# Adjusting internal diameters on ABS parts
**URL:** https://forum.lulzbot.com/t/adjusting-internal-diameters-on-abs-parts/451
**Category:** Hardware
**Created:** [February 2, 2014, 11:00am UTC](https://forum.lulzbot.com/t/adjusting-internal-diameters-on-abs-parts/451 "2014-02-02T11:00:27Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@camelopard](https://forum.lulzbot.com/u/camelopard)
#### Post date: [February 2, 2014, 11:00am UTC](https://forum.lulzbot.com/t/adjusting-internal-diameters-on-abs-parts/451/1 "2014-02-02T11:00:27Z")
I have noticed that holes tend to become a few tenths of a millimeter smaller than expected when I print things. I’m guessing this is due to the way ABS cool down, and maybe also how circles are actually regular polygons? Using a very high $fn in openscad doesn’t seem to make a difference for me.
What is the generally preferred way of dealing with this? Changing the source measurement to compensate, or is there some setting in slic3r or similar?
---
### Author:  [@piercet](https://forum.lulzbot.com/u/piercet)
#### Post date: [February 2, 2014, 6:33pm UTC](https://forum.lulzbot.com/t/adjusting-internal-diameters-on-abs-parts/451/2 "2014-02-02T18:33:31Z")
The tighter your X and y belts are, the more accurate your circle reproductions will be. If the dimensions are off in one particular direction or another, that’s the one you want to concentrate on tightening. Also upgrading to the newest version of Slicer if you haven’t already improves circle geometry specifically.
---
### Author:  [@gannon](https://forum.lulzbot.com/u/gannon)
#### Post date: [February 3, 2014, 2:06pm UTC](https://forum.lulzbot.com/t/adjusting-internal-diameters-on-abs-parts/451/3 "2014-02-03T14:06:36Z")
camelopard: You might find [this](http://hydraraptor.blogspot.com/2011/02/polyholes.html) interesting
---
### Author:  [@camelopard](https://forum.lulzbot.com/u/camelopard)
#### Post date: [February 3, 2014, 4:05pm UTC](https://forum.lulzbot.com/t/adjusting-internal-diameters-on-abs-parts/451/4 "2014-02-03T16:05:24Z")
> [@piercet](#):
>
> The tighter your X and y belts are, the more accurate your circle reproductions will be.
My belts are pretty tight, but I will keep that in mind, thanks.
> [@gannon](#):
>
> camelopard: You might find [this](http://hydraraptor.blogspot.com/2011/02/polyholes.html) interesting
Oh my, yes. That’s a really good article. And thorough. It explains a lot. Will try to use low vertex count polygonal holes and see what happens!