Bump bump bump, then... layer shift

[donaldlindsay]

Well, it didn't take me long to find another puzzle for the forum. I've got plenty of data on this one, and one or two theories. I've tried searches, but (perhaps due to my inexperience), can't come up with a search term to find a thread that covers a similar problem. So, here it is ;

Since I fixed the last problem, I've been hearing a "bump, bump, bump" when the head lowers to begin a print. At first things were going fine apart from that - I probably should have investigated, but since the prints were coming out lovely, I just carried on.

Yesterday, I had two consecutive failures. In both cases, at about 2mm height there was a "bump" again, and the print head seemed to move back about 10mm. It carried on printing here, meaning the print was misaligned and failing, so I cancelled.

When I run the skates up and down the towers with my hands, all three of them seem to slip into a "notch" around about the places where (on their way to begin a print) I hear a "bump". The height of the three "bump spots", measured from the top of the melamine lid of the base unit, to the top edge of the melamine on each skate, are ;

210mm, 287mm, and about 362mm

As I say, when I run the skates up the towers, they kind of slip into what feels like a notch at these exact points - same height for each skate, at all three locations. When I watch the head descend to print, it appears to be making the "bump" noise as it passes each of these points. I don't want to run it again until I resolve this issue, but I'd guess that the layer shift takes place around the lowest "bump spot".

So, on to possible reasons. I tightened the belts up when I was fixing the last problem. Is it possible I made them too tight?

I also tightened all the screws I could access - I wonder if I've misaligned anything?

I tightened the skates - I think they're right, according to everything I've read - but does this seem a likely consequence of them being too tight? Or even still too loose?

I'm going to loosen the belts a bit, as I've a feeling it has to be that. Does anyone have any thoughts?

[Polygonhell]

You need to isolate what is causing the notchy feeling. It's one of the extrusion, the wheels on the cheapskates or IMO more likely the belt itself is damaged and at that point it's going over the drive gears. Home the machine an do look at the underside of the inside belts.

You can certainly over tighten the belts, the stepper bearings aren't designed for a lot of side load, but it just makes the movement generally stiffer.

[bot]

I had problems with over-tightened belts that didn't manifest themselves in exactly that way, but they were strange. I'd readjust the belt tension and skate tension just to be sure. The belts should be just "tight" enough that they aren't completely loose.

[donaldlindsay]

It's the cheapskate bearings. Seems all six of the fixed left-side bearings have a flat in the same place. I've rotated them so that the flats aren't lined up anymore. The movement is still a bit bumpy, but no obvious 'notches'. Thanks for your help - sorry to be posting problems with such obvious answers! I'm not sure that the flat on the bearings explains the layer shift though. This is probably a separate issue. I'll run a test print and see how it goes.

[Polygonhell]

The layer shift is just one or more steppers missing steps because the force required to move is too great, it could be the flat spots, it could be belts too tight, it could be the motor drivers overheating.
There really ought not be obvious flat spots on the bearings, you might want to talk to SeeMe support.

[donaldlindsay]

Will do. I've been working the printer pretty hard, though - night and day for a week, and some quite long shifts leading up to that. Wouldn't be surprised if I need some new bits at some point soon.

Thanks for the tip about the stepper motors - it's probably the flat spot "notch" then, since it struck in the same place twice. I'll try a print now that I've distributed the flats, and see how it goes.

[donaldlindsay]

So... it seems the failure issue was a bug with Cura. I found another thread that deals with something similar, here ;

http://forum.seemecnc.com/viewtopic.php?f=62&t=6417

I've just reduced the z lift to cure a stringing problem (successfully), so rather than restore that setting, I tried simply reloading the models into Cura, and rearranging them a bit. Seems to have worked.