cnc router is out of alignment after a few movements

[vincent-visser]

Hello everyone,

A few months ago we started building our own cnc router with 1,3metre by 2,8metre workspace. Driven bij NEMA34 closed loop stepper and 3:1 gear reduction and rack&pinion mod2.

Machine is almost ready and we've done some testruns, we found out that the machine is not square after a job of a few minutes, i mean we measure a differance of 2-3 centimetre between y-axis and slave y-axis, and the Zero(home) of the X-axis has moved 2-3 centimetres behind de homing sensor.
First we thought we might had a fault in the Overall Turn Ratio but if we reset XYzero and then move 1200mm we measure 1199.5 millimetres so thats already really close en we don't think this causes our bad alignment after a job of 5 minutes.
Our current OTR is 15t pinion which has 30mm pitch-circle so together with 3:1 reduction that is (30*pi)/3= 31,41592 millimetre. Steps per Rev. is all set at 3200

Something strange we've noticed is when we run a job en the trail is on in graph mode 5 a 6 white dots are equal spaced and then 2 are very close to each other, with a constant feedrate of 2500mm/min, but samething happens with different speeds. Maybe something is wrong with the encoders of the steppers? isn't thats strange because is it happens with als the motors while they are brand new. See picture added.

hope someone can help us a little bit further.

Thanks
Vincent

[dpascucci_support]

Please also attach a fresh report for your system

[kb58]

In your post above it's not clear when the alignment was done in relation to when the gantry was measured as being out of alignment. Have you performed a fresh gantry alignment, run a job, then measured alignment again before powering down? Reason I ask is because I ran into this when I was using the simple homing method of "drive until any limit switch trips." The problem is that when servos or steppers are first powered up, they may or may not assume the exact same position as when powered down. This lead to higher and higher holding torque (leading to overheating) because the two now-misaligned servos may fight each other even though the gantry wasn't moving. The solution for me was to switch to the dynamic squaring/homing method, and all my servo issues went away.

[vincent-visser]

Please also attach a fresh report for your system

I attached a fresh report

[cnckeith]

vincent, disgard the graph that has nothing to do with out of sqaure( its your CNC PC CPU not keeping up with the graphics)

[cnckeith]

what gantry squaring method are you using?

1.) Manual squaring
2.) auto square with software paired
3.) auto square with hardware pairing

[cnckeith]

what gantry squaring method are you using?

1.) Manual squaring
2.) auto square with software paired
3.) auto square with hardware pairing

and post a link to a photo album of the machine so we all can see your setup.

[vincent-visser]

Hi Keith,

Thanks for your reply,
We use the auto-square with software paired axis. We have paired axis 4(Y-slave) to axis 2(Y master).
Here is an wetransfer link to some photos, good to know, all the wooden parts you see are temporary and will be replaced bij aluminium parts by the time the cnc can make themself. https://we.tl/t-f9bGyZkINb

One thing i noticed today was that the encoder is not back at 0 at the moment the axis is back at the homing-sensor and in this case that means the X axis is going further back then forward. We’ve mounted the x home sensor on the left of the gantry en an positive travel 0 to 1304mm. After a job of 5minutes we move the X back tot home sensor and then we still have a few millimetres left on screen at the moment the red light of the proximity sensor goes off.
I hope you understand what I mean haha.

Problem goes for Xaxis either as the y-master and y-slave. Those are all preach and pinion driven. Z-axis looks good for now but we’re not 100% clear about that one. Z-axis is ball screw.


We life in the Netherlands so sometimes is kinda hard to explain technical problems like this

Thanks!

[ShawnM]

Your photo link didn't work for me. Can you start a Google photo album or correct this one?

If after you home the machine and then jog the X axis back and forth and then jog it back to X0 it should never reach your proximity sensor because of your soft limit settings. The proximity switches you have setup to "homeall" should never trip after you home the machine with soft limits setup as yours are.

Did I understand your response correctly? You have a cnc machine made of wood? If you have a machine made of wood I'm not sure how you expect it to stay square or rigid.

[vincent-visser]

Your photo link didn't work for me. Can you start a Google photo album or correct this one?

If after you home the machine and then jog the X axis back and forth and then jog it back to X0 it should never reach your proximity sensor because of your soft limit settings. The proximity switches you have setup to "homeall" should never trip after you home the machine with soft limits setup as yours are.

Did I understand your response correctly? You have a cnc machine made of wood? If you have a machine made of wood I'm not sure how you expect it to stay square or rigid.

hope this link will work https://photos.app.goo.gl/jijctKskUTmWoee57
As you can see we have a steel frame combinated with aluminium extrusion profiles, but the connections between alu extrusion are temporary made out of wood until machine works fine and it can make his own parts out of aluminium.
proximity sensor Z X and Y-master are connected to homeAll on input1 en Y-slave is connected to SlavedHomeInput on Input 2.

We use Lichuan LCDA86H drivers and configured the Axis motor drive type settings in the Wizard as Leadshine DM Series, is that correct?