Z movement accuracy (2.4motion system) debate

Maurici · March 19, 2023

So, my functional OCD is back... and isn´t good.

I´ve been printing quite a lot with the micron+ (the shrinked 2.4).

playing with the extrusion factor and a bit of pressur advance and controling speeds, and I have a 0.05 accuracy to my X and Y axis.

However, I seem to be sistematically 0.1/0.15 off in the z axis no matter what I do. no matter what layer heigh do I use, what surface... what z-offset or how much do I squish the first layer...

So... I´ve started measuring.

(*Note, my DTI tool is far from a metrology certified instrument... is used to check valve clearances, setup camshaft lift and check planicity to brake disks where a 0.01mm accuracy onto a backyard build race engines is somewhat a dream rather than a need, and for purist metrologists, I know the setup is crap, I know I should have put both the DTI and the printer onto the same solid surface, probably removed the foam under the feet... i´ve made a judjament call to simplify the workload and I got plenty consistent results).

If I trust however the results, she moves spot on 0.01 deviation per each 5 of 10mm regardless of how much do i move the Z axis. (in fact if anything moves 0.01 less than too much, however, irrelevant vs the 0.1/0.15 deviation I´m getting.

I can´t do longer than 10mm movements as by the time I´ve preloaded the arm to avoid desplacement and so I´ve already eaten north of 10 of the 25mm of available travel in the DTI.

I´ve tested this setup in all 4 corners, and indeed all movements were spot on with a tendency (that I could easily blame to the measuring setup) of -0.01

Has someone made any Z accuracy measurements?

Failing the above... if you are having greater accuracy than me... any thoughts on what it could be? or what could I try?.

Edited March 19, 2023 by Maurici

SuperBoppy · March 22, 2023

I'm thinking, if your measurements are consistent, that the deviation could be the resolution of the stepper motors. Have you calculated the 1/200 of a rotation resolution (I think it's 200 steps, it may be something different) of the stepper motor and what that would be when translated through the pulleys and finally to the actual motion of the gantry? That may where the precision issues lie. I'm a little OCD myself, but I wouldn't loose sleep over 10 microns, when the smallest practical layer height is 50 microns.

Maurici · March 22, 2023

1 hour ago, SuperBoppy said:

I'm thinking, if your measurements are consistent, that the deviation could be the resolution of the stepper motors. Have you calculated the 1/200 of a rotation resolution (I think it's 200 steps, it may be something different) of the stepper motor and what that would be when translated through the pulleys and finally to the actual motion of the gantry? That may where the precision issues lie. I'm a little OCD myself, but I wouldn't loose sleep over 10 microns, when the smallest practical layer height is 50 microns.

Worth a check.

I dont lose the sleep over the 10 microns. They are plenty good... i lose the sleep over the 1 or 1.5 tenths im sistematically gaining in z axis...

ken226 · March 22, 2023

The coefficient of thermal expansion for PLA is about 68 microns per meter, per degree Celsius.

A 1cm tall PLA part would see a .13mm change in height after it cools from 220c to 20c.

Could this be some, or all, of what your seeing?

Edited March 22, 2023 by ken226

SuperBoppy · March 22, 2023

If there is a consistent accumulation of error as the z-axis position is increased, perhaps this could be related to the calibration value for the stepper motor in the config file? Just thinking out loud...

Demosth · March 23, 2023

19 hours ago, SuperBoppy said:

If there is a consistent accumulation of error as the z-axis position is increased, perhaps this could be related to the calibration value for the stepper motor in the config file? Just thinking out loud...

@SuperBoppy Great thought!

Since it appears consistent on @Maurici machine, I think SuperBoppy is thinking in the correct direction and should be addressable via the config file or through calibration.

Maurici · March 23, 2023

On 3/22/2023 at 1:36 PM, ken226 said:

The coefficient of thermal expansion for PLA is about 68 microns per meter, per degree Celsius.

A 1cm tall PLA part would see a .13mm change in height after it cools from 220c to 20c.

Could this be some, or all, of what your seeing?

Though about that but i happen to be oversized rather than "shrinked"...

21 hours ago, SuperBoppy said:

If there is a consistent accumulation of error as the z-axis position is increased, perhaps this could be related to the calibration value for the stepper motor in the config file? Just thinking out loud...

Obviously this is my best bet. That could be corrected "easily " with rotation distance values.

However, taking in account how hard is being to measure the actual deviation, unless is in finished parts, im not exaclty sure how i should modify said rotation distance. Specially because at the end im using 2.4 motion ratio with 2.4 config file... and this should be totally fixed?

Edited March 23, 2023 by Maurici

Demosth · March 23, 2023

6 hours ago, Maurici said:

Specially because at the end im using 2.4 motion ratio with 2.4 config file... and this should be totally fixed?

I think this is a "your mileage may vary" scenario - Quite likely, something is a tad off on your machine or components. For example - while it's not expected, maybe you torqued down on your Z belts, which are now stretched slightly, or they were manufactured just a wee bit off, or perhaps you have a z gear that is slightly out of spec (or even very dirty teeth on the gear) could cause a .01mm deviation over 10mm.

I think any of those could cause differences like you are experiencing.

Maurici · March 23, 2023

2 hours ago, Demosth said:

while it's not expected, maybe you torqued down on your Z belts

gilty on this. I didn't know how much really tension them so maybe I over-did it.

However, I think I've found the problem and is the expansion of the whole printer.

I was starting printing when the sensor was reading about 50 degrees... this is 20 minutes.

I've done different trials starting at minute 0 till minute 90 from preheating. temperatures ranging from 23 to 68 degrees.

after 30 minutes, all my measurements are spot on.

it seems the frame is expanding in all directions and sending the dimensions of the print off by 1 to 2 tenths...

this is the last cube after 90 minutes... is enough to me to call it a day. however, this is a massive bugger. 1h of soaking to print a 30 minutes part in ABS...