Welcome, Guest. Please login or register.

Username: Password:
Pages: [1] 2

Author Topic: harmonic drive directly driven stepper EQ mount with EQMOD compatibility...  (Read 16021 times)

pterodyne

  • Sr. Member
  • ****
  • Posts: 91
    • View Profile

Hi Tom,

Still thinking of (possibly crazy) uses for astroEQ.  One idea I had was a high ratio harmonic gear head on a .9 degree stepper (or higher if there is a bipolar one available.  All the 1000 step steppers I've seen require a custom driver (which obviously could be interfaced with astroEQ with the right changes)

there are a wealth of harmonic drives on ebay.  Recently I saw this one which had a 260:1 reduction ratio.  Mechanical parts aside, it appears as if one could use a 400 step stepper at 32 microsteps and get over 5 degrees a second at .38 arcsec per step.

I have been putting in as follows:

stepper degrees:  .9
microsteps 32
gearbox ratio 260:1
worm ratio 1:1

Is this making sense?  I didn't use your configuration program because I have to have an astroeq or arduino connected (any way to make it have a dummy board attached for this spec-ing out purpose?) but rather I put the variables into onstep.xls which I found at stellarjourney.com (a very similar project but without eqmod compatiblity).  I've in the past put in these variables into your configuration program before though and gotten similar results.

So the questions are:


1. Am I crazy?
2. Is there a way to use the config program without an astroEQ attached just to use the calculations?
3. Can the worm ratio be 1:1 or does this break something?  I'm assuming in the end your software looks at a total reduction.
4. is a way for the layman to tell how fast it will go (*sidereal) without stalling the motor or missing steps? (the onstep.xls file does provide a maximum slew rate as part of the calculation, but it is based on interrupt rate and I have no idea what that means in this context.
5. Trying to get up to 1000 steps to improve resolution requires different drivers.  Oriental motor sells 1000 step steppers and these controllers:
http://www.orientalmotor.com/products/pdfs/2012-2013/A/usa_st_crk.pdf

interfacing with a controller like this would allow for higher spec motors.  Just wondering if you would ever consider (even for hire) to swap out the driver components/firmware that would allow this type of controller? Or will this project always stay with the easystep type drivers.

Refer back to question 1 if need be.


Thanks!  just thinking outside the box. In my mind the directly drive harmonic drive would be most interesting way to go on a custom designed mount.  Minimum of machining to be done..
Logged

TCWORLD

  • Administrator
  • *****
  • Posts: 809
    • View Profile
    • AstroEQ

Hi,

1&3)Would the work gear ratio be 1? Are you talking about bypassing the worm completely and connecting the motor shaft directly to the axis??

2)I will add a test screen to the next version of the config utility.

4) No real way of working out the mechanically achievable top speed short of scientifically measuring how much torque it takes to rotate the shaft and then comparing that to the torque/speed characteristics of the motor.

In terms of interfacing different controllers, it is *relatively* simple to do as long as they use a step-direction type interface (i.e. have a step/clock signal and a direction signal). The 'Pulse input package' in that datasheet you posted does indeed have this interface. The only thing that would have to be tweaked in the firmware is the step angle settings which could be done.
Logged
Tom Carpenter (AstroEQ)

pterodyne

  • Sr. Member
  • ****
  • Posts: 91
    • View Profile

thanks for the reply, so yes, just a stepper driving a 160:1 harmonic gear head.  This would be inline like a chronos mount but instead of a Harmonic drive/AC servo/encoder, use a high count stepper.  No worm wheel or gear.  So in the configuration interface, either the worm or the gear ratio will be 160:1, as in one 160 rotations of the stepper = 1 rotation of the axis. This is why the number of steps has to be higher, even with microstepping. 

Logged

TCWORLD

  • Administrator
  • *****
  • Posts: 809
    • View Profile
    • AstroEQ

That makes sense. You would probably in this scenario put the 160:1 as the worm ratio as EQMOD uses that for its PEC calculations. You could also then do a quick 'hack' to get AstroEQ to work without firmware modifications by selecting a microstepping mode in the config utility to be 1/4 stepping (as it will disable the microstepping mode change for high-speed movements), and set the gear ratio box to be the difference, so for example if you set the driver to 1/100 stepping, the fake gear ratio would be 100/4 = 25.
In the calculations, the microstepping mode and gear ratio are only ever multiplied together, so if you make one half what it should be, you can double the other the compensate.
Logged
Tom Carpenter (AstroEQ)

pterodyne

  • Sr. Member
  • ****
  • Posts: 91
    • View Profile

I think I understand what you are saying about the math.  What I don't fully understand is why would that be necessary?
Logged

TCWORLD

  • Administrator
  • *****
  • Posts: 809
    • View Profile
    • AstroEQ

It would only be required if you wanted to use the current firmware/config utility without modification as the microstepping mode is selected from a dropdown containing only ones achievable by the DRV8825 and A4988 chips. As other driver boards will have different methods of mode selection modifications would have to be made to support that unless you stick with one microstepping mode and tweak the maths to compensate.
Logged
Tom Carpenter (AstroEQ)

pterodyne

  • Sr. Member
  • ****
  • Posts: 91
    • View Profile

Ok.. I see what you are saying..
Logged

pterodyne

  • Sr. Member
  • ****
  • Posts: 91
    • View Profile

this is the motor/gear Im thinking of:

http://catalog.orientalmotor.com/item/packages-stepping-motor-driver-packages-0-36-0-72/eries-closed-loop-stepping-motor-drivers-dc-input/ar98ak-h100-3?plpver=11&categid=3001133&prodid=3001123&origin=keyword&selectedindexes=MTgxMjowfDE5ODI6OHwxODQ4OjB8MTgyMzoxfDE4NjE6M3wxODYyOjF8MTgxOTowfDE4Mjk6MHwxODQ1OjB8MTgyNDowfDE4NTg6MHwxODIxOjB8MTgzNzowfDE4NDM6MHwxOTY2OjB8MjAwMzowfA2&stditem#

here are the AstroEQ metrics:

Motor Step Angle: .38
Motor Gear Ratio: 1:1
Worm Gear Ratio: 100:1
Ustepping: 32

Gets a resolution of .405
Max speed sets to 1064

At 1300x2 it could be a relatively cheap Chronos call it a "Clonos" type setup.  Machining the rest would be much easier than worm gears etc..

Ill post more plans as I develop the idea more.

Tom,

Im not sure on the tricking the driver part as you posted previously.  If I set the stepping to 4, and the motor gear ratio to 25, the IVal drops to 143 based on the above.  If I leave the ustepping at 32 it gets an Ival of 399.  Either way, this is just a thought exercise at the moment.

If you get a chance, and by no means in any hurry.  Check the motor/driver specs above.  Most interesting to me is the "resolver" which sounds like on the front end it's a DIR/STEP controller, but internally it's using both step counting and the resolver for feedback.  Do you think this setup would work with AstroEQ/EQmod.  These motors look to possibly be the basis for a very beefy mount.
« Last Edit: January 22, 2015, 16:10:30 by pterodyne »
Logged

pterodyne

  • Sr. Member
  • ****
  • Posts: 91
    • View Profile

Ive decided to move on this project after much time away.  In the mean time I converted a CG5 to astroeq with pulleys, and nema 17 steppers, and also a titan to harmonic servos (50:1) with Sitech Servo 2 controller.  Anyway so on to a new project with harmonic drives directly driving the axes.

Ive purchased 2 vintage 200:1 harmonic drives.  The current plan is to use TMC 2130 drivers (possibly on onstep, or maybe Ill reuse the astroeq from the CG5), and do a 2:1 pulley stage between the steppers and the 200:1 reducer.

This gives a resolution of around .5asec/step at 16uStep, but kind of a low max speed (518xSidereal).  I don't fully understand how the TMC2130 works with astroeq, as it does the 256 step interpolation, and not sure how the ustep gearing works as I think it has to be disabled in the astroEQ configuration utility.  Supposedly these drivers are the bees knees, and other forums about onestep and in the 3d printer world people are saying they get greater speeds and torque without changing steppers.  Im 99% certain for AstroEQ the TMC2130 cannot be run in SPI mode so Ill have to set the jumpers if I don't do Onstep. 

Anyone have thoughts on this?  How does the 256step interpolation inter-relate with AstroEQs internal resolution? Does it just make the output smoother/quieter than the theoretical 16ustep resolution implies? Can I use ustep gear changing and the TMC 2130 in the astroEQ?

Ill post more as it progresses. Mostly Im concerned with the mechanics and machining requirements.
Logged

pterodyne

  • Sr. Member
  • ****
  • Posts: 91
    • View Profile

here are the harmonic drives

https://www.ebay.com/itm/182997206729
Logged

TCWORLD

  • Administrator
  • *****
  • Posts: 809
    • View Profile
    • AstroEQ

Hi,

If you disable SPI mode on the TMC2130, it should operate just like the TMC2100, which is supported.

For the TMC2100, gear changing is supported for 1/16 ustep mode. It will use the 1/16 step spreadCycle mode with 1/256 interpolation for low speed, and the 1/2 step spreadCycle mode with 1/256 interpolation for high speed.

The 256step interpolation is done within the driver chip itself, meaning that as far as the AstroEQ is concerned it is still operating in 16ustep mode. It basically just makes the motion smoother.

Kind Regards,
Tom.

Logged
Tom Carpenter (AstroEQ)

pterodyne

  • Sr. Member
  • ****
  • Posts: 91
    • View Profile

For any who are interested I have continued to develop this and am almost complete.  I have switched to Onstep as a controller, but the mechanics are identical, and If my astro eq were working well I would probably just use it since I prefer EQMOD.

Anyhow, here is the link:

https://groups.io/g/onstep/topic/new_build_harmonic_drive/16745737
Logged

pterodyne

  • Sr. Member
  • ****
  • Posts: 91
    • View Profile

Im trying this mount with a 3axis TB6600 stepper driver from sainsmart and an AstroEQMega.  Really easy to setup, I didn't even put the EN line in.  Just DIR/PUL on both axis, 5V from the Mega 2560.  Can anyone think of a reason I have to have EN?  The drivers get disabled when an axis is given a command rather than enabled.  It obviously will use more power when idle with the axes always enabled.

Logged

pterodyne

  • Sr. Member
  • ****
  • Posts: 91
    • View Profile

So this didn't end up being a good solution.  The motor had too much vibration during tracking.  Not sure why.  Same motors with other drivers (8825 etc) even at 16uStep were pretty smooth.  Tried adjusting current through the full range, no help.

Back to onstep for now.  I have a RAMPS MKS V1.4 all in one board with onstep.  Maybe Ill try that.  In theory I could change the pins in the firmware...
Logged

pterodyne

  • Sr. Member
  • ****
  • Posts: 91
    • View Profile

Got it working with onstep and TMC2100.. Still wanting eqmod though.  I bought these drivers:

https://www.amazon.com/gp/product/B00QBTKHAC/ref=oh_aui_detailpage_o02_s01?ie=UTF8&psc=1

they do up to 256 microstepping.

Tom, you had previously stated:

Quote
You could also then do a quick 'hack' to get AstroEQ to work without firmware modifications by selecting a microstepping mode in the config utility to be 1/4 stepping (as it will disable the microstepping mode change for high-speed movements), and set the gear ratio box to be the difference, so for example if you set the driver to 1/100 stepping, the fake gear ratio would be 100/4 = 25.
In the calculations, the microstepping mode and gear ratio are only ever multiplied together, so if you make one half what it should be, you can double the other the compensate.

Now unlike the original example I am not using a 1:1 motor gear, the specs are:

1.8 degree steppers
200:1 harmonic drive
4:1 pulley motor reduction
driver with dip switches to set up to 1/256 ustepping.

So for the configuration utility Im assuming I leave uStep gear changing Disabled.

Motor step angle 1.8
worm gear ratio 200
that leaves Gear ratio and microstepping level.

if I wanted to use 256 (probably anything 32 or better is fine, but for this exercise), I would put the Motor microstep level at 4 uStep, multiply my actual motor gear by 4 (4*4=16) then divide the actual driver uStep mode by the result. 

as in:
Motor gear ratio 4*uStep mode of 4=16
driver uStep mode of256/16=16 final calulated motor gear ratio.

Tom does this pan out?

Ive attached a snip of the config utility.
Logged
Pages: [1] 2