X-Axis off by a few percent / Circles not round

[noa]

Hi,
i recently got myself a CTC Bizer with a dual extruder. I use Sailfish 7.7 as Firmware. From start on I'm not able to get round circles or the correct size in X-direction. I printed multiple cubes using different Slicers: ReplicatorG, Makerware, S3D and Slic3r - same Problem. My y related measurements are within the error range of my caliper, z is pretty close and x is too short by 5 to 6%

A 20x20x10 cube is about 18.52x20x10.03, a circle is getting flat parallel to the y-axis (undersized radius in x direction).

I tried different filaments: same result.
I checked for wobble: everything is stiff
Slicers: same resault
I print abs in a full enclosure, no problems with shrinkage at all.
I reviewed the few pieces i made with the original firmware, same Problem but these parts were too organic or too small to get exact values.
Dimensions stay consistent over 50+- layers.

When i did the firmware update i used the Flashforge machine settings in ReplicatorG, i flashed the firmware again with the Replicator Dual selected - same problem.

Has anyone a idea how i could solve this Problem?

[EagleSeven]

Hi,
i recently got myself a CTC Bizer with a dual extruder. I use Sailfish 7.7 as Firmware. From start on I'm not able to get round circles or the correct size in X-direction. I printed multiple cubes using different Slicers: ReplicatorG, Makerware, S3D and Slic3r - same Problem. My y related measurements are within the error range of my caliper, z is pretty close and x is too short by 5 to 6%

A 20x20x10 cube is about 18.52x20x10.03, a circle is getting flat parallel to the y-axis (undersized radius in x direction).

I tried different filaments: same result.
I checked for wobble: everything is stiff
Slicers: same resault
I print abs in a full enclosure, no problems with shrinkage at all.
I reviewed the few pieces i made with the original firmware, same Problem but these parts were too organic or too small to get exact values.
Dimensions stay consistent over 50+- layers.

When i did the firmware update i used the Flashforge machine settings in ReplicatorG, i flashed the firmware again with the Replicator Dual selected - same problem.

Has anyone a idea how i could solve this Problem?

My guess is a problem with a Stepper-motor or an output-signal from Motherboard.
To determine which, you could swap X and Y motors to see if problem followed the motor
or stayed with that Mobo channel.
(or replacing the X-motor may be easier, if you have or can get one with correct internal connections)
(not all motors are same internally, even though they look the same outside and have same number)

If it's Mobo channel, it could be the voltage-level is to Low,
so monitor voltage with a O-scope or meter to determine problem, if you can ?

Also, check all pulleys on motors and shafts to make sure they are same size.

[bigo93]

Do your circles come out like this?

20151226_195909.jpg

If so the problem I was having was the Y motor belt was loose.
Not the belts on the left or right for the Y axis, but the short belt connected to the motor at the back of the printer.
Loosen the screws holding the motor and push it down, but not too hard, and tighten the screws back up.
This will tighten the belt.

I would so this before trying to change steps/mm

[noa]

I think i will try to modify the firmware to compensate the steps per mm first.

When my new bearings arrive i have to take the whole thing apart anyway. I noticed today that my housing is not exactly square, this will add up to this issue too. At the same turn i can test these motors, with the casing around it's i a pain in the butt to find good spots for my dso probes. This together with the lack of any usable reference points for measuring the travel distances or arcs between the axis, I'm close to salvaging this printer and its parts for a more engineering friendly design.

I'm still questioning myself how these slicers calculate the paths when a cricle is needed? Is there any arc compensation? When i construct a circle with 20mm outer radius and 1mm wall thickness, S3D sets two oval circles overlapping each other. The inner Circles gets pretty round with this trick but the outside... I will take a few pictures tomorrow...

[noa]

Do your circles come out like this?

20151226_195909.jpg

If so the problem I was having was the Y motor belt was loose.
Not the belts on the left or right for the Y axis, but the short belt connected to the motor at the back of the printer.
Loosen the screws holding the motor and push it down, but not too hard, and tighten the screws back up.
This will tighten the belt.

I would so this before trying to change steps/mm

yes this very close to mine! I will try this right away! Thanks.

[noa]

Hi,
i fixed my problem.

The biggest impact of how round my circles could get had that short belt bigo93 mentioned! I just didn't notice that it was there when i checked my printer - thanks again.

Next step was to again check every screw on the housing - they get loose over time.

And to close the gab between expected and measured dimensions of my parts i used double sided tape to put a square on my print bed. With that i got a reference point to measure travel distances in jog mode. I was able to recalculate my steps per mm and updated the firmware with my new values. Now my dimensions are within the non significant digit of my cheapo vernier caliper.

[bigo93]

Great.

But how did you update the steps/mm on the printer?

I try changing it on makerware but it never saves it to the printer's firmware and resets when I restart the printer.

[noa]

Simply modified the xml flie ReplicaterG uses when uploading a new firmware to the printer:

Inside the ReplicatorG folder you will find it inside the "machines" folder - replicator-sailfish.xml if remember right.

you can modifiy this line for example: (I removed the xml tags, forum software usually deletes it anyways)
axis id="x" length="227" maxfeedrate="18000" homingfeedrate="2500" stepspermm="94.139704" endstops="max" Pulley dia: 10.82mm / 1/8 step = 1/(10.82 * pi / 1600)
axis id="y" length="148" maxfeedrate="18000" homingfeedrate="2500" stepspermm="94.139704" endstops="max"

[Geoff]

Guys, use makerware for the CTC, it works a crapload better than S3D on firmware 1.0.

[EagleSeven]

Guys, use makerware for the CTC, it works a crapload better than S3D on firmware 1.0.

Yes, I tried Many different Slicers and Many different versions of Slicers
and None worked as well as Makerware-2.4 .