300 should be able to maintain or even accelerate while taxiing with power levers at idle. Part of the reason for this is the thrust produced by the exhaust. Lowering this value would help the start up, but then taxi behaviour will become unrealistic and the total amount of horsepower produced by the engines will be incorrect (I could compensate for this, but then you get into entering many unrealistic values for other variables which I think is a drawback).

3. Condition levers

The twin otter only has on/off condition levers. If you do a 'ctrl-e' start you will notice that conditions levers are put into a low idle position (so unrealistic), but the engines don't surge as much. I think this is where the root of the FS bug lies. FS has been programmed to understand that all turboprops have a low and high idle posiiton and that low idle should be used for starts. Putting the lever all the way forward (as we should in the twin otter) makes FS understand that a lot of fuel should be added at once. Hence our problem...

Can we solve this problem?

It's interesting that Mathijs has not had this problem and Ken only has it intermittently, and it may be a reason to hope that we can solve it. I wonder if this bug has been fixed in Acceleration or SP2? I only have SP1 and no acceleration.

How can we work around the problem if no solution found?

1. Do a ctrl-e start. Then move your condition levers full forward.

2. Do a manual start, but only put your condition lever at low idle (or even less). Then once engine is started move it forward.

3. For start keep the props feathered. This is an approved method in the manual (apparently in real life the advantage of starting with prop levers full forward is that it results in a cooler start- something not modelled by FS). This will at least prevent your props from roaring (but torque still goes way up).

An unrelated problem

Finally, it sounds like some people have experienced that power gets stuck at 100% and doesn't come back down. This is almost certainly a different issue -probably joystick calibration-.

Anyway, hope this helps explain the situation and anyone who has ideas to help solve it please let me know!

Cheers,

John

[Aerosoft Team [Inactive Account] 51558]

I'm sorry, could you elaborate? What do you mean by "default turbine props?"

Cheers,

Ken

The aircraft with TurbineProps that come with FSX.

[JAL 0]

Ken and Mathijs,

Regarding the default turboprops: please see my post above and the section on the King Air. The start up surge also happens with it if you increase fuel flow gain in the cfg and put condition lever full forward at start. This confirms it is basically an FS problem.

Mathijs, when you start up the Otter do you do a manual start or ctrl-e? If manual do you push the conditions levers full forward too? If you still don't get a surge in torque and FF then that would be great news. Perhaps there is a solution...

Cheers,

John

[metzgergva 210]

From my experience designing such aircrafts (ATR72, Caravan, Cheyenne), I would like to make the following comment, hoping they contribute to the technical understanding and the limitations in FS:

The turboprop model is a very simple model in FS (FS9 or FSX) and does not at all represent very well a PT6 engine. As an example there is, that on the default model the TQ does not increase when you reduce prop RPM.

Furthermore FS engine design is not about modelling N2 or N1 and then getting an engine performance. Those values mean nothing for the thrust. FS is simple, you burn fuel at an engine specific efficiency which gives you a thrust on a jet and a TQ on a turboprop. All values on N1, N2, ITT are just added for visual impressions. Sorry if that ruins your perseptions.

As pointed out, the CL has a range and basically increases fuel flow in idle, which in consequence gives a higher turbine RPM (N2) and thus a higer Prop RPM (n1) and therefore more thrust.

The "fuel flow gain" is a control that controls changes on fuel flow. The lower the value the lower the speed of change. It is a constant in the turboprop model and it has a relation to N2 in jets which would be nice to have but we don't have available in turboprops. This control and it's setting has multiple effects.

During flight it regulates the speed of change on fuel flow depending on PL changes. Unfortunately a simple control so when you make it fast you get fast reaction on fuel flow, thus on N2 and TQ. But it overshoots the target value on TQ because it controls FF and not TQ. So it is a compromise between reaction time and overshooting TQ.

Secondly, if people push the PL full forward it opens the fuel flow valve fully and the TQ limiter then reduces it. Now when you have set the control to low reaction (=0.03) it takes long until the control drops down to the value that the TQ limiter already has, thus no reaction.

Thirdly, it affects start up.

A proper start procedure is to move the CL, thus opening the fuel valve, when the N2 is running at a minimum speed. Now if you have the control set to low it opens slowly and by that the raise in N1, N2 and TQ is slow.

If you set the control to fast (=0.08) it opens faster and the effect of overshooting can be observed.

If you put the CL into ON (either low or high idle which just makes a small difference in the max fluel flow at idle) prior to the N2 min is reached, FS puts the valve full open and the N2, N1 and TQ go up like crazy.

So my personal conclusion to that is ,that it is a big compromise and I prefer to have a proper start behaviour and a compromise on reaction time, though I know in reality it is much faster.

The slower rection I can plan in advance and you see on many videos that these control levers are all moved slowly.

I know that in consequence also the reduction on TQ/RPM during flare is too long and I anticipate that by pulling back the PL a bit earlier than in real life.

That is why I recommend values of 0.02 to 0.03 and I will set back the Cheyenne to those with the next update. Same for my Caravan modifications.

On the icing side I'm still learning what FS actually does. There are many different system in place and a drop on flying performance should only arrive from using bleed air either for having hot air circulation or building pressure in a duct. Any electrical device should only have an impact on the amps drawn but not have an effect on engine performance.

Lastly a comment on idle run. As far as I know all turboprops have enough thrust to start rolling and taxi quite fast with engines in idle. That is why some have the CL to use for adjusting taxi speed. This is not coming from the exhaust thrust, it comes from the idle RPM and the idle blade angle.

FS again is simple and has defined positions while on ground which are not regulated and not transitioned smoothly. They are revers angle, idle on ground, min governed. So for having it rolling in idle you can set the ground idle to a higher value like 5 or 10. That makes them roll. If you aply power it jumps to min governed which is typically 15 or more. So if the first on is set to small values like 1, this is why they start rolling immediatly with just a litlle PL move. Now the drawback of using a proper ground idle is that when you want to get out of reverse it does not do that slowly - it jumps from reverse angle to ground idle with the turbine still running high which then makes your plane jump forward. You see all compromises and personal developers taste on what seems to be more real, or close to or can be lived with.

[JAL 0]

Hi Alexander,

Nice post, and really interesting to hear your insights.

About the taxi behaviour on the real life Twin Otter. As I understand it it does require power inputs to get it moving, but once it is moving the power lever can be put back in idle. At this point it may need some beta inputs every once in a while to keep the taxi speed down. Thus it appears that the Otter does require a bit more coaxing to get going than some other aircraft like the King Air. Perhaps this is why the Twin Otter only needs an on/off condition lever, whereas the King air needs low idle to slow things down a bit.

That trade-off between the prop pitch on ground and proper reverse behaviour is annoying eh? I've been trying to find some ways to get around it and I seem to be making some progress. Hopefully, I'll get it to work.

Exhaust thrust in real life does affect taxi behaviour in addition to prop pitch and rpm (also true in the sim). In turboprops, as I'm sure you know, engine output is measured in shaft horsepower (SHP = engine power available to propeller) and equivalent shaft horsepower (ESHP= which is basically exhaust thrust plus SHP). For example, in the twin otter when the engines have no forward motion and they are run up to max values SHP = 620 and ESHP = 650. Thus exhaust gases add 30 horse power at this point (FS does a good job of modelling this when the correct values are given to it).

The situation is more complicated when the aircraft moves through the air. I've heard it explained a couple ways. When a turboprop aircraft moves forward faster than the speed of the exhaust gases from the turbine ESHP begins to decline (apparently the exhaust begins to function as a vacuum reducing the total horsepower output). The second explanation I've heard is that ram air drag from the turbine inlet area becomes greater than the exhaust gas output. FS also appears to have modelled this effect reasonably well judging by the output in AFSD.

I'm on the fence right now with the correct start-up vs. engine reaction time trade-off. I can see how the surge at start is annoying, but then again many would prefer to have the aircraft fly properly, especially in the flare. Oh well, maybe we can lobby Microsoft to get this fixed eh?

Cheers!

John

[Stefan Hoffmann 4160]

I have the problem with other planes too. This looks like an FSX internal problem. The sim or other joystick related software seems to forget were the throttle lever stays. When I understood the system right, the analog stick uses too a reference center, defined during calibration process. When this is lost somehow, the joystick raw data seems to be lost in the digital jungle.

Moving the throttle axis on the stick seems to reset the stick/relativ control software out of its sleep. But I have this only when FSX is my first game played after booting the system. After this all works fine. On the other side only FSX makes this problem. Had no simulator before where I had this fun.

[metzgergva 210]

@ JAL

I'm with you on the taxiing. Could well be that the Twin Otter needs a bit power to start rolling. From what I now of the ATR it is also very much load depending. Heavy very rare to start rolling and empty you are on the brakes all the time, even a litlle rev is then used to keep the right taxi speed. I was taxiing the Cheyenne for half an hour to find out and she is very fast, immediately starts rolling as soon as you release the brakes. That's were input from RW pilots is essential

In fact still is and always was that the roll resitance in FS is too much. But if you play with ground idle blade angle you will surely find the right setting.

Drop me a PM if you want to go into any detail on flight dynamics.