Bad bearing? Sometimes I’ve had a bearing grind in one direction but not the other…also, which printer are you running? Sounds like Ultimaker-like, with a descending print bed.
I guess I would throw a different mega at it to see if the problem persists. I’ve had strange things happen when my mega got too hot.
I think awhile back someone was trying to figure out a strange problem and it ended up being that he needed shielded wires for step and direction to remote stepper drivers. Does anyone else remember this?
Remote stepper drivers is quite uncommon.
I had problems with my stepper drivers overheating, they act odd when they do. Put some heatsinks and a big fan on it and things got better.
maKES TWO OF US. 3 weeks after we PAY extra for an assembled machine one of the two power resistors that heat the extruder burns out. Tech support sends directions on how to repair the extruder and offers to sell me the needed parts.
it sounds like it can be too high acceleration/initial speed.
@Ross_Hendrickson I’ve only tested going down, because that’s the only direction that counts as I don’t use auto bed leveling. I’ll look in to that.
@Francis_Lee In that case there would be a difference between the three Z-motors, wouldn’t it? There is none.
@Josef_Jelinek Yes, could be, but I’ve been there, done that.
It doesn’t matter. `Extremely slow to extremely fast, it stays the same.
@Daniel_Kruger At the moment the Z axis is controlled straight from the board with 2 Pololu’s, to see if the problem came from the external stepper. It wasn’t.
Another test you could try is small movements that equate to whole step values. I am wondering if, for some reason, when the motor stops after performing a microstepped move it resets to the nearest whole step position.
If this is the case the only thing I could think would be the cause is insufficient motor holding current. Is your power supply within specification?
@Neil_Darlow The power supply should be no problem. I’ve checked it while running, and the output is very stable.
I’m thinking that it can’t be something like that, because in that case I should see a bigger deviation with bigger movements, wouldn’t I? And it’s the other way around. I get erratic results only with small movements.
@Daniel_Kruger I’m thinking in that direction, trying another Mega. But I don’t have one to try. And buying one of these old 8-bit boards feels like throwing money away. I’d sooner go for a 32-bit solution. A cheap asmz-mini costs almost as much as a original Mega board.
@Rien_Stouten , I never had the same problem, but seems to me that your stepper skips steps on microstepping, surely 3steppers should be adequate, but if the load on the motor becomes too big, it doesn’t matter how much power you try to feed it. Beyond a certain limit(max torque setting)motors can no longer turns motions become erratic, severely compromising accuracy…which might what be happening to your z axis (pwm is not on max duty cycle and the sorts…since babysteps/microsteps doesn’t deliver the same amount of power as fullsteps)just a thought there
During printing you will get lots of small movements so my hypothesis is more likely to happen. Your Z moves will be at layer height which is small in comparison to the larger manual moves you described. Lots of small cumulative errors might account for the sizing error.
I want to thank you all for your input. It gave me the motivation to solve it.
As it turns out, it’s wobble.
One of the rods had a very small bend. And with my setup, two rods on the corners in front, and one in the middle in the back, it gave the seemingly erratic movements.
I’ve replaced the rods for some I had lying around, and although it is far from perfect, since 8X8 threaded rods will probably burn my small z-motors up, the difference is enormous. I can now move the bed with exact 0,1 mm steps. So at least I know what to do now.