Bizarre Marlin 2.1.1 Behavior


I have a TAZ 4 that I upgraded with a BMG-X2-M (only one hotend on it at the moment), auto bed leveling using a limit switch, and Marlin 2.1.1. This benchy is the first print I let the printer complete just to showcase the problem I’m having. It seems like there are layer shifts but only on part of a layer. Also, thee infill ( just a grid pattern) has parts of it that are curved and others like normal

If I print from an SD card instead of a USB cable to the computer I get the first layer to print normally, and then the rest are the same as the photos. The layer preview on Cura 5.1 shows everything normally, so my theory at the moment is that Marlin 2 is just a bit much for the 8bit Rambo board. I’m going to configure version 1.1.9 and see if that works.

If anyone is experiencing anything similar or knows how to solve this, please let me know.



If I’m not mistaken, the 4, 5, and 6 all use a Rambo board with the ATMEGA 2560 chip, and I’ve had zero issues like that with either of the 6s I use with 2.1.1.13

You say you’re running Cura 5.1 to slice, and without a native profile for the Taz 4, you’re leaving a lot of possibilities for an improper profile to be responsible. First thing to do is to use Cura LE with the latest firmware on your machine and do a test. Yes, the hot end doesn’t match the original, but with a 4, and manual leveling I don’t think it’ll be an issue. If you haven’t already, run a PID autotune and input the verify your e-steps on the machine. Check that Cura LE has the correct filament diameter and give it a test.

All the layer shifts appear to be in one axis. That motor (or its driver on the board) could be on its way out. If you’re handy with recompiling the firmware, you could swap drivers with that axis and the second extruder’s, and see if that corrects it.

Thanks for the tips, I’ll try it out today and report my results.

Although I doubt that Cura 5.1 is the issue. The only machine specific items that the slicer cares about are build volume, filament diameter, start/end g-code, printhead offsets, and firmware type. Also, I’ve used the Cura 5.1 profile before with the last version of stock Taz 5 firmware (Marlin 1. Something) and everything was working great with my E3D volcano.

The X-axis driver going bad does seem like a possibility.

Alright, I tested the stepper driver by just swapping the pin assignments between the x-axis and e1 steppers in the firmware and swapping the cables on the mainboard. This produced the same results, which tells me that the driver is not the issue, so I swapped everything back to normal.

The next thing I tried was going back to Cura LE and making a profile for my printer. At first, it seemed like everything was going normally, but about 15mm in, it started doing the same exact thing. (weird partial layer shifts)

I have a raspberry pi 3 B+ lying around, so I’m going to try and get klipper up and running, which will remove any and all computation that the Rambo has to do. Hopefully, that will fix my issue If going back to Marlin 1.1.9 doesn’t work.

I configured Marlin 1.1.9 and added the thermistor table and related items for the Slice Engineering 450C thermistor.
Unfortunately no change, the problem persists.

Tomorrow I’m going to go back to stock TAZ 5 firmware and put the E3D Volcano back on.

If the problem remains, I’ll get a new board.

If not, then I have no idea.

With the deletion, I guess that didn’t solve it?

It was partially right, turns out the belt pulley was a bit loose. I deleted the post because it was stupid of me not to check that in the first place.

1 Like