Drift / Backlash

Drift / Backlash

Postby emperorcow » Tue Jul 30, 2013 10:58 pm

So I'm not entirely sure what is causing this, but I've been having problems with what I've been calling drift on both my Y axis. Sadly, I've been having problems where on each pass of a cut the mill may vary the position as much at 1 - 2 mm along the Y, although sometimes its none. I'm hopeful that someone here may have some thoughts on what this might be, or additional steps I can take to do some troubleshooting. I've included a picture below of the problem on a cut I was doing to provide a visual. I also pulled out a dial indicator and, with it set to .1 in. to start I had the mill move back and forth .5 in. along the Y axis several times. I attached the results in a image so that they would stay nicely formatted below. These readings were taken while trying to change the settings on my TinyG in the hopes that it was the travel per rev setting causing the issue.

IMG_0647.JPG
Board showing drifting mill on each pass
IMG_0647.JPG (351.31 KiB) Viewed 2520 times

tests.png
Excel doc showing movement of .5 in. multiple times.
tests.png (8.26 KiB) Viewed 2520 times


For reference, my setup is as follows: expanded Y axis (1M), double Y motors, double X axis, and a DW660 spindle, otherwise everything is fairly stock. I have a TinyG controllerI included an attached picture for this as well. I'm actually trying to cut an electronics box I designed but can't get some of the pieces to cut properly. The ShapeOko has certainly been a self-perpetuating project, I can't get on to other things because I keep using it to try and upgrade itself!

IMG_0643.JPG
My setup
IMG_0643.JPG (311.42 KiB) Viewed 2520 times


I appreciate any thoughts anyone has, this is probably something really obvious that I'm just noobing up. Thanks!
emperorcow
 
Posts: 4
Joined: Fri Jun 07, 2013 8:08 pm

Re: Drift / Backlash

Postby cvoinescu » Tue Jul 30, 2013 11:43 pm

What are the units in your table? If they're millimeters, you're doing okay. If they're inches, it's not that good. I'm assuming it's inches.

It seems that your error accumulates, which could mean missed steps. It could be loose pulleys, too: either the grub screw not tight, or the plastic part of the pulley detached from the metal hub (I've seen that happen). I assume the machine is bolted to the table, but if it isn't, that could explain it too.

This is very circumstantial evidence, but: a missed step results in the motor skipping four full steps (or a multiple of four), regardless of your microstepping setting. If your motors are 200 steps per mm and you have the standard 18-tooth MXL pulleys, four full steps work out to:
0.08 inch/tooth * 18 tooth/rev / 200 steps/rev * 4 steps = 0.0288 inch
In your table, there's either barely any drift, or jumps of about 0.025, so it could very well be that.

The motors make a loud "thump" noise when they skip. Did you hear anything like that during your test? Given the numbers, they will have skipped four or five times.

The motors could skip because they don't have enough current for the acceleration you're trying to get. Either reduce the acceleration, or increase the current.

Another reason could be that you have the TinyG configured to power down X and Y motors when they're not moving. That feature causes problems on the Shapeoko. You can do it for the Z axis, but if you do it for X or Y, you'll get this kind of drift. Paths with curves and diagonal lines will work fine (both X and Y must stay on), but those with long straight lines parallel to the axes will suffer.

Yet another reason could be a mechanical problem. Check that you can move the Y axis smoothly with the motors not energized. It may be that the V-wheels are too tight, or maybe there's dirt on the rail or on the V-wheels.
Proud owner of ShapeOko #709, eShapeOko #0, and of store.amberspyglass.co.uk
cvoinescu
 
Posts: 4442
Joined: Thu Jul 19, 2012 6:50 pm
Location: Camberley, UK

Re: Drift / Backlash

Postby selektha » Wed Jul 31, 2013 1:59 pm

I would lower the feed and plunge rate and see if that lowers the "drift".
selektha
 
Posts: 42
Joined: Tue Jul 02, 2013 2:05 pm

Re: Drift / Backlash

Postby cvoinescu » Wed Jul 31, 2013 2:13 pm

selektha wrote:I would lower the feed and plunge rate and see if that lowers the "drift".

How does the feed rate affect the drift for traverse motion (G0, in his table), and what's the plunge rate got to do with it?
Proud owner of ShapeOko #709, eShapeOko #0, and of store.amberspyglass.co.uk
cvoinescu
 
Posts: 4442
Joined: Thu Jul 19, 2012 6:50 pm
Location: Camberley, UK

Re: Drift / Backlash

Postby Will Winder » Wed Jul 31, 2013 10:09 pm

cvoinescu wrote:
selektha wrote:I would lower the feed and plunge rate and see if that lowers the "drift".

How does the feed rate affect the drift for traverse motion (G0, in his table), and what's the plunge rate got to do with it?


When I run my machine too fast the (Dual) Y Axis skips steps, presumably my power supply (or more likely the pololu driver shared by both motors) can't keep up. I'm not sure if a similar problem exists with the TinyG though.
ShapeOko #367: Dual-Y drive, Belt on outside, 1000mm Y-Axis, DW660 Spindle, Nema-23 X/Y motors.

Primary developer on Universal Gcode Sender.
Will Winder
 
Posts: 435
Joined: Fri Apr 20, 2012 2:40 pm
Location: Hudson, MA

Re: Drift / Backlash

Postby emperorcow » Wed Jul 31, 2013 11:30 pm

So I finally got a chance to give a few things a try after work today and it was most definitely the power management settings on the TinyG. When I turned it on for the X and Y axis I was able to cut a few tests paths and the X and Y drift was reduced to almost non-existence and became very consistent. Just in-case anyone comes along and looks at this later with a similar issue, on the TinyG it was the $Xpm setting where X is your Axis. More info is here: https://github.com/synthetos/TinyG/wiki ... ement-mode.

I certainly appreciate everyone's thoughts, I had no idea this setting even existed! I did some additional testing and there's still a bit of drift, but on average (after 10 moves like before) its averaging .002 but within a significantly smaller range with almost everything in the +, so I need to play with my traverse settings to get that dialed in.

Will Winder wrote:
cvoinescu wrote:
selektha wrote:I would lower the feed and plunge rate and see if that lowers the "drift".

How does the feed rate affect the drift for traverse motion (G0, in his table), and what's the plunge rate got to do with it?

When I run my machine too fast the (Dual) Y Axis skips steps, presumably my power supply (or more likely the pololu driver shared by both motors) can't keep up. I'm not sure if a similar problem exists with the TinyG though.

I do know that the TinyG supports up to 4 motors, so I have the dual y setup with a separate driver for each of the two motors which should hopefully help with that. I'll have to play with the feed and travel rates to see if that makes any differences as well.

I've also found that, seemingly randomly, my Z axis will invert or start absolute valuing itself. Moving on the Z from -1 to 1 and back will work fine and then suddenly the axis will start ignoring the sign of the distance and will always move in the positive or negative. I don't think this is a physical issue, I'm going to see if it continues and if so see if anyone over on the TinyG Github has any thoughts.

I appreciate everyone's help, thanks!
emperorcow
 
Posts: 4
Joined: Fri Jun 07, 2013 8:08 pm

Re: Drift / Backlash

Postby cvoinescu » Thu Aug 01, 2013 12:12 am

emperorcow wrote:So I finally got a chance to give a few things a try after work today and it was most definitely the power management settings on the TinyG. When I turned it on for the X and Y axis I was able to cut a few tests paths and the X and Y drift was reduced to almost non-existence and became very consistent. Just in-case anyone comes along and looks at this later with a similar issue, on the TinyG it was the $Xpm setting where X is your Axis. More info is here: https://github.com/synthetos/TinyG/wiki ... ement-mode.

I certainly appreciate everyone's thoughts, I had no idea this setting even existed! I did some additional testing and there's still a bit of drift, but on average (after 10 moves like before) its averaging .002 but within a significantly smaller range with almost everything in the +, so I need to play with my traverse settings to get that dialed in.


Glad it was that! It occurred to me while I was typing the other possible causes.

Will Winder wrote:
cvoinescu wrote:
selektha wrote:I would lower the feed and plunge rate and see if that lowers the "drift".

How does the feed rate affect the drift for traverse motion (G0, in his table), and what's the plunge rate got to do with it?


When I run my machine too fast the (Dual) Y Axis skips steps, presumably my power supply (or more likely the pololu driver shared by both motors) can't keep up. I'm not sure if a similar problem exists with the TinyG though.


That is true, but selektha said specifically feed rate. That does not apply to G0 (traverse) moves, which is what emperorcow used to create his table. That may be nitpicking, but there wasn't really room in half an inch to reach any reasonable speed limit anyway, so even the traverse speed was likely irrelevant. Also, the Z wasn't involved in creating that table, so the plunge rate couldn't have been the problem. I was just looking at symptoms and eliminating unlikely causes.
Proud owner of ShapeOko #709, eShapeOko #0, and of store.amberspyglass.co.uk
cvoinescu
 
Posts: 4442
Joined: Thu Jul 19, 2012 6:50 pm
Location: Camberley, UK

Re: Drift / Backlash

Postby selektha » Thu Aug 01, 2013 6:22 am

cvoinescu: Sorry for that, I read the question to fast and answered to fast :) luckily there are people on this community that are extremely (in a good way) fast in correcting me :P
selektha
 
Posts: 42
Joined: Tue Jul 02, 2013 2:05 pm

Re: Drift / Backlash

Postby cvoinescu » Thu Aug 01, 2013 10:54 am

selektha wrote:cvoinescu: Sorry for that, I read the question to fast and answered to fast :) luckily there are people on this community that are extremely (in a good way) fast in correcting me :P

:P
Proud owner of ShapeOko #709, eShapeOko #0, and of store.amberspyglass.co.uk
cvoinescu
 
Posts: 4442
Joined: Thu Jul 19, 2012 6:50 pm
Location: Camberley, UK

Re: Drift / Backlash

Postby teohes » Tue Aug 06, 2013 8:40 am

re: I've also found that, seemingly randomly, my Z axis will invert or start absolute valuing itself. Moving on the Z from -1 to 1 and back will work fine and then suddenly the axis will start ignoring the sign of the distance and will always move in the positive or negative. I don't think this is a physical issue, I'm going to see if it continues and if so see if anyone over on the TinyG Github has any thoughts.

I am not running TinyG ... just the standard shapeoko kit with the default electronics. My unit was working OK when all of a sudden, it developed the above mentioned random Z-axis behaviour. I swapped the steppers etc to no avail. and was on the verge of ordering a new driver kit. Finally it was JUST !!! a loose wiring to the Z motor !! Steppers are lovely when they work ... !@#$% when they don't.
teohes
 
Posts: 3
Joined: Mon Jul 15, 2013 10:12 am

Next

Return to Assembly

Who is online

Users browsing this forum: No registered users and 1 guest