Strange PFC Yoke Aileron Behavior

[MichaelMcE]

Pete,

I hope you can help me troubleshoot this strange behavior ...

I have the PFC Jet Yoke, Rudder Pedals and use the 2 engine jet throttle quad. The PFC h/w is connected to my FS host via a USB-to-serial interface. I'm using PFC.dll v1.989.

I've been flying this setup for 2+ years with 767PIC and LDS 767 and have experienced no issues.

I recently purchased the PMDG 747. After the install I configured the pfc driver to use the 2 engine throttle quad to control the 4 engines. I enabled a user config, assigned the left throttle as throttle 1+2 and the right throttle as throttle 3+4. I also enabled the spoiler and thrust reverse levers. I closed down the pfc configuration utility and double checked my assignments by observing throttle, reverser and spoiler movement. All looks good.

I then get the aircraft airborne and initiate an easy right turn. The a/c violently rolled left, and then began to roll right. This behavior repeated everytime I would initiate a right roll. So, I ended the flight, restarted, and double-checked all my settings. When observing the aileron movement on the Flight Controls page, I noticed that as I slowly and continously input right aileron, the "dot" on the response curve would drop full negative and then return to the plotted curve. When inputing left aileron, all is fine. All is fine for up/down elevator, left/right rudder, brakes, etc. ONLY with right aileron to I get this sudden full down (negative) deflection. And it occurs at the same point in the deflection.

I decided to start over from scratch. First, I loaded the LDS 767 using the 2 engine throttle quadrant config. I repeated the test of control movements that I described above. Everything worked as it should. I ended that flight and started the PMDG 747, still using the 2 engine throttle quadrant config. Repeated the test again, everything worked just fine.

I then selected a user config, but did not make any assignments. I checked control throws and everything was fine. I then assigned throttle 1 to engine 1+2 and throttle 2 to engine 3+4. When I checked control throws, the full negative deflection behavior reappeared.

I'm not sure where to go from here. I have started from a clean FS9 start, using clean (archived) versions of pfc.ini, and regardless of user config I use to assign the controls, I get the same behavior. Do you have any ideas on what might be going on; is there some data that I can log to help explain what might be going on????

I'm stumped????!!!!!

Thanks.

-michael

[Pete Dowson]

When observing the aileron movement on the Flight Controls page, I noticed that as I slowly and continously input right aileron, the "dot" on the response curve would drop full negative and then return to the plotted curve. When inputing left aileron, all is fine. All is fine for up/down elevator, left/right rudder, brakes, etc. ONLY with right aileron to I get this sudden full down (negative) deflection. And it occurs at the same point in the deflection.

Hmmm. I'm not at all sure how to interpret the information you've provided. Are you actually saying it ONLY occurs with the PMDG 747?

The Flight Controls page in the PFC driver Options should really only be reflecting what is being input from the hardware, so I would first suspect the pot on the yoke -- a dirty section perhaps. But then you'd expect to be able to see this no matter what aircraft you are flying. Certainly, whilst in the PFC options the current aircraft loaded in FS shouldn't matter at all -- unless for some reason its activity is interfering with the PFC link to your PC, but I really can't see how that could be. Even if it were, you'd expect that to play rather random havoc, not a specific type of error.

I then selected a user config, but did not make any assignments. I checked control throws and everything was fine. I then assigned throttle 1 to engine 1+2 and throttle 2 to engine 3+4. When I checked control throws, the full negative deflection behavior reappeared.

So you are possibly deducing that it is related to the mapping of the throttles, independent of the aircraft itself? Can you verify that with, for instance, the default 747? Is it only since you purchased the PMDG 747 that you have been mapping the throttles linke this?

I'm not set up at present to conduct experiments on this here, but I will within the next few days. Meanwhile, any further information you may find will be useful.

Oh, could you try the current "official issue" PFC.DLL (from http://www.schiratti.com/dowson) just in case it is some problem introduced over the last year or so. Thanks.

Regards

Pete

[MichaelMcE]

So you are possibly deducing that it is related to the mapping of the throttles, independent of the aircraft itself? Can you verify that with, for instance, the default 747? Is it only since you purchased the PMDG 747 that you have been mapping the throttles linke this?

Oh, could you try the current "official issue" PFC.DLL (from http://www.schiratti.com/dowson) just in case it is some problem introduced over the last year or so. Thanks.

Regards

Pete

Pete,

I verified that the behavior exists independent of aircraft. I just ran a series of tests with pfc.dll v1.989 and v1.92, using the default 777, default 747, and the LDS 767. The behavior existed in all tests.

And yes, prior to acquiring the 747 I never tried mapping throttles in this manner. I figured I would give it a try to see how I like it before investing in the 4-engine throttle quad.

Thanks for the help. Let me know if I can help further at this time.

-michael

[Pete Dowson]

I verified that the behavior exists independent of aircraft. I just ran a series of tests with pfc.dll v1.989 and v1.92, using the default 777, default 747, and the LDS 767. The behavior existed in all tests.

... Let me know if I can help further at this time.

Could you Zip up and send me the PFC.INI file from the FS Modules folder, please? (petedowson@btconnect.com). I'll try checking this out tomorrow -- if I can't reproduce it here I may ask you to do some tests for me with different options.

Seems odd that no one has reported it before if it is in the old version too, so maybe it is related to specific options, slopes, etc.

Thanks,

Pete

[Frank.O]

Hi,

I've had something similar with my Cirrus II. Only the dot did not go fully left but only a small portion. If I used a certain amount of deflection I could actually make the plane bank left by turning the yoke to the right.

This happened for elevator and rudder also (but the rudders are very old and I know they have a faulty pot).

But I have never done anything special with assignments. I only ever fly the Baron and am using the twin prop quadrant exclusively.

My Cirrus II did have some serious faults and after I had it sent for repairs the strange behaviour was gone. At some point it returned and I repaired the console myself and it was gone again. I haven't been flying for quite a long time now so I can't tell exactly how it developed but I allways suspected the hardware to be responsible. I seem to recall that now and then I had to deal with it by making the dead zones in a little wider.

Regards,

Frank

[Pete Dowson]

My Cirrus II did have some serious faults and after I had it sent for repairs the strange behaviour was gone. At some point it returned and I repaired the console myself and it was gone again.

Yes, from Michael's description it does sound as if it MUST be hardware in origin, as I think the calibration displays are showing the raw inputbut maybe the "slope" tables come into it. I need to check. What is so odd in Michael's case isn't so much the wrong deflection (which can, I think, be a little "dead" spot on the pot or something else electrical in the controller) but that, by his testing, it only happens when he has mapped the throttles!

I need just to be sure there's no corruption occurring internally to the driver. Once I have checked that here and found no sign of it, then I will have to ask Michael to turn on some Comms logging to check the raw data arriving from the PFC controller.

May I ask, in your own console repair, what did you actually do? Was it a change or clean of potentiometer? Or something electronic?

Regards,

Pete

[Pete Dowson]

Let me know if I can help further at this time.

Well, I've tried all sorts of things and cannot make anything go wrong here. I am using a slightly later version of PFC.DLL, but the changes are only in relation to PMsystems and overgead switches in the PFC 737NG.

I definitely need your PFC.INI so I can check with the same options and selections. Tell me also:

1) which user config in your setup you are selecting

2) using Throttles R12, R34 or just forward thrust 12, 34?

3) Flitering on or off?

4) Slope selected or flat?

You say that the little dot on the Slope depiction zooms off the wrong side, but that is manipulated by the driver. Please watch the "input" value for the aileron, the first figure under the little "RVS" checkbox. Does that undergo similar sudden changes? If so, then it is definitely something on the hardware side, as that is reflecting the raw input.

Let me know these things, then we'll try logging to get hard data.

I attach PFC DLL 1.992, just to be sure we are using the same version. Please do the tests with this from now on. (Note that the Version Info on this build says 1.991 -- there's no difference, both were internal builds on the same day).

Regards,

Pete

pfcdll1992.zip

[Frank.O]

May I ask, in your own console repair, what did you actually do?

There were several things. When I first sent it in they said that it was a broken wire on the aileron pot. The second time they couldn't find anything and replaced the controller board. When I got it back it was still faulty so I opened it and found a bad protection diode in the power line for the avionics, a bad avionics master switch (both together made the avionics flash on and off wildley and display silly things), a bad soldering spot on the controller board and I think I resoldered some of the wires too.

Regards,

Frank

[Pete Dowson]

When I first sent it in they said that it was a broken wire on the aileron pot.

That's the sort of thing which would give the results Michael is getting. On my Jetliner yoke the left side of the aileron is playing up too -- jitters up and down a bit near the centre of the left side movement. I have proven it is definitely the hardware input -- I think it needs a new pot, or at least a clean in there somehow. I don't currently use it for real flying only testing, so I've not asked PFC about it yet. The one in the 737NG is fine.

The second time they couldn't find anything and replaced the controller board. When I got it back it was still faulty so I opened it and found a bad protection diode in the power line for the avionics, a bad avionics master switch (both together made the avionics flash on and off wildley and display silly things), a bad soldering spot on the controller board and I think I resoldered some of the wires too.

Phew! That doesn't sound good. Their build quality is usually very good indeed -- you must have been very unlucky!

Were any of those other things responsible for jitter in the aileron, though? Surely that was all down to the pot?

Regards,

Pete

[Frank.O]

Were any of those other things responsible for jitter in the aileron, though?

No, I don't think so. The jitter was surely caused by the pot (mainly) and the wires. May be the cold spot on the controller board had some influence. It was on one of the pins of the connector where the wires from the switches and pots come in so it's possible but I can't say for sure.

A clear indicator that this was hardware related was the fact that sometimes a decent tap on the side of the console could "fix" it (the flashing of the avionics and/or the jitter).

Regards,

Frank

[MichaelMcE]

Could you Zip up and send me the PFC.INI file from the FS Modules folder, please? (petedowson@btconnect.com). I'll try checking this out tomorrow -- if I can't reproduce it here I may ask you to do some tests for me with different options.

Thanks,

Pete

Will do so when I get home tonight.

-michael

[MichaelMcE]

I definitely need your PFC.INI so I can check with the same options and selections. Tell me also:

1) which user config in your setup you are selecting

2) using Throttles R12, R34 or just forward thrust 12, 34?

3) Flitering on or off?

4) Slope selected or flat?

Pete

1. I tried it with several user configs: 1, 2, 3, 15.

2. Just forward thrust 12 and 34.

3. Filtering is on.

4. Slope is selected - about 2 clicks.

You say that the little dot on the Slope depiction zooms off the wrong side, but that is manipulated by the driver. Please watch the "input" value for the aileron, the first figure under the little "RVS" checkbox. Does that undergo similar sudden changes? If so, then it is definitely something on the hardware side, as that is reflecting the raw input.

Let me know these things, then we'll try logging to get hard data.

Pete

Ah yes. Good point. I will check the actual values to see what is happening with the raw input.

I will also try it with zero slope.

BTW: I initially suspected h/w (and have not ruled it out :-) ). I had to return the yoke to PFC 3 months after I purchased it to have them address an intermittent problem with the left/left rocker randomly triggering input on the other rocker switches. Since I use that rocker for trim, I was getting all kinds of random actions occurring during a flight when manually flying - very frustrating to say the least. I used the "test" page to document what was happening (thanks Pete!). Since then, the h/w has performed flawlessly.

I will make a visual check of the board in the TQ Console and also pull the board at the base of the yoke - PFC had me do that when I first reported the left/left rocker issue.

I attach PFC DLL 1.992, just to be sure we are using the same version. Please do the tests with this from now on. (Note that the Version Info on this build says 1.991 -- there's no difference, both were internal builds on the same day).

Pete

Very good. It will be a long day for me, I wont get to test this until about 11PM tonight. Look for it tomorrow.

Thanks.

-michael

[Pete Dowson]

I will also try it with zero slope.

And no filtering, just in case, please.

Regards,

Pete

[MichaelMcE]

I will also try it with zero slope.

And no filtering, just in case, please.

Regards,

Pete

Will do.

-michael

[MichaelMcE]

Please watch the "input" value for the aileron, the first figure under the little "RVS" checkbox. Does that undergo similar sudden changes? If so, then it is definitely something on the hardware side, as that is reflecting the raw input.

Pete

Pete,

We can eliminate hardware as an issue. I checked the raw input value as I deflected the yoke. As the raw input value progressed from 85 to 97, the 'scaled output' went full negative. At one point I was able to hold the yoke steady with a raw input value of 87, and the scaled output remained locked on full negative.

But, then the really interesting part ...

The above test was made with a response curve. When I set the response to linear, I could not reproduce the problem. I then made a single click on the response curve and was able to reproduce the problem. I then made a second click on the response curve and was not able to reproduce the problem. After a series of tests on the response curve, I can only reproduce the problem if I make a single click to the response curve from the linear setting.

The tests were run with pfc.dll v1.992.

-michael

[Pete Dowson]

I checked the raw input value as I deflected the yoke. As the raw input value progressed from 85 to 97, the 'scaled output' went full negative. At one point I was able to hold the yoke steady with a raw input value of 87, and the scaled output remained locked on full negative.

... When I set the response to linear, I could not reproduce the problem. I then made a single click on the response curve and was able to reproduce the problem. I then made a second click on the response curve and was not able to reproduce the problem. After a series of tests on the response curve, I can only reproduce the problem if I make a single click to the response curve from the linear setting.

Okay -- that's good information. It indicates a corruption in the table for the one curve. With your INI (now received) I should be able to locate that and fix it.

For now, use one of the response curves that works! ;-)

Regards,

Pete

[Pete Dowson]

Further on this:

I checked the raw input value as I deflected the yoke. As the raw input value progressed from 85 to 97, the 'scaled output' went full negative. At one point I was able to hold the yoke steady with a raw input value of 87, and the scaled output remained locked on full negative.

Hurrah! I can reproduce this. It appears to be irrelevant whether throttles are mapped -- in fact the phenomenon is completely independent of the quadrant assignments, once it occurs.

It is obviously a corruption in the 1st Slope table (after the default linear setting, that is), as it affects that #1 slope selection identically on any axis to which it is applied, not just aileron!

The interesting thing is that it is not due to the action of selecting a quadrant setting with mapped throttles, nor even using it subsequently. The problem only arises if such a quadrant is active when FS is first loaded. So it is something to do with initialisation of the slope tables.

RightI have a lot of clues, now to dive into the code! ;-)

Regards,

Pete

[Pete Dowson]

Hi Michael,

Found it! A very old error for sure. Please try the attached PFC DLL version 1.993. If this is okay for you too I will put it up with the Interim releases at the top of the Forum.

Thanks for your help,

Pete

PFCDLL1993.zip

[MichaelMcE]

Hi Michael,

Found it!

Pete

Good man!

A very old error for sure. Please try the attached PFC DLL version 1.993. If this is okay for you too I will put it up with the Interim releases at the top of the Forum.

Thanks for your help

Pete

Thank you. I know how these strange issues can eat at you and it's impossible to rest until you solve them! ;-)

I will test it tonight and get back to you tomorrow.

-michael

[MichaelMcE]

Hi Michael,

Found it! A very old error for sure. Please try the attached PFC DLL version 1.993. If this is okay for you too I will put it up with the Interim releases at the top of the Forum.

Thanks for your help,

Pete

Pete,

Glad to report that I can not reproduce the error with 1.993. Thanks for the quick turnaround.

-michael

[Pete Dowson]

Glad to report that I can not reproduce the error with 1.993. Thanks for the quick turnaround.

Thanks. I've released it now in the "Interim Versions" announcement above.

Regards,

Pete