Calibration: holes 1.2 mm smaller

[pablobell 0]

Hi everyone,

I am having a strange problem. The parts I build have very accurate measurements on the outside, but the holes are smaller than expected. I printed some test parts with Netfabb, 2.9mm and a filament packing density of 84%:

All holes have a diameter about 1.2mm smaller, independent of hole size. The small object on the left is odd because its circular groove has the convex border with the right measurement, but the concave is 1.2mm smaller, so the groove itself is 0.6mm thinner.

Does anyone know how to correct this? Is this hardware or software related?

Thanks!

PS: the picture was taken up close, so it presents some optical distortion (the straight borders look round).

[billdempsey 0]

Actually, that's not strange at all. It's typical. Plastic shrinks as it cools, which makes the holes in particular shrink. For example, any time I want an 8 mm hole, I have to create it as 8.4 mm. A 3 mm hole has to be 3.1 or 3.2 depending on the filament I'm using. Sadly, the shrinkage isn't linear, or we could just calculate it. The larger the hole, the shrink percentage increases. It has been a trial and error process to figure out what dimensions to use for various holes I regularly put in objects. The good news is, you get better at instinctively accommodating the shrinkage the longer you use the printer.

EDIT: Actually, if your shrinkage is remaining constant for every size hole, that IS strange. The holes look round, so it is probably not belts or motors as those usually create distortion in one direction or another. My money is on software. I'd download Cura and try it. I purchased NetFABB, but it never worked right for me, so I abandoned it in favor of Cura. Unfortunately, it's not perfect either. I've created a model which just won't slice correctly in Cura no matter what I do.

[owen 19]

It's actually not so much shrinkage as you would get it over the whole model and not just the holes. someone explained it here before but basically you will get it more if your circle has less points.

It is very common and it's usually closer to a fixed amount than a percentage. For example a 30mm diameter hole may come out at 28.6mm, that is 1.4mm less than expected and a 3mm hole comes out at 1.6mm.

Some slicers may cater for it like slic3r I hear but most others don't seem to.

You just have to design your model with a bigger hole than you want or get your drill out.

It only happens with horizontal holes (vertical ones can get it but much less)

Other things you can do in your designs are make "C" shapes which won't close up as much and have a bit of spring to accommodate whatever shaft etc. that is going in to it.

One other thing you can do is print a template that has many holes in it from 1mm to 50mm etc. going up in size 1,1.5,2 ... 10,11,12 ... 20,22,24 ... 30,35,40,45,50. Some people have put them on thingiverse.

Then you measure your holes and record them in a spreadsheet or whatever for future reference when you are designing something.

I've found when I slice with KISSlicer mine all end up about 1.4mm less diameter so I just go that much bigger plus or minus depending if I want a tight fit.

[joergen 2]

Hi everyone,

I am having a strange problem. The parts I build have very accurate measurements on the outside, but the holes are smaller than expected. I printed some test parts with Netfabb, 2.9mm and a filament packing density of 84%:

[attachment=0]small holes.jpg[/attachment]

All holes have a diameter about 1.2mm smaller, independent of hole size. The small object on the left is odd because its circular groove has the convex border with the right measurement, but the concave is 1.2mm smaller, so the groove itself is 0.6mm thinner.

Does anyone know how to correct this? Is this hardware or software related?

Thanks!

PS: the picture was taken up close, so it presents some optical distortion (the straight borders look round).

just as a starter, I would suggest to look into your basic printer calibration: if you have the new V3 bolt, and the bertho extruder, your e-steps/mm are probably closer to 835 (from the previous 866), resulting in a 4% over extrusion.

second, I thought netfabb has changed the packing density to flow rate? anyway, PLA needs 1.00 (100%), and ABS a hair more, 1.03 (103%) (or 97%, if you see it as packing density), together with proper measurement of the diameter.

but in general, you should be relatively OK with 835, flow 100%, pack 100% with standard 2.85mm PLA

I haven't used netfabb in the past year or so, so I can't comment on the hole precision. I know that cura/SF tried to compensate for hole diameter, but it never really worked. slic3r has a dedicated small hole adjustment code, which does an awesome job making holes the exact diameter. KS does an OK job with holes, but they are generally a hair smaller than expected, but not too terrible. a 3mm holes is tight, but still big enough to push a M3 screw through (horizontally).

overall vertical holes tend to be more filled in than horizontal holes.

based on your observations, I would say netfabb does not have any small hole compensation code.

[pablobell 0]

Thank you all guys for your responses! It's a relief to know I wasn't doing something wrong.

I printed the same objects using Cura 12.12A and got the exact same dimension as with NetFabb, so I confirm Cura is not correcting for small holes either. I'll try slic3r and KS.

About the packing density, NetFabb v4.9.5 calls it "calibration Fine tuning ("filament packing density")". It's a confusing name, as smaller numbers give less extrusion, so it works as flow rate. I was using 2.90mm diameter and 84% density for that cheap clear PLA. The filament actually measures (2.92+-2) mm, but at 84% the extrusion was well calibrated. When I changed to pigmented PLA I had to turn the flow rate up again to 100%.

[pablobell 0]

I have been looking into this problem, and at least some of it might be due to loose belts.

As you can only tighten them so much, does anybody know a good procedure for measuring hysteresis? And whether M98/M99 actually work?

 

[msurunner 2]

I am always hesitant to recommend tightening the belts more. They need to be tight, but there's definitely more forgiveness there than elsewhere, and it's all too often used as a lazy answer for another problem in the system. Over tightening them is a serious possibility and it can lead to issues with pulley slippage and stepper resistance/heating. Are your holes irregular or do they appear to be generally round? If they are pretty dang close, I would check pulleys and what not before tightening the belts and I would only do it in small steps. Also, the small belt tensioners you can find on thingiverse work great for belt tension, as it's usually the small ones that end up pulling the stepper motor up the frame and loosing tension. The tensioners require that a bolt turn, which is awesome. I have had students working on the printer for months now and haven't had to adjust tension once.

 

[pablobell 0]

I use these banana belt tensioners made by MoonCactus. Great stuff. And I agree with you in that the belts from the motor tend lo loosen more often (these were the loose belts I was talking about).

I just wanted to know if hysteresis is accounted for. At fast print head speeds, the inertia when stopping or changing directions would make it difficult to calculate a precise position. But at lower printing speeds it could explain the very well known issue with smaller hole sizes, which is the topic of this thread.