However I noticed that the front idler pulleys have some friction before they start moving. Feels like it requires to much force before they start spinning. Are there any improvements that can be done here?

I will take them apart and see how it looks like, maybe they just needs some grease to reduce the friction.

Using bearings on that shaft meant I could increase the tension in the belts another notch without having the y motor skip steps due to too much friction. I believe that it's kept backlash and resonance from the belt drive to a minimum

Like I said I haven't tried this myself yet and don't know if it'll work, but talking to jagowilson and wardjr about this I think I have a decent idea of why it is helpful.

The point is not to just replace the two bushings on each side of the frame to lessen the friction on the rod.
You'll also replace the pillow block bushing found in Solidoodle 4's with a bearing, and for Solidoodle 3 add one (or two) like pirvan has done on his machine.

The pillow block which was added in Solidoodle 4 I think was to reduce the bending of the back 6mm rod caused by the motor belts and the side Y belts pulling it. But since Solidoodle decided to use a bushing instead of a bearing the pillow I think does more harm than good by introducing more friction, even if reducing the rod bending, because that bushing is closer to the origin of the bend and so more friction is created.

In short we want to decrease friction and rod bending (which also results in friction...) at the same time and see how it goes.
When well reduce friction, we can make the Y rod tighter and get rid of any remaining backlash without causing the motor to skip steps.
Unless friction can affect stepper motors in any other way except step skipping (which we don't have), I don't think this will help by itself, without tightening the belt some more.

Pirvan also had two pillow bearings in his Solidoodle 3 before he made the y motor tensioner mod. It's possible that also played a role for his success.

I have ordered a set of flanged bearings for the Y rod as well.
Did it make any difference for you?

Using bearings on that shaft meant I could increase the tension in the belts another notch without having the y motor skip steps due to too much friction. I believe that it's kept backlash and resonance from the belt drive to a minimum (now it's all about making sure the hot end mounting and frame are stiff; a few challenges there).

They're great for the last little bit of fine-tuning, but unfortunately I would not expect the difference between bushings and bearings will be large enough to explain the amount of slop shown in your circle test print.

Pirvan's tension arm is worth a shot on either the SD3 or SD4, as they both have that drivebelt. I think it's a useful and not overly difficult mod (and would be straight onto it myself if I hadn't already mounted the y motor directly to the shaft outside the case!). While I couldn't guarantee it's going to fix your issue, it's an idea worth trying if you're keen, and can only improve the situation (even if only a little).

At the moment I think we've only been looking at your belt drive; perhaps we can get more creative. Is there any slop/play in the carriage or mounting of the hot end? Try pushing it backwards and forwards while cold, and see whether there's any movement (e.g. the hotend mounting is loose? the carriage rocks on the rails? etc.). The SD4 is quite different in design here to the SD3, so forgive me for not being able to suggest anything in more detail.

Did it make any difference for you?

If not, try to get 3 6x13x5 flanged bearings and replace the 3 bushings on the back Y rod with them. Should just come off if you push and then you could push the bearings in their place then.
Something similar to what the first thread grob linked to suggests to try. As Solidoodle 4 users we already have the holder for the bearing/bushing so no need toprint one I think.

Thanks for the tip of the tension thing by pirvan, I will also try it out. Even thoght it didn't resolve the problem for redbarret.

Judging from that circle test, it looks to me like there's a fair bit of backlash in the y axis (see how it's misaligned front-back at the start of the circle?), this is likely at least partly the cause of the gaps in the previous prints.

If you've tensioned and balanced the y-belts, have you also considered the tension of the small drivebelt on the y axis driveshaft? You can adjust the tension a little on this one by loosening the 4 bolts holding the y motor to the case, pushing the motor down, and re-tightening the bolts.

Also, have a read through these two threads, there may be some inspiration there:
http://www.soliforum.com/topic/4824/how … -for-ever/
http://www.soliforum.com/topic/8357/y-s … belt-slop/

Any ideas on why the prints are not solid when they should? You think that is caused by the same thing that makes the circles uneven?

I printed the circle stl, it's attached.

Thanks
Peter

I have done some adjustment and I do see improvements, but I'm not satisfied yet.

The Y belts were loose so I have tighten them and aligned them again.
I also calibrated the extruder, it was extruding 136mm instead of 100mm so I modified the extruder steps per mm setting in the EEPROM so that's fine now.

The flow rate was also adjusted slightly. Thanks for the .48 tip, I had no idea about that.

But still the circles are still not round and I'm getting "holes" in the prints.
I have attached a picture. I think my problem with the roundness has to do with the X axis as I see issues at 9 and 3, is that correct? The X belt is tight, are there any further adjustments that can be made?