737 TQ Trim Wheel Motors and Control Questions

[MistyBlue]

For those of you who have motorized your 737 TQ trim wheels, what components worked for you? 

I rebuilt my RevSim TQ, but ended up stripping teh internal gears in one of the Modelcraft motors.  Likely because the instant stop/reversal of direction of the trim wheels was too much to take.  I have a feeling I may have burned out the motors too, or maybe the Phidgets servo card.  Or both.

What motors are you using?  And how are you handling sudden trim wheel stops and reversals?  A clutch of some sort?

I was thinking of putting something like a polyclutch into the drivetrain somewhere so an instant reverse of the wheels isn't so hard on the components.

I'm also wondering if there are better motors to use...something with a much stronger internal gearing.

TIA

[FredK]

Actually....finding a better way to do the trim wheel motion is going to be my major winter project this year. However, I really have not given it any thought as yet.

Right now I have a standard DC robotic motor that drives the chain drive. The issue is that it does not spin the wheels as fast as they should....and as you have described the wheels do not stop instantaneously and the sudden reversals place a strain on the setup.  The setup has lasted for 7 years now but my gear drive is slightly out of alignment. So I want to at least fix that which means that I have to remove the entire TQ assembly which is a major project.  So if I have to go to all that effort I might as well do it right.

My interface is an OpenCockpits DCMotorsPlus card with a custom SIOC script I wrote.  The interface has performed well.

I will be out at Mike Sherick's at the end of January. I am very curious to experience how his operates.  I know his trim wheel setup is more eloquently engineered than mine regarding the drive setup.

I am eager to see what this thread produces.

Fred K

[MistyBlue]

Hey Fred, looking forward to what you find out.

I've purposefully stayed away from SIOC for a number of reasons and went with the Phidgets DC motor card instead. I think it'll suffice since I don't need multiple trim wheel speeds on this TQ.  I'm planning on replacing the TQ with something better next year.  I just want to get simple forward and reverse working.

My RevSim setup has two trim wheel motors geared together (see pic) which should be enough to handle quick starts and stops if I keep the trim wheel speed down.  I think I just had it revving up too high and the strain on the components during reverses was probably too much.

I saw in an earlier post that MickC had gone with Pololu 1:10 motors (https://www.cockpitbuilders.com/index.php?topic=7894.msg56772#msg56772).  I may give those a try.  I think the helical gear coming off the motor might be a bit stronger than the spur gear I previously had which broke.

RevSim originally had a solenoid with a plastic "bumper" that would pop out and rub against one of the trim wheels to help stop it from spinning more quikcly.  I removed that during my rebuild because I had no way to provide logic to tell it when to deploy.  That's probably where SIOC would have been handy.

I'm thinking if I keep the speed down I can probably get by as is without damaging anything.  I'm just wondering about the Phidgets card.  When you reverse direction on the trim wheels, the momentum keeps them going and as I understand turns the motors into generators.  I "think" the card can absorb that via capacitors (dynamic braking?), I'm not sure.

Something seemingly simple apparently isn't as simple as I thought.  Sounds familiar, haha.

[MistyBlue]

I talked with the Phidgets folks about my motor braking concerns.  They recommended their PowerGuard in-between the motors and the power supply.  It's meant to prevent back-voltage during motor breaking from making it back to the components including the power supply.

He said yes, the motors will burn out over time with sudden changes in direction, more quickly if heavier loads are applied but that's normal.  I figure if I keep trim wheel speeds down it'll save on wear and tear and they'll last longer.

I'm going to try their newer DC motor control with the PowerGuard and and see if I can get these wheels a'turning again.

[sagrada737]

Trying to emulate a real 737 Trim Wheel operation represents a big challenge when converting a real 737 TQ for our Sims.

The DC motor "burn out" problem is not a surprise, if the control electronics are not designed for this function of "instant reversing" of the Trim Wheels.  Back EMF can be difficult to control.

Perhaps a Servo Controller/Servo Motor would be a good solution to accomplish the instant reverse and speeds of the Trim Wheel.  That said, keep in mind that there is MASS involved in this type of mechanical system, and although the Trim Wheels are not that heavy, there is quite a bit of inertia to overcome when starting/stopping/reversing of the Trim Wheel system at its high rotational speeds  -- further complicating the mechanical systems to be able to hold up under such activity.  No wonder the 737 TQ is built like a tank!

What Fred mentioned on our TQ conversions has more to do with the sync of the Trim Wheel activity, as a function of how the SIOC code is currently written.  In Fred's and my case, we are using a small, low RPM 12 VDC motor, which in fact works fine to "give a hint" of Trim Wheel motion.  Presently, in our case, the Trim Wheels don't always start/stop/run perfectly in sync with the 737 flight characteristics.  However, it does give a nice sense of Trim Wheel activity, which adds to the overall cockpit Sim immersion.

If I had to redo the Trim Wheel drive system in my Sim to accurately emulate the 737 Trim Wheel function, I would be tempted to use a Servo Drive system.  But the system integration challenges of such a system are too much for me to tackle for my Sim.  I'm quite happy with my current simple setup using the 12 VDC gear Motor/sprocket, which is dirt simple.  Thanks to Fred's SIOC code, it does a good job of moving the Trim Wheels.

Looking forward to seeing you in January Fred...

Mike

[MistyBlue]

Thanks Mike.

It turns out the Phidgets PowerGuard is a no-go for back-EMF.  Needs to always be connected to a VINT hub and manually enabled through the Phidgets control panel every time the app is opened (no "remember previous state" option).  Since ProSim doesn't well support VINT hubs and certainly not the PowerGuard, that won't work.

Fortunately it appears my Phidgets DC motor card (1604) is still working fine.  Phidgets says it has a diode built in which should be enough to protect against back-EMF on these smaller motors.  I guess I could add a simple diode across the motor leads as added protection.  I just have no idea what sized diode I need.  If anyone knows, I'd appreciate the advice.

[Mike.Powell]

It may not be so simple as a single diode across the motor. A bidirectional DC motor controller can/will put reverse voltage on the motor to reverse its rotation. If there's a diode across the motor, it will short out the motor power on one of the rotational directions. If you need to clamp back EMF you can use four diodes (two per motor lead) that restrict the voltage swings of each motor lead to between ground and the power supply voltage, but I doubt that is needed. I suspect that the Phidgets 1064 uses H-bridge drivers and they often include such diodes.

I admit that I'm shooting from the hip here, but I question if back EMF is actually the root cause of your issue. It seems more likely that back EMF would kill the motor driver itself rather than the motor. Is  this a brushed motor that is possibly getting over current damage during low RPM while reversing direction?

[MistyBlue]

Thanks Mike.  Yeah, I don't think back EMF is an issue.  Seems like the Phidgets card is just fine.  And the motors seem to be as well, aside from a stripped gear in the one gearbox.

I just tried the 10:1 motors from Pololu that I remember Mick mentioned but they just don't seem to have the torque I need to get the wheels moving and stopped fast enough.  I'm going to give the 1:100 gearing a try.  Lots of torque and slow enough speed so it shouldn't break anything.

I'll update here when I figure out my solution, for what it's worth.

[MistyBlue]

I've been doing some more testing.

The original Revsim design has two geared motors driving a center sprocket and chain to the trim wheels as in the Gearing.jpg graphic attached.  I noticed, however, that both geared motors didn't turn at the same speed which may have contributed to some of the problems I was having.

So, I decided to remove one of the motors and drive the trim wheels with just one Modelcraft 1:200 geared motor as in Gearing2.jpg.  That didn't last long.  A spur gear inside the motor broke rather quickly.  These Modelcraft motors only handle up to 18 Kg/cm torque, which apparently isn't enough. 

The 1:100 Pololu's I'm waiting on are 34 Kg/cm so I'm hoping that will work better, but I still think it may not be enough.  Pololu has an even higher 49 Kg/cm torque motor so that may be my final resort.  Of course the higher the torque, the lower the RPM, but even their lowest RPM motor is still adequate in this configuration.

Regardless, I think the Revsim design isn't very good because you have large gears driving small gears instead of the other way around.  I think it puts too much load on the motors.

I may reverse the design and try putting a small pinion gear on the motor driving a bigger main gear connected to the chain.  Means the motor will have to turn faster, but maybe the loads on the motor will be less?

Or, I may have to find a powerful motor that isn't geared.  I just don't know if I can find any with the torque I need.

A last resort may be putting some sort of poly clutch somewhere in the drive train to help absorb sudden changes in direction.

[sagrada737]

Hi Tony,

Great job on your Sim web site.  Well done!  Interesting background and project status...

Regarding the Trim Wheel gearing...   The two motor configuration seems a overly complicated.  I should ask what kind of Trim Wheel specification are you aiming for, in terms of speed, and start/stop capability.  This in turn really defines the hardware and software specs.  It would be better to explain your design goals in order to better provide comments.

Mike

[FredK]

Keep in mind that we are comparing apples to oranges here as well. In the real OEM case the sprocket gear for the trim wheels (4-5 inch diameter or so) is typically a much larger diameter than for the drive motor sprocket. This is the opposite of what is depicted in the drawings. The torque requirements etc. for each case will differ markedly relative to the gearing arrangement. Engineering a rapid spin effect for the real TQ becomes a bit more challenging in that regard given the cramped working space.

Fred K

[MistyBlue]

Great job on your Sim web site.  Well done!  Interesting background and project status...

Thanks Mike!  I sincerely appreciate it.  I'm humbled as you are way better than this than I.

Fred, I didn't realize the sprocket gear was that large.  Thanks for that info. So yeah, apples to oranges as you say.

Mike, for requirements here's what I'm thinking.  Just a single speed.  Top RPM somewhere in the 85 range which, if I recall, is around where Flaps Up A/P on trim speed is.  Stop start doesn't have to be instantaneous, but fairly quickly (< 1.5 sec I would think).

FYI, I got a very helpful response from Patrick at Pololu yesterday eve regarding motors.  He said all their 37D motors have the same gearbox and that max instantaneous torque is 25 Kg-cm (350 oz-in) with constant max continuous load of 10 Kg-cm (150 oz-in).  He suggested planetary geared motors would be the strongest and was kind enough to point me to other places that sell them.  However, nothing may be good enough if I can't figure out how much torque is applied to the motor when those wheels are stopped/reversed quickly.

I have been racking my brain on is how much torque it takes to stop the trim wheels spinning.  It's a physics problem that I'm just not smart enough on yet to know the answer to, but I'm working on it.  For now it's trial and error.

At this point I'm thinking of switching to a simple design using a single powerful geared motor with a poly clutch and appropriately sized sprocket, as in pic attached. 

Regardless, the motors themselves will need to have high torque to get things moving and stopped quick enough, which means a high gear ratio/low RPM.  Right now I'm guessing the large sprocket at the bottom is currently about a 3:1 ratio with the small sprocket above.  So a max RPM of about 30 RPM on the motor should get me about 90 RPM at the top in this configuration.

I think a polyclutch is going to be the only way to help keep the gearing on the motor from breaking.  I have bought some Dynatect multi-plate adjustable poly clutches for the TQ levers and they work great so I may do the same here.

[MistyBlue]

I found this interesting video that shows how to determine the torque required to slow a spinning disk, which I think is what I need.

https://www.youtube.com/watch?v=LYsG78Ba1Xw

I'm basing my calculation on assuming the trim wheels are made of epoxy resin which has a mass around 1.2 g/cm3.  I also assumed the trim wheels are a solid cylinder to make calculations easier, so mass is higher than it actually is.

At 85 RPM, if I'm doing my math right, to slow both trim wheels to zero in 1 second, takes around 27.9 oz-in of torque.  Seems a bit low but stopping them in 0.5 sec doubles the oz-in required, so maybe it's not far off.  If the instantaneous load on the Pololu 37D motors is 350 oz-in then it should be more than enough.

That's if a motor were connected directly to the trim wheel axles, but we have gearing involved which I'm sure changes things.   We'll find out.

[bernard S]

naa  ya maths is wrong   its done this way   wet the pointy finger and  place in air above your head now turn around and say let's do some sketchy sgite l and it will be right lol

[MistyBlue]

naa  ya maths is wrong   its done this way   wet the pointy finger and  place in air above your head now turn around and say let's do some sketchy sgite l and it will be right lol

Lol, I feel like that's how it's been done until now.

[sagrada737]

Hello Tony,

Here are a few observations...  With all respect, I think you are over-working the Trim Wheel issue/problem.  On the real 737 TQ Trim Wheels, disconnected from its drive source, the amount of breakaway torque required to move the Trim Wheel is fairly low.  But keep in mind that we are adapting a real 737 TQ Trim Wheel to the needs for our Sims.  Thus, all we need is to "emulate" the Trim Wheel movement, such that it does a reasonable job of moving the Trim Wheel in sync with the flight sim, giving an added sense of Sim immersion to the cockpit experience.

In the real 737 aircraft the actual/physical Trim Wheels are a small part of a very complex electro/mechanical/hydraulic aircraft Trim System, that involves numerous sub-systems.  Did I say complex?

Check this out...  If you want to understand a bit about the 737 Trim System, here is a good overview that technically discusses the "Trim Functions", as related to the nature/history of "runaway trim controls".

https://www.satcom.guru/2018/11/stabilizer-trim.html

I am guessing that you are not trying to duplicate all these highly complex systems, and that you only want to move the Trim Wheel in sync within the 737 flight characteristics.

If that is the case, then my advice is to "keep it simple"!  Use a small 12 VDC gear motor, 10 to 20 RPM, coupled to a chain sprocket (20 to 30 teeth), and control the motor via a simple bi-directional DC Motor Controller that responds to FSUIPC trim status/requirements (start/stop/trim Up/trim Down).  This will get the job done with minimal effort and expense.  In essence...  Keep it simple.

Now...  Some folks don't want to be told to keep things simple -- and that's fine.  In this Sim hobby, we do exactly what we want.  More power to ya!  That's part of fun and pleasure in building our Sims -- we get to do what we want.  But I would say this...  The more complexity and realism we try to build into our Sim, the longer it takes to complete our Sim projects, and more expensive they become, and the more maintenance and trouble shooting these systems involve.  All reasons why some folks don't ever finish their Sim Projects.  To a large degree, this boils down to building a Sim that falls within our capabilities and resources. My 2 cents...

On a practical note...  It might be best for you to build a "bench prototype" so you can experiment with different motors and gear ratios, and control schemes as interfaced to your Sim.  Once you are happy with the bench prototype performance, you can then integrate that design into your actual 737 TQ. 

Best wishes on your Trim Wheel project.

Mike

[MistyBlue]

I hear ya Mike.  Not trying to over complicate.  Just trying to work with/improve what was already in this TQ if I can.

I just designed a new main gear that integrate the adjustable single disc Dynatect polyclutches I have lying around into the hub. They should be a 1 for 1 swap of the existing gears and take the load off the geared motors.

After the holidays when things quiet down I'll 3D print them and see if they work the way I'm hoping. 

I like the idea of a bench test, but in this case I think it would take more time to set up than testing it on the actual TQ.

Thanks, btw, for the link to the page on the stab trim.  Very interesting read!

Happy holidays everyone!

[bernard S]

oh now thats tasty I have to admit .. will be going down same path kinda when it comes to my TQ as there's two and linked hydraulic and cables  will be fun so I feel for you Mistyblue

[blueskydriver]

Hey Tony,

Been watching some YT videos that might be what you're looking for...Strain Wave Gear or aka Harmonic Drive.

Here are two videos that I thought were very informative:

https://youtu.be/xlnNj9F37MA

https://youtu.be/Emvo3bLT-Z4

IMHO, you're looking for little to no backlash, but with enough torque to satisfy the 737TQ Trim Wheel Gear; thus, this type of gearing could solve that. However, in these videos they're using 3D printed materials, so you might have to consider changing some aspects of these items.

Still, this concept is much easier since it uses stepper motors which are more capable of handling the drastic stop/start characteristics of the trim wheel.

Of course you could just go with a 12v car door window motor and get the simulated factor, which is perfectly fine, as there are a good amount of builders here who go this route...myself included.

Albeit, if you wanna stay a step up, go with this type of gearing and a stepper motor...lol...pun intended. 

John

Edit: Not sure if this will get the right RPM, but torque would be there....

[MistyBlue]

Well, sometimes the simplest solution is to stick with what you already have. 

I tried 3D printing the two replacement main gears with the single disc polyclutches I had, as in my image above.  The gears were a perfect fit, but the polyclutches were way too weak.

Rather than spend a couple hundred bucks on a couple of stronger poly clutches from Dynatect (which takes weeks to get), I decided to try the new 1:100 Pololu motors that just came in with the original gearing (no clutches).  Keeping the speed low enough seems to be working.

Hopefully it'll last until I replace this TQ with something better down the road.  If not, Pololu has some 1:150 geared motors that handle quite a bit more torque, so I can give that a try.  Two motors at $25/ea is still cheaper than two poly clutches at $100+ each.

Thanks all for the advice and indulging me!  Hope you all had a great Christmas and an even better New Year!

[FredK]

Tony

I have my TQ pulled out and have been working on some improvements in mechanical design.

As part of that I have reengineered the trim wheel motor mounting and drive system to deliver a smoother rotation.  I replaced my old DC spur gear motor (253 rpm) with a "heavy duty" planetary type (432 rpm) from Servo City. That will up the maximum trim wheel rotation speed to just over 200 rpm which is plenty enough to mimic the real thing. The question of course as you raised earlier is will it hold up with abrupt reversals of motion. My old motor was still working after nine years but at half the trim wheel rpms of the new one.

If it doesn't, I am going to look at the following:

https://www.mcmaster.com/clutches/torque-limiters-for-chain-and-belt-drives-4/for-shaft-diameter~3-8/maximum-torque~72in-lbs-/

It is a simple solution that involves swapping out the drive sprocket, but is a bit of an expensive proposition.  So we will see.

BTW...McMaster-Carr is an excellent resource for mechanical parts.  They have pretty much anything you can think of, and orders are processed same day with inexpensive next day delivery.

Fred K

[MistyBlue]

Thanks Fred.  I was told by Pololu that the planetary gear motors are much stronger so you'll probably be fine, but I'm curious how it holds up with the higher speeds.  If it continues to work I may go that route.  Of course, it depends also on how you have geared the trim wheels as to how much load the drive motors sees.

Interestingly, on the first 1:100 Pololu motor I tried the spur gear broke loose from the shaft.  Didn't break any gearbox gears but the pinion gear was just spinning on the motor shaft.  No way to fix that and Pololu doesn't sell replacement parts. Kind of surprised they don't put a D shaft on the motor.

Anyway, I've since turned down the motor speed further and limited the trim movement in the Prosim config and I'm getting  enough movement to indicate the trim is working without overloading anything so far.

A clutch, or perhaps the wiper motor John mentioned earlier, would be a much stronger solution.  That reminds me, I need to double-check I put a flyback diode on the drive motor.  I think the Phidgets servo card has protection built in, but it doesn't hurt to put one in.

Yes, I'm very familiar with McMaster Carr...fantastic selection of mechanical parts for just about anything!