Dual Drive Setup (with driveshaft)

[admin]

All the parts came in yesterday and today so tonight I set about
assembling the system.
However, before disassembling my machine I forget to cut the mount
plate for the bearing. (on the exploded view drawing this is item #16,
SM003).
Instead of putting the machine back together and cutting the piece I
decided to just mock it up and see how it looked. In the process I
found a couple of things out:
1.) MOST IMPORTANT: We don't need a bearing plate (you know, the one I
forgot to cut)! We just need a different bearing than the one I had
used for the design. Because I'm thinking of using a flanged bearing,
and there is already a hole in the plate, why didn't I just use a
bearing with a matching OD? I don't know the answer to that question.
But, I ordered one from here: http://www.rcbearings.com/5x13x4-f695zz-flanged/
it should be delivered sometime this week.
2.) If you order the mounts from pololu, order the bolts that go with
it! They's #4x40 bolts only about .25" long. I'll need to go to Ace
Hardware on Monday and pick some up to finish the prototype.
3.) Because both Pololu mounts are threaded, you'll have to drill out
one set of threads. Seems obvious in retrospect, but you can't thread
two threaded pieces together! It's an easy fix because it's aluminium
(soft). I used an 1/8" bit from my index and reamed the holes out.
Now, one set of mounts are threaded, and the other isn't. Just like we
want it to be.
4.) There isn't enough room at the end of the motor shaft for both the
pulley to sit in line with idlers and for the mounting hub to bolt on.
Because of that, the mounting needs to be done first, then the pulley,
and in order to keep everything lined up, we'll have to move the motor
off the plate a bit. Ultimately that shouldn't be an issue if the
right hardware / spacers are available. We'll have to stand off the
drive side motor by about 1.5". I'll get the exact dimensions once I
get the rest of the pieces. Looks like maybe an M3 x 40mm or 45mm bolt
will do the trick.
Even with all of those hiccups I managed to mock everything up and am
pleased to say that I still believe the design is going to work quite
well. I'll report back once I get all the *new* pieces in.
-Edward

[kf4mot]

My vote is for shaft drive. Assuming the performance is just as good or better than a dual motor setup.

[Enraged]

Is the cost of the driveshaft and all associated parts less than the $16.95 for the stepper from Sparkfun (and associated pulleys, belts, etc)?

these might be an option for the driveshaft: http://www.sparkfun.com/products/10006

[edwardrford]

I'm with you on that one @enraged. For me, I'd rather spend the extra couple dollars and have two drives. But, for some, breaking the 'best practice' rule of 1 stepper motor per stepper driver, is too far out of the comfort zone and that's OK too.

If all works out well, either method will work just fine. Then, if at some point and time someone would decide to add a second motor (perhaps they upgraded driver boards to support 4 axis) they would only be out the cost of the rod + couplers. A net total of about $10.

[sandro]

My concern with the driveshaft is introducing twist in the shaft, and having that translate into backlash on the one side of the mill. I suppose if forces aren't large that's not too much of a problem?

It seems like more and more driver boards (tinyG, G540, etc) actually have 4 drivers (to support extruders?), so it seems like a dual-motor solution will get more cost-effective as time goes on.

Has anyone asked Ril3y or Alden about the possibility of a 4 driver version of GrblShield?

[Enraged]

another option is running 2 Y drivers on the new version of the buildlog stepper shield, rather than 2 motors sharing a driver.

[aldenhart]

I like the shaft idea over the dual motors. I have machines in both configs. The main issue with a dual drive design is racking - where the motors get out of square and the machine won't move at all. Another issue is real homing. Truly homing a dual drive axis is a way more complicated operation if you do it right, involving gradually moving both motors against isolated (i.e. not shared) homing switches. Most people just eyeball it and let it go at that, which is less than optimal. The shaft design avoids these problems. The shaft design is also cheaper and less moving mass. I think the shaft can easily be made rigid enough to avoid twist for the loads Shapeoko is capable of

As for a 4 axis controller: as you know, TinyG has 4 axes now. If there is enough call for a 4 axis grblshield we could certainly get one out there. The 3 axis version will ways be cheaper, though.

[JCPhlux]

I am thinking larger scale so the cost of a shaft is going to go up for a larger size. Once I setup the dual drive it does not matter what width is it will work.

[nogthree]

I really like the shaft design for its simplicity and it also solves the problems of the sides getting out of sync like Alden indicates. If we do find that under torque the two get out of sync I guess you can always reinforce the frame between somehow.

[ForestRoyse]

another option is running 2 Y drivers on the new version of the buildlog stepper shield, rather than 2 motors sharing a driver.

I will be going this route this once my Pololu drivers arrive. I'm hoping racking won't be too much of a problem but will report if it is.