Slots print out with different widths despite being the same

[Wangmaster 0]

The part I'm printing has two slots with the same width of 1.75mm to loose-fit a piece of PCB (see attached pic). Several prints later, I found that my U3 consistently prints the two slots with different widths, the left one at 1.56mm and right one at 1.64mm. Now I understand that holes and gaps always come out slightly undersized which I can compensate for, but what I don't understand is how two exactly the same gaps can print out differently. Why is this happening, and how can I identify when this will happen in the future?

 

Thanks!

 

Block.stl

[gr5 2,192]

So this part looks like it's functional and doesn't need to be pretty.  I'm hoping UM comes out with a functional/accuracy profile some day.  Until then, try this.  Parts will look worse, but fit better.

 

All printing speeds at 40mm/sec (or 25mm/sec for super high quality).  This is critical as at the speed change from infill to shell it will over extrude briefly and underextrude when speeding back up.

disable acceleration control

disable jerk control

set initial horizontal expansion to -0.2 (optional - this is a bad idea for some prints with thin walls on the bottom layer)

All line widths to 0.4

wall thickness 1.2

top/bottom thickness 1.2

walls 3

horizontal expansion -0.03

 

Also in CAD make vertical holes .4 to .5mm larger than desired hole size (or drill them out after printing).

 

 

 

[gr5 2,192]

I'm guessing the one on the right is a wider gap.

 

It's the same reason vertical holes are too small.

 

PLA acts like a liquid rubber band milliseconds out of the nozzle.  It's like snot.  It's also rapidly shrinking before it becomes a solid.  So it pulls inward when you make a small circle or square or corner.  Because it pulls inward corners and things tend to "shrink" inward.  This happens before it even solidifies.  It's much worse on corners as it's pulling inwards.  It doesn't affect flat sides so much but can reach pretty far from corners.

 

That wall on the side of the left slot helps anchor things and keeps the slot accurate at the wall.

 

You may think it's a pain to compensate for these errors in CAD (the best solution) but keep in mind for injection molded parts you have to do all kinds of crazy design considerations and hacks/adjustments as well.  It's just that you don't know about these because the engineers at the injection molding factory are doing the adjustments for you (and it costs a lot).  e.g. 90 degree corners - if you want to be accurate have to be adjusted by a degree or so in cad.  Crazy stuff.

[Wangmaster 0]

Sorry for not being clear but left and right are with respect to the picture. So the slot supported by the wall is 0.08mm tighter than the free-floating slot. May not seem like much but its very noticeable when fitting small parts.

 

13 hours ago, gr5 said:

That wall on the side of the left slot helps anchor things and keeps the slot accurate at the wall.

 

That would make sense but like I said, the anchored slot was less accurate than the free floating slot, aka it was actually tighter.

 

13 hours ago, gr5 said:

So this part looks like it's functional and doesn't need to be pretty.  I'm hoping UM comes out with a functional/accuracy profile some day.  Until then, try this.  Parts will look worse, but fit better.

 

All printing speeds at 40mm/sec (or 25mm/sec for super high quality).  This is critical as at the speed change from infill to shell it will over extrude briefly and underextrude when speeding back up.

disable acceleration control

disable jerk control

set initial horizontal expansion to -0.2 (optional - this is a bad idea for some prints with thin walls on the bottom layer)

All line widths to 0.4

wall thickness 1.2

top/bottom thickness 1.2

walls 3

horizontal expansion -0.03

 

Seems like sound advice, but are these parameters suitable for every print? Or only for my specific case?

[gr5 2,192]

That's as expected and as I predicted - the slot with the wall was tighter. I'm not sure why they are tighter than expected but probably because the nozzle is 0.4 but the default line width is something like 0.35.  This is a less precise way of printing.

 

The above parameters are for all prints where you don't care how it looks but you care how well it fits.  You can get even better quality by slowing it down to 25mm/sec or faster, worse quality by speeding it up.  Notice I didn't specify the layer height as most people understand exactly what that gives you.  I usually do 0.2mm layer height for practical (not pretty) things as the best way to speed up a print is thicker layers. In my opinion.

 

The thing that the above suggestions will hurt the most is ringing.  So if you have text on the side of your print it will look pretty bad.  But when you measure the area of the ringing with calipers it will be quite good.

 

Even with this more accurate profile, sometimes you will want greater accuracy and you will have to fix the last little bit by tweaking the cad model.