Marlin sometimes does not move motor for small steps

Hello,

I have RostockMini which is running perfectly fine using repetier FW. I now want to switch to Marlin to be able to use FSR sensor for automatic bed leveling. Everything is working beside doing small steps.
Example:
G28; Motors are homing
M114; Just to make me and RepetierHost happy
G0 Z100; Motor moves
G0 Z100.1; Motors doesn't move
G0 Z100.2; Motor sometimes doesn't move
G0 Z100; Motor moves

I did above scenario with RepetierHost and minicom. Same result. Marlin FW (deltabot and fsr) randomly refuses to move motors when doing small steps equally which axis!

I've attached the Repetier as well as the Marlin Configuration.h

Anyone has a clue what I'm doing wrong?

Forgot to mention that if I read out the position using M114 Marlin tells me that it's at the position it's supposed to be (but unfortunately isn't)

If you want to move less that ~0.1mm you can use G1 instead of G0

AFAIK on most CNC machines G0 is used for fast moves, to get to a position as fast as possible.
[gnipsel.com]

Thanks for you answer. However, my problem is not the movement itself but the fact that the printer in approx. 1/3 cases not moves when using Marlin FW and doing small steps. Equally which axis involved.

Edited 1 time(s). Last edit at 03/31/2014 06:11AM by RostockMini4tw.

How are you determining you're not moving. Moving only a tenth of a millimeter is only a few steps and may not be discernable. I'm not saying that it is moving, I'm just saying it might be tough to determine.

First of all I sometimes do not hear the motors moving when doing very small steps. If I press step 0.1mm in any direction I can hear them moving. With Marlin this is not always the case.
I first noticed it when trying to calibrate my Rostock Mini with a sheet of paper. Calibrating was not possible.
I then did more test. The most simple scenario was

1. G28 maximum z-height is set the same in both Configuration.h
2. Move the head to x0y0z1
3. Use step 0.1mm down in Repetier Host software 10 times
4. measure the difference to the printing bed

With Repetier FW this works fine. With Marlin I always end up a bit above the printing bed.

Edited 1 time(s). Last edit at 03/31/2014 08:02PM by RostockMini4tw.

Anyone else got an idea? I've the wonderful force sensor laying around but can only try them if Marlin firmware runs stable.

Seems like it might be a weird rounding problem--there's an awful lot of sqrt(sq()) rounding oportunities in the Delta code.

Can you measure the actual movement with a dial indicator?

With the 44.4444 steps/mm in your file, it looks like a 0.1mm should be 4.4444 steps per motion. I wonder what you'd see with a 100 steps of 0.01mm.

Sorry as I'm out for a business trip until end of the month I won't be able to do this. I will do when I'm back.

Sorry but it took me some time to come back to this topic. Anyway, here is what I did:
- Move the nozzle down in 0.1 steps to one object printed earlier until z_min (FSR sensor attached to bed) is triggered.
- G91
- F100
- 100 times G1 Z0.1
- 95 times G1 Z-0.1
As before, and in contrast to repetier fw, I don't hear the motors move for each step (if done manually ofc)
Then I counted the steps until z_min is triggered again. There were five. I repeated the above 5 times every time with the same result.
I repeated the above again using 1mm steps. Same result. I always end up with the sensor triggered after the correct steps.

Next test I did:
- Move the nozzle down on an object with known height until z_min is triggered. M114 reports z 25mm
- Removed the object
- Counter the steps 0.1 needed until nozzle hits printing bed. I used a caliper to measure the object's height. Its a bit smaller than 20mm (approx 0.1, all I can say with my poor tools is that it is smaller). It takes 201 steps until z_min is triggered again. M114 reports Z 4,9
I repeated the above three times. Every time with the same result.

Summary: In total the sum of the error is negligible. I just fear when printing with 0.1mm layer that they get mixed-up.

Just in case someone finds this topic with a similar problem. Yes, Marlin firmware seems to have a rounding problem with my set-up. I got GT2-16 timing pulley now and now motors are moved for every 0.1mm step.