Y-AXIS drifting / machine not usable for PNP

[S G.]

Hello,

I get a 1-3mm drift on the Y-Axis after 30 movements or so, which makes the machine totally unusuable.

i had to built 2 machines to get away from the Y-axis drift but still no luck.

My second machine is built with Marlin, reprap board, rails , 3M belts, nema23 motors and tb6600 drivers.

all calibrations were made accurately.

I even use the i2c position encoders function for closed loop motor driving with an 180LPI encoder strip attached to the Y-axis and the H9730 optical eye.

I am out of ideas and really need help how to solve that issue.

Thanks.

Sly

[JW]

If you're confident in the closed loop control of your steppers, it must be mechanical.

This thread is nearly identical, turned out to be a loose set screw on a belt pulley.

[S G.]

Thanks for the link.

I have reviewed all the necessary settings, but the drifts still occur.

Interestingly, I tested the axis on Repetier, moving it forward and backward for 10 minutes. Not a single drift occurred, even without the encoder. This suggests that the issue might be related to OpenPnP, possibly something in the motion control settings.

Additionally, I noticed that selecting constant acceleration or other options does not affect the movements.

The firmware details are as follows:

• Firmware Name: Marlin bugfix-1.1.x (GitHub)
• Source Code URL: Marlin Firmware on GitHub
• Protocol Version: 1.0
• Machine Type: OpenPnP
• Extruder Count: 1
• UUID: cede2a2f-41a2-4748-9b12-c55c62f367ff

It seems there might be a compatibility issue between OpenPnP2 and this firmware.

Any suggestions?

[Mike Menci]

Do you have the same settings in configuration file  with compare to OpenPnP - Axis settings? 
https://github.com/openpnp/openpnp/wiki/Machine-Axes#referencecontrolleraxis

[mark maker]

> Thanks for the link.

> I have reviewed all the necessary settings, but the drifts still occur.

The link's solution was not about settings. Have you checked the set-screws, if you have any?

_Mark

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/8b1df0a0-bdbc-4f62-816b-339a5d2bd82cn%40googlegroups.com.

[S G.]

@Mike: Yes all settings are the same. For HTD3 belt with 18t Pulley the steps per mm is 59.259259. Set in openpnp Axes and in Marlin Configuration.h file.

@Mark:   Yes i checked the set-screws on Pulleys etc. Everything is fine. It is not a mechanical issue.

[bert shivaan]

So a couple of things you will need to understand. OpenPNP is mature. Meaning there is a HUGE user base running it everyday without any drift. The reason I point this out is not to tell you there are no bugs, as there very well may be, but more so to help you remember that is not the actual issue most likely.

Having said that, there is clearly something going on since it works with Repetier but not OpenPNP.

My guess is it is something in the accels or jerk settings.

Or there could be some accumulated error from the odd step size. but that should not make it drift. It would just be wrong all the time, but the same amount wrong.

Keep in mind the way this works is really quite simple. OpenPNP tells your controller where to move to. It has no idea if it moved there or not. If you post your log file we can see if OpenPNP is telling the control to move to the correct place every time or if it is telling it to move somewhere different causing the drift (I really doubt this, but keeping all minds open)

Assuming OpenPNP is sending the correct commands, the question becomes why does the machine/control not repeat when OpenPNP is sending the commands.

Maybe it is not completing the moves when it reverses?

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/d325467f-8aac-4ea5-939a-772d01695af0n%40googlegroups.com.

[S G.]

@cncmachineguy:

Apologies for mentioning Repetier. I did some high-speed tests with Repetier, and the issue persists.

Regarding the jerk settings: Jerk is set to 0.

As you noted, the moves do not complete when the axis reverses, causing it to drift on the next move.

Do you have any advice on how to resolve this?

Thanks.

[bert shivaan]

I am not sure this will be an issue when you actually start to use the machine. There are not too many cases I can think of where the axis just oscillates back and forth without doing something at the end of the move. You do need to check if there is exact position mode in Marlin.

In LinuxCNC this is G61. Basically it tells how close the position has to be to call it complete. The point is to allow it to round corners for accel/decel reasons. In practice, we hurry to a spot in order to do something - pick/place/vision which means we must make that position.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/f7b95922-d188-472d-af04-f62179b63f9bn%40googlegroups.com.

[S G.]

I believe the issue is related to the Marlin 1.1.x firmware (MarlinOnRamps4Openpnp)  . After flashing the Marlin 2.1 firmware  from crashmatt and making the necessary changes in the configuration, my M119 command no longer shows the X and Y endstops. The axes move, but I cannot send them to home. :(

On a positive note, the drift problem disappeared with this firmware.

Endstops are set as usual with:

#define USE_XMIN_PLUG
#define USE_YMIN_PLUG

#define USE_ZMIN_PLUG

Has anyone else experienced this issue or have any suggestions?

[S G.]

Additionally, if the problem with Marlin persists, I am considering switching to another board instead of RAMPS. Can anyone recommend a board that offers precision and speed, with minimal configuration required? I'd prefer a solution where I can just flash the firmware, make basic settings, and start running the PNP. Any advice would be greatly appreciate.

---

I noticed that the OpenPnP wiki mentions RepRapFirmware. The duet3 is a bit expensive. What If I purchase an SKR Pro v1.2, can I just flash the firmware and have it work, or are there many settings to configure?  Or is the configuration easier and more straightforward with the Duet3?

Thanks for your time & help.

[mark maker]

> The duet3 is a bit expensive. What If I purchase an SKR Pro v1.2,

I vote for Duet. Quality has a fair price.

But before you buy a new board are you absolutely sure you don't have lost steps/closed loop fails?

What happens when you set the Speed % to half or quarter speed. Still drifting?

Note, the essential thing is to try with lower acceleration, so before Speed % can effectively work, be sure to have followed Issues & Solutions and accepted the ModeratedConstantAcceleration motion control type, and accepted all the G-code that can actually send it.

Ask back if that is not clear.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/6bc4db39-6d6a-4172-850d-c57936c52c81n%40googlegroups.com.

[S G.]

Thanks for your advice Mark.

Do you use a Duet in your machine?

I really did everything to check if its hardware realted or not. I tightened all the set screws, glued even pulleys to the shafts that nothing slips without my knowledge.

For the last step, i have to get a new board to be sure if its Marlin settings or something else. At the moment i just have the Ramps board and only a working Marlin 1.1.x. So before i order another unreliable board i decided to ask here for your opinions.

I will  try tonight the axes movements for the last time with a simple Arduino GRBL setup to check what's causing the drifts.

[Krzysztof Kozłowski]

Have you checked onboard IC’s temperatures? 

I never had this kind of problem in pnp machine, but I had similiar problem when machine was loosing steps because of overheating stepper motor drivers.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/377a4071-5070-4d71-b519-06409631cacen%40googlegroups.com.

[Krzysztof Kozłowski]

That problem was in DIY 3D printer when I had controller completely enclosed without air movement.  

[S G.]

I finally configured Marlin 2.1.x to recognize the X and Y endstops. After testing for 10 minutes, there are no more drifts on the X and Y axes. The issue was due to Marlin 1.1.x settings causing the drifts, though I'm not sure which specific settings were wrong. Now, I can finally use my machine for Pick and Place with OPENPNP. Thanks to all the supporters. Cheers!