~ 0.025 deviation in any length

Hi there!

I am very new here, but not only in EMC, also in English and CNC! :)

I have built an axis which is 400mm length, driven by trapezoid leadscrew 14x4mm, 1Nm stepper motor and an industrial motor driver. I have plastic nut (zero play), but if I say to EMC jog 1mm or 5mm or 10mm every travel shorter than cca. 0.025 - 0.050mm.

I thought that, maybe the pitch of the leadscrew less than 4mm, but than there have to be more difference after a few turns.

I know, tha 10mm is only two and a half turn, but the leadscrew pitch difference has to be added together, so if after 1mm there is 0.025, than after 2,5 turns the difference is 0,025 but not, it is still 0,0025.

It is like I have a nut which plays a bit on start and stop, but my one is not!

And an another observation:
It does no mather, how many mm moved away and back, the meter always get back on the zero dot!

I cant imagine the reason, maybe here is somebody, who was already played around this problem.

Thanks,
Alex

Quote:

---

Originally Posted by s_volenszki

Hi there!

I am very new here, but not only in EMC, also in English and CNC! :)

I have built an axis which is 400mm length, driven by trapezoid leadscrew 14x4mm, 1Nm stepper motor and an industrial motor driver. I have plastic nut (zero play), but if I say to EMC jog 1mm or 5mm or 10mm every travel shorter than cca. 0.025 - 0.050mm.

I thought that, maybe the pitch of the leadscrew less than 4mm, but than there have to be more difference after a few turns.

I know, tha 10mm is only two and a half turn, but the leadscrew pitch difference has to be added together, so if after 1mm there is 0.025, than after 2,5 turns the difference is 0,025 but not, it is still 0,0025.

It is like I have a nut which plays a bit on start and stop, but my one is not!

---

Alex,

I would say that the first thing is to find out is if the difference is precisely the same on every move 0.025mm or if it varies, that is for a 1mm move it is off 0.025mm but for a 10mm move it is off 0.050mm.

Is the error on every axis or only on one axis? Is it the same error on each axis or does the amount vary depending on which axis you measure?

Some "precision" leadscrews are guaranteed to have no more than 0.001" error in 12" (0.03937mm in 304.8mm). Many slightly less precision are rated at 0.003" in 12" (0.118mm in 304.8mm) and others the manufacturers don't even bother to tell you what the error is. So the error could actually be in the leadscrew.

Finally, you are talking about a very small error. It could be in the bearings on your leadscrews, the mounting of the nut or the flex of some part of your machine.

Alan

Dear Alan

Yes, the train of your thought is the same as my, and I will try to fix a 200mm long digital meter (scanning 0.01mm) on the axes to see the error on the full length.

This is a positioning system, it has only one axes, so I haven't got any comparison with another axes, built form the same leadscrew.

The point of interest is that, if I positioning the nut on the leadscrew 1mm than it has cca. 0.025-0.05mm error, when the move is 2mm, than the error rather 0.025mm and on one full tur (4mm) near nothing, and the same on 2 full turns (8mm) and on 2.5 turns (10mm).

If I set up the machine unit 400.25 instead of 400 wich is the default (1600step on one turn/4mm => 400), than in 4mm (so on one full turn) the axes absoluteley work out the fail, and on the next 1.5 turn it is not incrementing.

Like tolerance:
from start-up to 1 turn: ~0.025-0.05
after on full turn: ~: 0.005

I will try to measure the possible movment of the leadscrew in axial directioin. It can confirms that is the bearing causing the fail.

So thank you for the ideas, I will be back with the measurements!

Alex

Is there any freeplay in the lead screw mounting? (the fact that the error disappears after a longer movement sounds to me like backlass somewhere)

I will set up my measuring unit at the and of the leadscrew (I dismount the handweel) and measure at that place!

Thaks for your comment,
Alex

Hi Men!

I'm back and the cause has discovered! I tell the story:

Today I had some free time so I started to mount the metter at the end of the leadscrew. I measured only cca. 0,002mm in one full turn, so I thought the error is not about the backlash of the bearing.

After I mesured 100mm travel length, and the error was less than 0.01mm.

And than fortuitously I have got a sight of the clutch, and you won't belive, but the clutch in the stepper side is starting to revolve sooner than on the screw side!

Here is a video, to see what happens (look the fixing screws): [ame="http://video.google.com/videoplay?docid=1791872994763790927&hl=en"]http://video.google.com/videoplay?docid=1791872994763790927&hl=en[/ame]

So I think I need biger clutch (this one is 1Nm like the stepper) and no problem!

Thank you for the help and ideas,
Alex

Aha, well spotted :)

Looks like you need a stiffer (less compliant) coupling.

Glad you've found the problem.

Hi again!

I have new coupling (aluminum instead of plastic) and everything is all right!

This one:
http://farm4.static.flickr.com/3170/...ee8cdd.jpg?v=0

Bye