Bumpy Stepping Motion (2 Dof Flight Chair)

Shortened the levers to 45mm and it's still quite rough. I can feel each single gear, if that's in fact what it is. I think I'll try to come up with some sort of dampener.

One last thought from my end would be to try the rig without you sitting in there.
Reason being, just to see if your current setup can handle all the weight that it is now bearing. Now sure how you'll feel the bumps without being on the chair though...


Also yes, Flypt can be tricky to setup, but you get used to it. I'll post some links which can be useful to you.

LeBois Racing shows the Flypt setup.


my own thread here as i was struggling with Flypt and got loads of help from the forum.
https://www.xsimulator.net/community/threads/2dof-all-controls-inverted-flypt.19143/

I also came across Departed Reality Sim Manager, which is supposedly simpler than Flypt. It has a demo which you can check out.

Much appreciated. I'll take a look at the video later today.

I've tried off the seat and it still steps. The video I posted earlier where you can see the stepping was without me on the chair.

I still am not 100% sold on the cogging possibility because I didn't notice it behaving this way when I used my first iteration of the chair. It felt kinda mushy and lumpy, but not like hard steps. I'm kind of wondering the motors or gears were damaged from some hard motions or something? I'm still coming up with a few more things to try before I start thinking about a different motor or gear solution. If anyone has any more ideas, please keep them coming. I'd hate to replace the motors only to find out it wasn't the problem.

You can always connect the motor directly to the 24v dc supply - that is without the IBT. Disconnect the linkages when you do that and you'll see the motor shaft turning, you'll have an idea whether the motor steps due to the gears and not due to the electronics.

I've disconnected the linkage and pot belt and let it spin at the slowest setting I could get it to go by adjusting the pot with my fingers and it felt smooth, but it was still quite a bit faster than what the flight sim would achieve in a gentle climb.

If I was a good programmer (which I am not), I could modify the smc3 code to send a constant low pwm signal to have the motor turn at a very slow speed. That could be very informative. Any idea how I would do that? I have experience with python, but just basic stuff.

IMHO I don't think what you're seeing is cogging. The movements look too rough. In my case the levers move smoothly I can't see any visible jerkiness. What I did feel was a series if micro-accelerations and decelerations, which I couldn't even feel in a not-sterile environment (in MSFS flight), not sop-and-goes as your video shows. that would be impossible to ignore.

I also think that there is plenty of power with those motors, especially for a 2DOF.
I see from your topic that you're struggling to get a slow movement for testing. It's easy with FlyPT Mover: just add a source of type "Loop" and choose Sine wave type, then just assign it to roll or pitch and you can control the speed and amplitude as you want. You can make it really slow.
One other thought is IBT2s are notoriously unreliable with 24V, that's the reason I'm not using them in my 6DOF (I used them in 2DOF with 12V PGSAW motors just fine). It's possible that the issue has something to do with IBT2s as well. Many people operate IBT2s around 19V because of their tendency to burn out at 24v. If you're doing that, the motor may not have enough power to move properly.
Also, your SMC3 calibration may be wrong. In theory, it can behave this way if it tries to move the motor too weakly and increases the PWM in steps until the motor finally moves to the desired angle...

I suggest you start with a sine wave in FlyPT Mover, and completely unload the motor - disconnect the arm. See what the motion is like. If it's smooth, the issue is definitely not cogging.
Oh, I just thought of another potential cause - somebody had an overheating problem with one of his Yalu motors and after disassembly of the gearbox saw a huge buildup of grease on the side of the gears, that made the large gears sticky. He removed it from the side and the motor stopped overheating. I guess the gears could be sticking and lead to your problem...

Thanks so much Ronan, I wasn't aware of that function in Flypt. I couldn't find ANY documentation on how to set it up however, other than the video that was posted previously here. I guess I need to start there.

I was also having the squishy sloppy motion as well, but was willing to live with it. However, now with this new problem which didn't occur before (or at least that I noticed), I just bought two of these NEMA17 motors which have a much higher ratio. They won't be here for a week though so I'll go ahead and try get get flypt to send a wave to an unloaded motor to see what I can see. Thanks again.

https://www.phidgets.com/?prodid=1080#Tab_Connection__Compatibility

Try zeroing kd. I know when I zeroed kd and ki things got ltos better and they were set to very low values.

I've tried just about every combination of settings I can think of.

While I'm waiting for the new motors, I'm working on figuring out flypt (and failing so far). Honestly I have no idea how people are figuring out how to use this program. I can't find a single tutorial anywhere on how to set up from scratch. I'm kind of fumbling my way through via the very limited info on their website, guessing at what I should do next. And the UI is so tiny (4k monitor), it's difficult to read and interact with it. However, I do see the incredible potential with being able to do pretty much anything you can think of with it, so I'll keep pushing on. The YT videos I've found are mostly European and I haven't found any that are a beginner tutorial. I feel like I'm missing something that is obvious to everyone else.

I'm making some progress, discovered 3d view and tool tips so I could actually create the rig. I've got a sine wave going to the 3d view of the rig, but I don't know how to connect the rig to the serial output? The serial output seems to be connected to the ard, but not sure.

Ok, after much effort, I got flypt going and sent a slow sine to the motors without a load. Here's the video. I tried Kp from 150 to 300. I tried adjusting PWMmin from 0 to 40. I tried Fpwm from 15 to 30. Nothing made a significant change.

https://www.dropbox.com/scl/fi/uxqt...ey=dis78ixdbidy0v1u9i60jlpme&st=sbkgdx9f&dl=0

I do have an extra IBT2 iirc... I could swap one out and see if that changes behavior.

Tried swapping ibt2, no change. I hope changing to these new 56:1 motors will make everything buttery smooth. I noticed the new motors have a built-in hall sensor, which is great if I can manually set the 0 position. They should be here Thursday.

Hi Tim, looks like you've exhausted the software route. i guess the motors are next then.
I'm surprised how 2 motors can have their gearboxes damaged at the same time - mainly a common cause then - as this would have been highly unlikely.

all the best.

Just an update. The new motors are brushless and way too low power anyway, so that won't work. I'll be sending them back. I'd really like to find a way to get these motors I have to work so I don't have to dump much more money into this project. I took them off the chair and cut all the wiring as I was expecting to replace them. I shortened one of the 4080 base pieces to reconfigure for the new motors. It'd be a pain to replace that piece if I end up putting these motors back on. Anyway...

I took one of the motor gear boxes apart and cleaned out the excess grease. I don't see any damage or wearing. It all feels pretty smooth except for the minor cog vibrations. I also tried to assess the gear ratio and found that it is somewhere around 50:1 instead of the 5:1 that I thought it was. Which makes me wonder if a different 50:1 motor would behave differently.

I hooked it back up and see no difference. There is a lot of low amplitude noise from the sensor but I just have the wires twisted together. When I pinch them tight, it improves. Unplugged in the PSU's does not change the noise.

I know I could dump a bunch of money and build a couple of the DIY actuators for smooth motion, but I'm getting some heat from the wife already, and I don't really feel like rebuilding yet again. I want to fly, but I can't fly it like this.

I have even been looking up using a cog and linear rail system which would definitely smooth it out, but those linear track blocks are expensive themselves for the larger ones. Or I could put an additional gear box on it to cut it in half or something. I feel like I've been researching for days and keep hitting dead ends.

Thanks for the input! I'll consider improving each of these, but firstly, the problem lies with the motors with or without load, so it's not a balancing issue. During high-speed motions, the rig operates quite well and feels smooth. It is well-balanced when I'm sitting on it. Please refer to the previous video I posted that shows the visible jitter of the rotation at slow speed.

This may be my next thing to try. I keep reading people saying they "burn through" Ibt2's but that hasn't been my experience. My PSU's are running at a measured 24v with a fan on the drivers and I haven't noticed any issues. But I'm wondering if they may be damaged? How would I know? Like I said, the motor seems to run fine at higher speeds. I just replaced one of the drivers yesterday with the third one that came in the original order and noticed no change in behavior. Maybe it's a PSU problem? Again, how could I diagnose this? I know less than I need to about electronics.

I need to do my due diligence to try to get these motors to work. I don't want to dump more money into this project.

Last night I measured the voltage going to the motor while it was running at slow speed and it oscillated with the sine wave between +/- 0.6v. If I had a power supply that could supply enough current, I could run it slow and see if it still behaves that way. Perhaps the PSU or the IBT2's are just not capable of running it that slow? Maybe it's not a problem with the motors after all? Just an idea that I'm currently investigating. Trying to locate someone with an adjustable power supply. I also ordered an ammeter to, at the very least, give me more data to understand the situation.

I finally got the attention of a friend who is a guru all things electronic/mechanical and he says that this motor isn't going to ever run smoothly at slow speeds with pwm due to the high gear ratio. Makes sense. Back to the drawing board.

Hi, Sorry to be late to the party.
I'm having the same issue with a DOF Reality P3 machine. From what I've be able to gather, the problem is a combination of 2 issues that tend to present while slow flying. The founder of DR had a conversation with mark from Flight Sim Hanger regarding this issue, and their fix for it. Basically they are using the 3000 RPM motor in combination with a planetary gearbox and a ballscrew. This is supposed to eliminate both the backlash (position searching) and the issues with the 2 pole PMDC motor (cogging). Their new machines have this configuration, and they are supposed to be coming out with an upgrade kit for older units. I emailed them, and they wanted $520 +++ for just the ballscrew update. Can't imagine what the entire upgrade would cost. I've chosen to build my own upgrade kit. Here's a link to my progress. https://www.xsimulator.net/community/threads/sort-of-diy.19294/ I had to wait for a longer end-mill (that arrived today) to make cuts in the ballscrew mount. I hope to have at least one of them done in the next week or two. Here's the interview