Round surfaces not smooth anymore

[Nicolinux 288]

Hi,

I have an interesting problem and I'd need your opinion.

These two objects have been printed with similar print settings.

Yet the first one has a weird surface. Looks like some layers are retracted inwards or missing. Not sure if this is under extrusion. I've run the cylinder test and it came out perfectly (same filament):

I noticed this behavior on very fine prints with curved surfaces before. I can't say when it appeared first but it might coincide with the switch to the Olsson Block.

What do you think?

[neotko 1,417]

Since your machine it's doom from day one I would try to sell it and get a new one... Sorry but it's a curse?

Ok now seriusly, sell it...

Nah, now for real... Sell it...

Just kidding!

sell it.....

Ok now for real, back to your problem. SELL IT!

Stuff you need to check someday. Since your machine I think many of us know that you have review almost everything. Could you check the belts tension? UM2 inner belt tensioner IMO it's really really weak. Also, just to add more stuff that you might not have check. Do any of the pulleys rotate on a non circular way? Maybe you have a bad set of pulleys on the x/y. Also, even that um2 has that block spacers on the pulleys, have you check if the belts/shafts are on perfect alignment? Like with the umo's

Also, just another crazy idea, have you check if the sideblocks do wooble? Just in case...

Imagine this but on a um2 and move x/y by gcode.

Hopefully this error this time it's just bad luck, but omg. Your um2 it's cursed!

BTW very clean 18mm3/s test!

Also no changes on the slicer program?

Edited March 5, 2016 by Guest

[peggyb 171]

you can check the outside of your nozzle, even the slightest scratch or burnt residue can influence the surface. And this is probably Colorfabb? Always feel that it prints more 'fat', less 'crisp' then other brands, but this may depend on the type of model, speed, temp, etc...

[Nicolinux 288]

Thanks for the hints everyone.

@neotko: Sadly I can say "check" to everything you mentioned. But these things are very minimal.

@peggyb: Alright, I did another test and printed 5 times the same object (same .gcode file) but with different nozzles/filament. Btw. here is the .gcode file (prints at 30mm/s and takes around an hour to complete): Link

1. E3D 0.4 Nozzle, colorFabb Orange

2. Jet 0.4 Nozzle (long tip), colorFabb Orange

3. Jet 0.4 Nozzle (long tip), Faberdashery White

4. Jet 0.4 Nozzle (short tip), Faberdashery White

5. Jet 0.4 Nozzle (short tip), colorFabb Orange

I'd say the nozzle and filament makes a bit of a difference, but some "horizontal gaps" are still there.

Edited March 11, 2016 by Guest

[neotko 1,417]

Umm. Short tip/long tip means that the z it's different? I think I see a z pattern there. At slow z speed or better put, when the z changes little... Umm

This it's just a gut felling (most probably wrong).

Do a z home. Then. Could you by using both hands under the bed push it up/release. If the bed moves up/down and does 'tack' then you have one of the z nuts that have playroom. My 3 umo+ beds, one had a big almost 1mm playroom and because the bearings where bad the bed wasn't able to keep moving the bed at the same rate.. My English isn't good enough to explain this properly.

Basically. Test if the bed does 'tack' when pushing it. Clean the z smooth rods by doing z gome, clean, release, home, clean, release (with something like a eyeglasses cleaner that don't leaves fibers).

Even better. Add half a kilo weight to the bed and print again. If it doesn't change, this isn't related.

[IRobertI 521]

How are the temps? Stable?

[Nicolinux 288]

@neotko: The nozzles have a different height, but I re-leveled the bed upon each nozzle change. Long tip looks like this:

Sort tip like this:

I'll test the bed (and try to fix the other possible causes you mentioned in your first post).

@IRobertI: Temps are very stable. Ever since I switched to a 35W heater and a new temp sensor (also performed PID Autotune), the temp barely twitches even when the fans come on.

[Nicolinux 288]

Little update. I have installed the Extrusion Upgrade Kit and the issue is much less pronounced. It appears some filament sorts hide it better than others.

[neotko 1,417]

Maybe it was preassure feedback from the bowden angle. @ultiarjan rail system should help with that.

Edited March 21, 2016 by Guest

[Nicolinux 288]

@neotko: I'll keep this in mind if the new z-nut and linear bearings won't work

And I noticed that I started two threads for the same issue. I'll close this post now and will continue here:

https://ultimaker.com/en/community/20188-smooth-walls