Infill issue: low infill bands in y every +/- 40mm

[coen 11]

After many (3) years of semi-reliable service my Ultimaker Original has now developed a problem (or it has become more apparent) that I can't seem to fix. I was printing a wedding gift for friends of mine: a plate with their initials. In this large flat object a regular banding became apparent: good infill, bad infill, good infill, bad infill, roughly in y-direction. This effect shows from the bottom layer, to the top, and also through the whole prints.

Pictures tell a thousands words, so without further ado, here's a complete novel (apologies for the coordinate direction not being the same in all pictures, but be assured: the pattern is in y-direction):

On the small one the problem wasn't that apparent on the top, but I still need to fix it with Putty.

However, on the big plate it is really bad:

notice also the loose support material: it would be a really good idea for Ultimaker to make round corners instead of the perpendicular ones: perpendicular motions put a peeling force on the deposited filament.

I can go through the motions and put a whole load of info about cura settings here, but as it has happened now in two big prints I'm pretty convinced it's a hardware issue related to the driving system in x/y. Maybe banana rods lifting the printhead at a regular interval, or perhaps a speeding issue due to offcentered pulleys...

The extruder train seems unlikely as you would then expect the effect to be regular in extruded length, not in y-coordinate. Issues with the bed would show more heavily in the bottom layers, and disappear at the top layers, which is not the case.

My current approach would be to take the machine apart, inspect all parts, replace where necessary and then put it back together again. Any idea of what to check before I start doing that?

Edited July 6, 2015 by Guest

[maartenw 6]

Coen, first of all: beautifull design!

But about the problem:: from the first 1 pictures it is seen that the blue tape is often used, and maybe worn out. The banding is in the same direction as the taperuns. So could you replace the tape (everywhere) and make a first layer to see whether the band still exists?

[IRobertI 520]

How does the bed look, is it perfectly flat? An easy way to check is to hold a ruler against the surface and hold it up to the light and see if light seeps through.

How about the rods of the gantry? Are they all true or do they wobble?

[coen 11]

Coen, first of all: beautifull design!

Thanks! But I can't take credit for the logo, I copied it from the invitation :-).

But about the problem:: from the first 1 pictures it is seen that the blue tape is often used, and maybe worn out. The banding is in the same direction as the taperuns. So could you replace the tape (everywhere) and make a first layer to see whether the band still exists?

I actually replaced the tape between printing the square tray and the plate. The plate has a very large flat surface and I was worried about how ugly it would be with all the cracks and holes ;-). So this sadly couldn't be it, as the issue didn't improve between the two prints.

How does the bed look, is it perfectly flat? An easy way to check is to hold a ruler against the surface and hold it up to the light and see if light seeps through.

It is kind of hollow, both in x and y. Only very slightly, if I press down on the ruler it sinks into the tape and the light disappears.

How about the rods of the gantry? Are they all true or do they wobble?

Yesterday I tried printing a full size flat object but without filament to check this out. By eye I can't make out any wobble during the printing on the gantry axes. I'll take it apart tonight though to see how they hold up when I roll em on a flat table.

[coen 11]

mmmm.... it's pretty hard to judge whether the rods are curved or not. I can see a small airgap when I roll em on the table or put the on a ruler... but you would expect this gap to change in size if they are curved and you roll them, but this does not happen, or at least not in a way that I can see it. I don't have a flat table available hear, but I might take them to work to check it out...

what's the best way to test them for curvedness? I can try putting them in a drill and then checking at higher speeds?

[maartenw 6]

However this may not be a solution, would it be possible to increase the thickness of the bottom and top layer?

This should be considered as a workaround, not a solution, but I guess the wedding will not wait for your beautifull plate to be ready.

[coen 11]

However this may not be a solution, would it be possible to increase the thickness of the bottom and top layer?

This should be considered as a workaround, not a solution, but I guess the wedding will not wait for your beautifull plate to be ready.

 

The wedding already took place... I gave the plates with bad infill and asked to get them back after to fix them up properly... so I'm now working with Putty...

Thicker layers don't help, it will just add more layers with holes in them... flow helps a little bit, but then the regions that have proper infill now get 'plowing marks'. These could be sanded down, but yeah :(

As for the rods: I took them to work today and rolled them on a granite block... they're perfectly straight.... that leaves the pulleys or some other issue as the possible cause...

Suggestions welcome!

[gr5 2,167]

Your issue would make total sense if one or more Y axis pulleys aren't centered.  Please measure the distance between good or bad areas and see if it's the same as the circumference (from outer surface of the belt's point of view).

This would make sense because the the Y movement would be non-linear and when commanding it to move 1mm it would actually move slightly more than that or less than that depending on the orientation of the pulley.  When it moves too far you get underextrusion.  Too short you get overextrusion.

It could be any of the 6 Y pulleys including the motor one and the other one on the short belt so don't just look at 4 pulleys.

Also if you tell the printer to draw a single diagonal line I would expect it to be crooked back and forth as it passes through these areas of more and less extrusion.  Like a sine wave but subtle.

Edited July 22, 2015 by Guest

[coen 11]

Your issue would make total sense if one or more Y axis pulleys aren't centered.  Please measure the distance between good or bad areas and see if it's the same as the circumference (from outer surface of the belt's point of view).

 

The diameter of pulley+belt is 13.7mm, times pi gives 43mm, which is rather perfectly in the range of the spacing issue... I'll inspect the pulleys to see if I can see anything on them from the outside. Do you know if Ultimaker has since upgraded them? I've seen a topic on the forum about custom wellcentered pulleys... I'd rather order those then get a new crooked batch from UM...

here's the custom pulley thread: http://ultimaker.com/en/community/view/4110-top-quality-pulleys-what-do-you-think-is-an-appropriate-price

[gr5 2,167]

I've seen that thread before. Evidently the pulleys are perfect but the hole drilled through them is not. This makes it kind of obvious if the pulley isn't centered as you mainly just look at the hole.

[coen 11]

After reading through the pulleys thread fully I decided to go for a GT2 pulley/belt system as an upgrade to fix this issue. As that needs new printer parts I firstly reassembled my printer. First results: the clicking sound that had been plagueing it before is gone . Print quality also is nice, but I am yet to print a full surface to see if the infill issue has been reduced or not (adjusted belt tensions too).

The major find however: there is a major offcentre issue visible on the pulley on my y-driving stepper... as well as that on some pulleys the screws push the pulleys offcenter too. So yeah, GT2 upgrade to be ordered this afternoon and until that time: work with what I've got.

Edited August 10, 2015 by Guest