What's going on here then..?

Upgraded from 0.78d-dc42 to 0.96i-zpl and noticed that my prints were a bit distorted, seemed a bit short in the Y-axis. I noticed when my holes weren't coming out circular. First thought was my Y-belt had slipped, which it had a little, but after tightening, the problem persists.

I made a quick 20x20x2 box with a 5mm hole and set it printing. Calipers say X=20.01 and Y=19.65. I checked my steps/mm with an M92 command and got 87.4-something for both.

I tried 0.78za-dc42 and had the same issue. I tried rolling back to 0.78d-dc42 but I don't appear to have the 0.98 web interface files saved anywhere and 1.04 can't see the SD card contents or tell me the firmware version/display the config.g.

I've been using slic3r 1.1.7 before and after the firmware upgrade, but I tried slicing the test piece with 1.1.6 anyway, same result.

Can't think of anything else to check! Any pointers from you lovely people?

Many thanks!

Found the 0.98 web interface files so put them in the www folder and flashed 0.78d-dc42 again, still can't read the SD contents or report message log/firmware version etc. I can control the steppers though. What the...?

This behaviour is the reason I upgraded to 0.96i-zpl, I had just uploaded a gcode file and when the progress bar reached the end and the file list refreshed everything was gone... I figured it was time to upgrade anyway so when it started working again with 0.96i I thought nothing of it until now.

After deleting 50MB of gcode files from the SD the web interface can read it properly again with 0.78d. I guess I just had too much stuff on there for this particular firmware to cope with.

I've adjusted my Y steps/mm to make the print 20x20 in both X and Y but the holes are still rugby-ball-shaped, with the longest axis at 45 degrees between X and Y, so something else is at play here, I just can't think what! Failing stepper motor? Failing stepper driver on the Duet?

I've attached my test piece gcode in case there's anything glaringly obvious in there. If someone is feeling nice enough to print it (after pasting in their own start/end gcode I guess) they could verify if the skew is in the produced gcode rather than my printer. Any help gratefully received

for me that Gcode puts the model half way across the top edge of my bed. I can't print it.
Not sure how to help, but you can try to swap the stepper motors and see if the problem follows the stepper.
Switch the X and Y stepper and see if the out of round switches from the Y axis to the X axis. If so the stepper is bad.
If the problem stays the same then it may be the duet.
Swap the Y and the extruder steppers is another option.
We are just trying to find the broken item and hoping the problem will follow the part.

---

Socrates ~ The Amsterdamman
slic3r-1.2.9

Thanks for the suggestion. I've swapped out the X and Y steppers and the issue remains the same. All of the test pieces I've printed are consistently distorted. Could it be the driver?

I overlooked the fact my bed is a non-standard size so that gcode would be offset for most people. Apologies! I've resized it for a 200x200 bed and reattached.

I can't print that Gcode! It is WAY too fast for my machine, it didn't even stick. Assuming your machine runs the same speeds, I think your running very fast!
If you upload the STL I can slice it myself.

So the steppers don't seem to be the issue. Every hole size printed is distorted? Both printed flat or vertical along or across the X axis? Can you upload some photos?

---

Socrates ~ The Amsterdamman
slic3r-1.2.9

I have printed the 20x20 test and it looks fine. I Used your first gcode file, and used search and replace on the gcode file to subtract 100mm from the Y values. It over-extruder a bit but that would be down to different slic3r settings.

Edited 1 time(s). Last edit at 01/15/2015 07:08AM by MikeyD.

Thanks, I think that rules out the STL or slic3r being the problem.

I've attached an image of a few of the test pieces I've printed so far. The issue is more pronounced in some pieces than others, but the hole is always noticeably oval and measurement reveals the Y dimension to be short.

Additionally, I calculated the steps/mm I SHOULD have had to set to get the y-axis correct: 87x(19.65 / 20) = 88.55. The value I ended up setting (through trial and error) was something like 94.

I think I shall be contacting RRP next week to see if they have any insight. I really hope I don't need to replace my Duet...

Do the filament fall straight down if your extrude in the air? - if you zero the nozzle a bit lower than bed surface (make the nozzle build up pressure and burp), are the filament blobs of the same width on both side of the nozzles centerline?

Erik

@static65. Just a thought, have you carried out the Orthogonal Axis Compensation correctly? It is very important that the XYZ axis are all square to each other and compensated for if necessary.

---

appjaws - Core XYUV Duet Ethernet Duex5
firmware 3.1.1 Web Interface 3.1.1
Ormerod 1-converted to laser engraver, Duet wifi
OpenSCAD version 2020.07
slic3r-1.3.0, Simplify3D 4.1.2, Cura-4.4.1

As another suggestion, try printing the test piece extremely slowly (10mm/s) and see if the hole shape improves. This will show whether your axis are overshooting at speed. (In fact it is the acceleration that should be reduced, but reducing the speed is more intuitive).

Dave

App jaws may be on to something. Is it possible there are axis compensation entries in configuration that should not be there?
did you replace or change your configuration files with the firmware upgrades?

I haven't actually got any axis compensation, I spent a long time getting the thing damn near perfect when I built it. I need to check it's still lined up though as when I swapped the stepper motors over I had to move the printer about a bit which always makes me paranoid that something will have shifted. In any case, the problem is exactly the same before and after the motor swap so it wasn't that. I'm also using all the same config files as I was on the previous firmware. The printer has been running excellently for months up until this point, and nothing noticeable has changed about the hardware/config files which makes this super frustrating.

I'll try Erik's nozzle check and a 10mm/s print tomorrow (my usual speed is 40mm/s but I can whack it up to 80mm/s if I'm just drafting something) and let you know the result. Thanks all for the help!

Oh, something I just thought of! Is it possible to change which stepper driver is assigned to which axis? That way I could swap the X and Y cables over and if the problem shifts to the X axis it could indicate a duet fault?

Finally got some time to try printing it at 10mm/s. The result is the same but perhaps a bit less obvious. There's still 0.2mm difference between the X and Y dimensions.

Would increasing motor currents help? I already bumped them up during the build (X1000 Y1000 Z800 E1000). I'm not experiencing any skipped steps though. What's the current limit?

Does the firmware upload process blank the Duet's memory completely or is there a tool I can use to do that? Might be a rogue value somewhere?

Clutching at straws here :/ Thanks for all the help and suggestions so far!

@static65, you never answered my comment "Have you carried out the Orthogonal Axis Compensation correctly?" From what I've read in this thread, this is probably your problem and could do with another check.

---

appjaws - Core XYUV Duet Ethernet Duex5
firmware 3.1.1 Web Interface 3.1.1
Ormerod 1-converted to laser engraver, Duet wifi
OpenSCAD version 2020.07
slic3r-1.3.0, Simplify3D 4.1.2, Cura-4.4.1

No, increasing motor currents would not help. So long as the motors never skip a step (which would have a different symptom), your currents are high enough and increasing them will make no difference.

Dave

@appjaws1, I did say I haven't got any orthogonal axis compensation. I've checked and rechecked with a set square and it's still as straight as the day I built it, but I will print the axis compensation piece again to be sure.

Printed the orthogonal measurement pieces. There was some minor XZ compensation to be had which will benefit tall prints (I have a silly 825mm tall build volume). Unfortunately it hasn't fixed the distortion issues

Quote
static65
Oh, something I just thought of! Is it possible to change which stepper driver is assigned to which axis? That way I could swap the X and Y cables over and if the problem shifts to the X axis it could indicate a duet fault?

I would expect the path of each axis to be hardwired through the PCB, see no reason why it shouldn't, but not something I know of - but at least you got a bump on that Q

Erik

This may be a really dumb question, but when last did you oil the smooth steel rods on your Ormerod? I had endless problems in the Z-direction, and after spending hours with polishing it with steel wool to no avail, 2 drops of 3-in-1 oil resolved my problems completely. Perhaps your Y direction rods are binding slightly, just enough to cause a problem?

Quote
pandionx
This may be a really dumb question, but when last did you oil the smooth steel rods on your Ormerod? I had endless problems in the Z-direction, and after spending hours with polishing it with steel wool to no avail, 2 drops of 3-in-1 oil resolved my problems completely. Perhaps your Y direction rods are binding slightly, just enough to cause a problem?

Have to second this - I started with insufficient oil on the smooth rods and Z threaded rod. Will cause many issues.

---

Socrates ~ The Amsterdamman
slic3r-1.2.9

Quote
pandionx
This may be a really dumb question, but when last did you oil the smooth steel rods on your Ormerod?...

That is a very valid question!, the force needed to move the rubber seal of the bearing on a oiled vs dry smooth rod is huge - when changing the direction, the seal will bite the smooth rod and have to expand to move from digging to dragging mode, if well oiled, the seal will not have to this "flip flop" action and just float on the oil

Edit: (not very clear) pictures of flip-flopping seal on dry smooth rod



Erik

Edited 1 time(s). Last edit at 01/20/2015 03:37PM by ormerod168.

A fine suggestion! It was probably due an oiling anyway as I haven't done it since I built the extended axes. Hasn't solved the issue though!

Disappointingly, a couple of the seals broke on the linear bearings within a couple of weeks after building the original kit, one on the X axis and one on the Y. Perhaps I should replace these as if debris has been getting in there no amount of oil will unclog it. If that doesn't solve my problem I don't know what will