737 trim

[ivar hestnes]

I have an issue with my trim-function. When on autopilot, the trim is correcting every movement due to wind or whatsoever. (software and then hardware also of course). And it always trim too much, so it has to correct back again. The trimwheels almost never stop to run. Forward and back, all the time. Sometimes I have to go around because of this. Because when I use approach it trim so much that I actually miss the glideslope;)

If it was possible to add a delay or something to the trim it would be nice.

Anyone else has issues with this? Or improvements to share:)

[Michael Carter]

The 727 only does this in rough conditions. Normally, you'll only hear the trim when setting flaps or a power change, or both.

I noticed it trimming quite a bit in your video. Man, that chain drive gets on my nerves in rough weather.

[NicD]

Yes - we have the same issues. At the moment we have disabled trim wheel movement while the autopilot is connected. In the longer term we'll be programming SIOC to handle this. A few ideas here:

1) use a delay between writes so that the trim isn’t running constantly
so logic could be – once a trim command has occurred to the trim then don’t do another command until X seconds has elapsed

2) There also needs to be trend logic to eliminate constant small corrections. So don’t do a trim command if the commanded change is less than X% of trim.

[ivar hestnes]

Yes Michael, the trimwheel chain is having a really hard life;). At least I have a switch to cut off the motor, but the relays still make noise though.

NicD: You definitely are spot on with your suggestions here. But I dont know how to solve it. Thats for sure. Could there be a way to configure this in the flight-model?

I tried FSX with default 737 a while ago, and if I remember correct, it was much smoother there.

My first guess is that maybe this is flight-model related. I use pmdg as the flight model in my setup. But I cant remember the trim acted like this when desktop flying earlier. Is it possible that it is the combination of pmdg and PM that cause this? hmmmm. Lots of questions;)

Anyone using an other 737 flight-model? Experiences?

:)

[NicD]

Yes that's right Ivar - from my testing it's the PM AFDS that is overactive... the A/T is just too twitchy, and the movements a bit coarse. Maybe we should talk to PM about smoothing the A/T logic?

[mauriceb]

I am using Phidgets rate of change function to control my trim motors with my logic board. With that function, you can adjust how sensitive the trim wheel motion is and I don't really have a problem with that setup.
Yes, the A/P does try to trim a bit too often but it settles down quickly and it is not an issue for me.
I tried without using the rate of change function and it was a disaster since the trim wheel was moving almost constantly.

Anyway, here is my circuit and some notes I made. I used my own logic board, but you can do the same now with a Phidgets 2 relays board where the relays can be controlled by an offset. I'm sure there must be other ways to make trim function properly, but this works for me ;)

Maurice


http://thumb0.webshots.net/t/24/565/3/16/26/2923316260051510320IjlOKk_th.jpg (http://home-and-garden.webshots.com/photo/2923316260051510320IjlOKk)

http://thumb0.webshots.net/t/60/660/6/86/33/2307686330051510320BKrkPi_th.jpg (http://home-and-garden.webshots.com/photo/2307686330051510320BKrkPi)

[ivar hestnes]

Yes, we should definitely talk to PM about this issue. Maybe they can make a solution for it. I will send a mail later today. Maybe you can do the same NicD:)

Thanks for explaining what you did Maurice. My trim is setup with phidgets, so I will check the "phidgets/FS rate of change" function. Thanks.

:)

[ivar hestnes]

Hi there.

I tried to "uncheck" the use FS rate of change box. But it didnt actually help. The trimwheel now only go in one direction. But less movement on the wheel though. Still indicator moves alot, like before.

I use same offset as per your drawing Maurice.

I sent an e-mail to PM-support and a link to this thread.

:)

[mauriceb]

Hi there.

I tried to "uncheck" the use FS rate of change box. But it didnt actually help. The trimwheel now only go in one direction. But less movement on the wheel though. Still indicator moves alot, like before.

I use same offset as per your drawing Maurice.

I sent an e-mail to PM-support and a link to this thread.

:)


What do you mean by 'uncheck? That setting should be checked in both inputs. I will send you screenshots of mt settings as soon as I can.

Maurice

[mauriceb]

Here are my Phidgets settings. If you decide to try that, you will have to play with the sliders to get the right trim wheel movements (adjust sensitivity). Of course, it would be a lot easier if PM fixed their PM trim offsets :D

Maurice

http://i32.photobucket.com/albums/d38/mau_ben/TrimOffsetsettings.jpg

http://i32.photobucket.com/albums/d38/mau_ben/DigitalOutput7.jpg

http://i32.photobucket.com/albums/d38/mau_ben/DigitalOutput4.jpg

[JonathanRichardson]

Hi

We received a support e-mail asking to react to this thread. . .

The background information is that the MCP software is reacting to the instability of the flight model. It will try and keep the a/c in trim and this requires many changes small and large to achieve this. This I’m afraid is inherent in most MSFS flight models. Normally you don’t notice until you connect large motors and trim wheels and see / hear every single change of trim. But it is going on all the time when the AP is engaged, just use the CDU to monitor the raw trim values and you will see it changing, if it were not, then the AP would not be able to fly the flight model precisely. The flight model instability can be effected/improved or made worse by various parameters in the flight model, and also the way the a/c is loaded. But essentially it is inherent and there, and I doubt there is a way to completely get rid of it. To get around this as others have suggested, you need to add a section so that your trim motor/wheel will only trigger after a given amount of MSFS trim change. This is how we do it in the simulator here.

I have a feeling that if we were to change the MCP code and took out the fine trim control, you would get the a/c constantly drifting from level flight, undulating, and in turns or level changes capturing altitude or maintaining altitudes would be not be possible. Also, what might work for one very good flight model would not work for others and so forth. We looked at this a couple of years ago in detail and at that time came to the conclusions I outline here.

Regards
Jonathan Richardson
Project Magenta

[ivar hestnes]

Here are my Phidgets settings. If you decide to try that, you will have to play with the sliders to get the right trim wheel movements (adjust sensitivity). Of course, it would be a lot easier if PM fixed their PM trim offsets :D

Maurice

http://i32.photobucket.com/albums/d38/mau_ben/TrimOffsetsettings.jpg

http://i32.photobucket.com/albums/d38/mau_ben/DigitalOutput7.jpg

http://i32.photobucket.com/albums/d38/mau_ben/DigitalOutput4.jpg

Thanks for the pictures Maurice:) Actually my boxes was checked all the time. Tried first yesterday to uncheck. They are now checked again. Will give it a try to move those sliders.

But one thing: I have circuit 1 on both digital outputs. And you use circuit 0. What would be the difference here? Hmmmm, I used Thomas Richters 737 program to find the offsets and circuits, but could still got it wrong LOL.



@ Jonathan: Thank you very much for jumping in with your explanation:)

[mauriceb]

.

But one thing: I have circuit 1 on both digital outputs. And you use circuit 0. What would be the difference here? Hmmmm, I used Thomas Richters 737 program to find the offsets and circuits, but could still got it wrong LOL.




I'm pretty sure the circuit number makes no difference. As far as offsets are concerned, all I know is that 0BC0 works for me :). It's a bit tricky to get the sliders at the right values and it's just a matter of trial & error...no 'scientific' way to do that as far as I know.

Also, I have selected Invert for the signal but you may not have to (I needed to do that to drive the transistors that operate the relays)

Maurice

[mauriceb]

One more thing about that. For the calibration, I did not use the A/P. I simply used the trim switches. The idea is that when you press the trim switch up or down, there is a delay before the trim motors operate. If you get the right delay, then the A/P will also not actuate the trim motors as frequently but it will still 'hunt' a bit back & forth when you do major flight level changes although it settles down quite quickly with the right delays.

Maurice