Treatise on Y axis Racking error (Stop Calling it Backlash)

discussion of design changes / improvements / suggestions
Post Reply
minorthreat
Posts: 43
Joined: Tue Apr 24, 2012 4:34 pm
Location: Boston, Ma

Treatise on Y axis Racking error (Stop Calling it Backlash)

Post by minorthreat » Sat Dec 29, 2012 7:16 pm

On a stock shapeoko all of us are seeing ~2-3mm of Y axis racking error (@5lbs on the opposite side of the belt drive). This has been mislabeled "backlash", or "lag" a few times. Backlash is strictly a motion dead-space due to teeth tolerance in transmission elements and lag is more of a side effect of spring rate dependent racking.

The easiest way to see this is to grab one side of your shapeoko Y gantry and pull back and forth on the on the other side. You might be tempted to think with the right V-Wheel preload and belt tension you could tune out this error, but you would be wrong. Racking is a spring rate error, ie it takes force to activate the mechanism. Increasing the Eccentric nut V-wheel preload creates more friction (more force) making the racking worse. In fact the racking is still evident if you clamp one side of the Y axis directly to the makerslide. Our 'spring" is simply the bending of the thin motor mount plate.

There are several easy ways to fix this:
Dual Motor mod (most popular)
Drive shaft mod (my fav)
Thicker Motor Mount plate

I chose to implement the drive shaft mod, since I do not like the over constraint of the Dual motor mod (jamming will happen some day).

So the question is has anyone had success with the 3rd option (Thicker Motor Mount plate)? This seems like the lowest production cost improvement and could be easily rolled into the next batch.

cvoinescu
Posts: 4442
Joined: Thu Jul 19, 2012 6:50 pm
Location: Camberley, UK
Contact:

Re: Treatise on Y axis Racking error (Stop Calling it Backla

Post by cvoinescu » Sat Dec 29, 2012 10:25 pm

Thank you for the correct term. (I've never called it backlash, but I did call it lag because I didn't know the proper word.)

For the record, I really like the dual motor approach. If one side skips and the other doesn't (which doesn't happen all that often) it's easy to re-align, whether it was the motor or the belt that skipped; if your belt skips on one side, it's not that easy. Also, on my machine, which has the X rail further back, a shaft would have to go through the spindle...

Anyway: my motor plates are more flexible (again, because the X rail is attached to a corner) than the ShapeOko plates; the thickness is roughly the same (mine are 2 mm, and the original ShapeOko plates are about 1.9 mm). The increased flexibility makes the racking worse, no doubt about that. So one more data point for you, albeit in the wrong direction. :)

I don't think any reasonably thick plates would solve the problem to my satisfaction, so I'm not going to try. That said, I was surprised how rigid the standard MakerSlide carriage plate was. Maybe some thicker aluminum in the same alloy as that would do a better job than steel, and without making the plates too heavy. Also, with belts on the outside or on top of the rail, you could have the motor plates less tall, and raise the Y rail to get the same clearance under the X carriage.
Proud owner of ShapeOko #709, eShapeOko #0, and of store.amberspyglass.co.uk

edwardrford
Posts: 1250
Joined: Mon Apr 09, 2012 5:40 pm
Location: Dixon, IL
Contact:

Re: Treatise on Y axis Racking error (Stop Calling it Backla

Post by edwardrford » Sun Dec 30, 2012 12:42 am

Interesting timing on this post.

A few days back I was talking to someone about the racking issue. And told them this story:

My first iteration of machine used end plates and motor mounts made from 3/16" (0.1875") A36 mild steel. I immediately said "these are too heavy" before even mounting them to my machine and ordered another set of plates made form 10ga (0.1345"). Those worked just fine, with no noticeable racking. At one point I tried out a really thin pieces of steel (16ga), and it was terrible. Even with dual drives the gantry was racking (actually it was twisting) like crazy due to the flimsy motor mounts. I finally settled on the 12ga plates for the production machines as I felt that size was the best cost to performance thickness.

However, when I started selling the beta upgrade kits, I went back to 10ga steel, thinking that most people would be buying the kit to extend their machines, thus applying more force, thus needing a heavier duty plate. It was really only then (when I applied an upgrade kit to one of my own machines a few weeks ago) that I realized the thicker plates were the reason I wasn't seeing any (or seeing very minimal) racking during my prototyping stage..... So, long story short, thicker plates DO indeed help with the racking issue.

Another option, is to make more connections between the motor mount plates, which is something that i've been really interested in lately. The dual-x upgrade helps this. Another option that I really like and was suggested by Ril3y, was to run a 5mm threaded rod through the rear smooth idler across the gantry through the other hole. Even without dual drives OR a drive-shaft it's reduced racking considerably. I'm working up a newer prototype the connects the plates at the bottom as well, we'll see how that works out.

-Edward
Shapeoko 1 #0 - a couple of upgrades.
Shapeoko 2 #0 - a couple of upgrades.
Shapeoko 3 #2 - Stock

chips1776
Posts: 5
Joined: Fri Sep 21, 2012 12:02 am

Re: Treatise on Y axis Racking error (Stop Calling it Backla

Post by chips1776 » Sun Dec 30, 2012 3:17 am

edwardrford wrote:Interesting timing on this post.
Another option, is to make more connections between the motor mount plates, which is something that i've been really interested in lately. The dual-x upgrade helps this. Another option that I really like and was suggested by Ril3y, was to run a 5mm threaded rod through the rear smooth idler across the gantry through the other hole. Even without dual drives OR a drive-shaft it's reduced racking considerably. I'm working up a newer prototype the connects the plates at the bottom as well, we'll see how that works out.

-Edward
I made two changes to reduce the racking I was seeing. One was adding a second motor and the other was to add brackets to the x makerslide ends. The latter considerably reduced the amount of flex I was seeing.

minorthreat
Posts: 43
Joined: Tue Apr 24, 2012 4:34 pm
Location: Boston, Ma

Re: Treatise on Y axis Racking error (Stop Calling it Backla

Post by minorthreat » Sun Dec 30, 2012 4:26 am

Wow, I haven't thought of the extra stiffness across the gantry thru the idler (or any other) hole. That's a brillant (cheap) idea! you could even preload a tube or something with the thread rod to lock the 2 plates together.

I'll run to ace tomorrow and test it out.

Post Reply