×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

INDI focuser driver for Waveshare Stepper Motor HAT for Raspberry Pi / Rock Pi

  • Posts: 1119
  • Thank you received: 182

My 2 cents on this is that upon microstepping down to 1/32 and using only one step for each move, the motor does not have enough torque to reliably move that one microstep.
That will be very difficult to fix, except for replacing the motor with a much heavier NEMA 23 or even NEMA27, or putting an elaborate transmission in place to reduce the motor torque, or by replacing the NEMA17 with a geared motor (which you are planning to do, using an 18:1 gear ration, I presume).
It really comes down to the problem that the helical focuser on the RedCat has A LOT of internal resistance, especially going out, and that the focal plane is very narrow. Also, the angular movements of the helical focuser have to be exquisitely small, given the narrow focal plane of the lens.
I doubt a belt-driven focuser can achieve that consistently. Even deepskydad says that their focuser needs to be moved out 5 mm at least, lest there is too much friction on the tube (I can confirm the friction part).

As for the STL files: I made the designs using FreeCAD, not Fusion360, happy to share those as well, but the STL files should be translatable immediately without further modifications.

The only thing you would have to forego moving the tube all the way back at the end of the night. Not necessary, IMO. But you can always do that manually, if you are worried about the tube sticking out 10 mm or so.

Hope that helps.

Jo
3 years 10 months ago #54510

Please Log in or Create an account to join the conversation.

  • Posts: 1119
  • Thank you received: 182

Direct drive, no gears on the motor. Gear ratio is approximately 1:2, due to moving out the secondary gear segment with the extended arm of the gear ring clamped to the helical focusing ring of the telescope.
3 years 10 months ago #54511

Please Log in or Create an account to join the conversation.

  • Posts: 421
  • Thank you received: 102

When using the Waveshare INDI driver that I wrote, a single step in the INDI control panel will equate to 32 microsteps at the motor. I always end on a full step boundary. I use microsteps for smoothness and quietness, not for increased resolution. If it were to stop on less than a full step, then the motor would need to have power applied to it constantly to hold that position, thus using more power and causing the motor to get warm (or very hot if you haven't tuned the current limit). But with my approach, that allows me to turn off power to the motor when not actively moving for focus.

So I'm not sure if that is esb's problem of it not always moving the same amount with a single step. Does that also happen if you disconnect the belt, just letting the motor run free? If it works fine when the motor is running free, but misses a step when connected to the belt, then it is most likely a lack of torque. You can try slowing down the motor with the delay settings, since stepper motors lose torque when running faster. You can also try using less microstepping via the DIP switches. You will need to modify the source code to match the micrstepping setting of the DIP switches. Eventually I'll make that a configurable option in the INDI control panel.
3 years 10 months ago #54512

Please Log in or Create an account to join the conversation.

  • Posts: 59
  • Thank you received: 3
I'm guessing you're getting about a 10:1 gear ratio since the drive gear on your motor looks to have +/- 11 teeth and I'd estimate that if the gear on the scope was a completely circle, it would have ~100+ teeth. In my case, I have a 6.5:1 ratio (20 teeth on the drive gear and 130 teeth on the scope side). And now I'm adding a stepper with a 5.13:1 gearbox.

I did some additional experiments with the geared stepper and discovered that the problems I was seeing with inconsistent (even backwards) movement went away when I increased the step delay in the WMH driver. I had been using a 40us delay per step, and am now using 100us. I'm guessing it was trying to micro step too quickly. I bet that's the issue I was getting on my regular stepper as well. I'll have to see if it performs better with the higher step delay. I also have the option of turning micro stepping off completely. The only reason it was used in the first place was to smooth out the motion and not to actually increase resolution. But if the stepper loses torque in the process, it might be better to just use the native step size of the motor. The gearing should improve both torque and resolution.

I had been looking for your original CAD files to adapt your gear design to my layout. However I can do that pretty readily in Fusion360 if I need to. I also probably have a Nema bracket around somewhere if I end up trying out your design. I'm still hopeful that I can get the belt driven solution to work.
3 years 10 months ago #54513

Please Log in or Create an account to join the conversation.

  • Posts: 59
  • Thank you received: 3

Ha... our messages crossed. I think the timing is indeed my problem. With the geared stepper, the inconsistent (even reversed) movement happens under zero load. I think I'm just pulsing the motor too quickly.
3 years 10 months ago #54514

Please Log in or Create an account to join the conversation.

  • Posts: 421
  • Thank you received: 102
Hah! Looks like esb was all over it! We both posted essentially the same thing!
3 years 10 months ago #54515

Please Log in or Create an account to join the conversation.

  • Posts: 1119
  • Thank you received: 182

Correct, the ratio in that case is 13:1, the full circle of gears on the focus ring is 130 gears, if I remember correctly.
But you can take A LOT of strain off the motor when moving out, using the rubber band support see in my earlier pictures. That can reduce the strain on the motor by an additional 50% or more, depending on how strong the rubber bands are you choose to install.

Anyway, there is a lot you can do mechanically to support a weak motor with a poor transmission ratio.

And don't forget the reduced backlash. That is something that is difficult to eliminate using the belt drive.

I also added the FreeCAD files to the stl folder. Not sure you can translate them into Fusion360 format, but if you can, there they are!

Best

Jo
Last edit: 3 years 10 months ago by Jose Corazon.
3 years 10 months ago #54516

Please Log in or Create an account to join the conversation.

  • Posts: 59
  • Thank you received: 3
And a problem like this wouldn't be complete without a facepalm. I just noticed with the stepper attached to my RedCat, if I ran it from 0 (fully retracted) to nearly max (500) and then back to 0, it kept missing the zero point by a few degrees (it was always short). It turns out the grub/set screw for the pulley attached to the stepper is loose. Double-d'oh! That means I have at least 10+ steps of backlash. I'm guessing it will work much better with a tight pulley.
3 years 10 months ago #54518

Please Log in or Create an account to join the conversation.

  • Posts: 1119
  • Thank you received: 182
Well, all problems in astrophotography are mechanical or optical in the end....

:)
Last edit: 3 years 10 months ago by Jose Corazon.
3 years 10 months ago #54520

Please Log in or Create an account to join the conversation.

  • Posts: 421
  • Thank you received: 102

Consider yourself getting off easy! When I did the Rowan belt mod for my Orion Atlas mount, one of the set screws on a pully deep inside the mount wasn't tight enough. Took me a month to figure out why I was getting such bad backlash in Dec movements! So I had to tear the mount apart to tighten that screw. I put blue Loc-Tite on those screws, just because I was mad at them!
3 years 10 months ago #54521

Please Log in or Create an account to join the conversation.

  • Posts: 59
  • Thank you received: 3
With my existing stepper, backlash now seems to be around 1 step. I can see clear consistent movement for even a single step and when changing directions at most that first step doesn't show full movement. I can see that with large moves in one direction followed by single step moves in the other. When does darkness get here? I really can't way to try it out again.
3 years 10 months ago #54522

Please Log in or Create an account to join the conversation.

  • Posts: 59
  • Thank you received: 3

Its especially ironic because the grub screw is one of the first things I advise others to check for many different kinds of 3d printer issues. I need to take my own advice.
3 years 10 months ago #54523

Please Log in or Create an account to join the conversation.

Time to create page: 0.943 seconds