[ajayre]

I got my CNC machine kit in 2008 and used it until 2010 without problems. No anti-backlash nuts but backlash on the long axis (Y) was 0.13mm and I used EMC2 software compensation to reduce it to 0.02mm.

I then packed my machine up until 2015 before reassembling it again. Now after reassembly I have backlash of 0.24mm on the Y axis. I tried EMC2 software compensation but it didn't make a difference. I expected a small improvement at least.

I disassembled the gantry and the leadnut had a noticeable amount of play on the screw (ACME 3/8-10-2).

I removed the leadnut and installed a brand new DumpsterCNC anti-backlash nut. Now there is no noticeable play on the screw.

But I am dumbfounded because the backlash is still at 0.24mm!

So now I think I must be making a fundamental error somewhere, but I don't know what.

I am using 1/4 microstepping. I measured the backlash with a feedrate of 100mm/min. I am using a digital dial indicator for the measurement with a commanded distance of 0.04mm. The procedure is:

- step into the dial indicator
- zero the dial indicator
- step 10 times into the dial indicator 0.04mm a time
- step 10 times away from the dial indicator 0.04mm a time
- look at the reading on the dial indicator (0.24mm)

This is consistent and reproducible. I can see and feel that the first five steps or so away from the dial indicator the leadscrew moves but the gantry does not.

Any ideas? Thanks!

[691175002]

Well something somewhere must be loose.

I'd put an indicator on the end of the leadscrew to see if it has any axial play. If the bearings/couplers are loose the screw could be moving back and forth instead of the gantry.

Assuming the screw is properly mounted the issue has to be the connection between the nut and the carriage. There aren't really any shortcuts here, you need to measure and check everything until you find the problem.

[ger21]

How much weight are you moving? Those plastic anti backlash nuts are really only good for about 20-25lbs of force.

[LazyMan]

To start with, I would test using full steps only. What kind of couplers are you using? They can flex and have slop and, as gerry said, if its a heavy load, the nuts will flex. Everything else will flex too.

[Devastator]

With stepper power off, turn the leadscrew by hand back and forth and watch where the leadscrew goes into the nut and also make sure the nut is firmly attached. You should be able to feel and see if there is any backlash by rocking the leadscrew back and forth. After you've checked that, move to all of your mounting points, like where you nut attaches, and then your stepper shaft coupler.

[louieatienza]

First I'd make sure your lm bearings run smooth. Sounds like your coupler however. If a spider coupling, make sure the spider is tight, or use one with higher durometer rating. Helical couplers can wind up and cause lash. Bellows or Oldham type couplers may be a better choice.

[ajayre]

I made a bit more progress with this.

There is no movement along the screw axis during operation - zero.

If I try to jiggle the gantry by hand there is zero slop.

The grub screws are in tight in the solid brass coupler.

If I take the gantry off and just attach the gantry base then the backlash is 0.04mm. This is the same regardless of whether I attach the gantry base to the sliders or not.

As soon as I put the rest of the gantry on the gantry base the backlash increases to 0.24mm, so weight is a factor. The weight of the gantry hasn't changed since I got the kit years ago and backlash was OK back then.

If I jog 0.2mm at a time then I can can visually see when I reverse direction that the screw moves but the gantry does not for the first few jogs. I am using a two start, five turns per inch screw. By my reckoning that means a backlash of 0.24mm equals a turn on the screw of 17 degrees. I am positive that when I reverse direction the screw always moves. I think a stall of 17 degrees would be quite noticeable. So I don' think it is the coupler.

The gantry is attached to the leadscrew using two wood screws. I have these in as tight as possible. My current thinking is that the repeated removing and installing of the wood screws has created just enough slop that it only becomes apparent when the weight is increased.

I am thinking of drilling through the holes and replacing the wood screws with M3 bolts and nylock nuts, but once I do that there is no going back.

[wizard]

I made a bit more progress with this.

There is no movement along the screw axis during operation - zero.

If I try to jiggle the gantry by hand there is zero slop.

You might not be able to feel this. It isn't like slamming something up against a hard stop. For problems like this a dial indicator can be very useful. Obviously you already have made measurements I just wonder if the setup is correct as you should Measure something.

The grub screws are in tight in the solid brass coupler.

If I take the gantry off and just attach the gantry base then the backlash is 0.04mm. This is the same regardless of whether I attach the gantry base to the sliders or not.

I'm trying to imagine what this even means. Maybe a picture might help. By the way what type of bearings are you using for linear motion? If they are round bearings do realize that you can loose balls out of them effectively adding slop. You should be able to feel and see this when checking out the machine though. Actually you can loose balls out of most linear bearings.

As soon as I put the rest of the gantry on the gantry base the backlash increases to 0.24mm, so weight is a factor. The weight of the gantry hasn't changed since I got the kit years ago and backlash was OK back then.

Have you visually inspected the plastic nuts?

If I jog 0.2mm at a time then I can can visually see when I reverse direction that the screw moves but the gantry does not for the first few jogs.

If the screw is turning and you aren't getting motion then you have isolated the area of concern significantly. You will want to look at the leadscrew nuts and the thrust bearings. I'd especially look at the nuts due to this problem being load dependent.

I am using a two start, five turns per inch screw. By my reckoning that means a backlash of 0.24mm equals a turn on the screw of 17 degrees. I am positive that when I reverse direction the screw always moves. I think a stall of 17 degrees would be quite noticeable. So I don' think it is the coupler.

Sometimes it helps to mark things up to see what is turning and if relation ships between part change. Sometimes a sharpie works but I've also have used old fashion typewriter correctin fluid. You could also use paint markers. One of our extractor manufactures from Japan marks every single screw, nut, bolt, coupling and just about everything else with the potential to move with marking paint. If anything moves you know instantly.

For something like a coupling, registration marks can highlight a coupling spinning on a shaft. Couplings can do this even if the set screws are tight. An index mark on a leadscrew would allow you to verify that the screw is turning as much as you think it is.

By the way I did not verify your calculations but your backlash is the sum of all of the backlashes in all of the mechanical components in the drive chain. Some of those backlashes have existed since day one. In all likelihood the part with the most backlash, has backlash less than the 0.24 mm.

The gantry is attached to the leadscrew using two wood screws. I have these in as tight as possible. My current thinking is that the repeated removing and installing of the wood screws has created just enough slop that it only becomes apparent when the weight is increased.

You should be able to see movement though. Using your own calculations the nuts would have to rotate a considerable amount.

I am thinking of drilling through the holes and replacing the wood screws with M3 bolts and nylock nuts, but once I do that there is no going back.

Don't make changes until you have satisfied yourself to the point that you are certain you are correcting a real problem. Either your nuts are mounted tightly or they aren't . Verify that they are indeed rocking back and forth before making changes.

In any event if I was working on this problem, with limited access to tools, I'd mark everything up to try to get a handle on where I'm loosing movement. Your observation that the leadscrew is indeed rotating is good as it can narrow things down for you. Just don't get hung up on the idea that it is rotation enough or isn't moving axially.

Sometimes you have to scrap your previous conclusions about what is or isn't wrong with a machine. I know I often get hung up this way at work after a long shift.

[Devastator]

It sounds like you don't have enough preload on the leadscrew. You can test this by grabbing the gantry and tugging it back and forth. If you get a clunk, clunk, add more preload until it doesn't do that.

[ajayre]

Thanks for all the replies.

If I tug on the gantry there is no clunking, no movement.

I see that I made a mistake in my OP. The backlash after reassembly this year was 0.465mm and adding anti-backlash nuts reduced it to 0.24mm.

I replaced the wood screws with M3 bolts and mylock nuts and the backlash reduced to 0.15mm. I've used software compensation to reduce it down to 0.04mm. I think that is probably the best I can get out of my machine so I will see how it goes.