Welcome! Log In Create A New Profile

Advanced

All objects ever so slightly too small… e_steps_per_mm value to blame?

Posted by Pointedstick 
All objects ever so slightly too small… e_steps_per_mm value to blame?
October 15, 2011 01:29PM
I just discovered that everything I print is about 3.5% too small in the X and Y dimensions (Z is fine). My e_steps_per_mm value for the X and Y axes were determined programmatically using the forumlas on the wiki and I arrived at a value of 80 for both. I'm willing to change it to my newly-calculated correct value of 82.96 for each axis, but if I needed to do that begin with, doesn't that mean there's something wrong with the formula? Or is this just something to be expected given the imprecision of printed fasteners and gears?
Re: All objects ever so slightly too small… e_steps_per_mm value to blame?
October 15, 2011 07:36PM
The formula looks right and it doesn't depend on the accuracy of any printed parts, just the pitch of the belt.


[www.hydraraptor.blogspot.com]
Re: All objects ever so slightly too small… e_steps_per_mm value to blame?
October 15, 2011 07:37PM
How big are these parts. Better yet, what is the actual error? Besides the actual extruded filament width possibly being smaller than it should be, there's also shrinkage as the plastic cools down that would make parts smaller.
Re: All objects ever so slightly too small… e_steps_per_mm value to blame?
October 15, 2011 09:04PM
A gear I printed couldn't fit the M8 bolt it's supposed to fit. That tipped me off, so I printed the nickel calibration piece and found that it can't actually fit a nickel. It's slightly too small. This is with white Ultimachine PLA printed on blue tape.



Thingfarm claims that my belts are T5, and that's what I plugged in to get an e_steps_per_mm value of 80 for my X and Y axes.
Re: All objects ever so slightly too small… e_steps_per_mm value to blame?
October 15, 2011 10:13PM
Circles come out smaller on a reprap printer for a number of reasons. Nophead has a few blog posts on this subject. But, one of the main reasons is that circles are actually printed as a polygon inscribed inside a circle which would have a smaller diameter. I also think that the other problem that happens if you're printing above 20 mm/s feed rate is that the Arduino is not able to keep the steppers moving continuously when it has to execute a lot of short moves that make up a circle. The time it takes to move is shorter than the time it takes the Arduino to prepare the next step. So, the head pauses during this time. These pauses results is small blobs if the distance is large enough (radius of 10 mm or so) but it is not visible if the distance is small (very small radius). This results in thicker lines or blobs when drawing circles and reduces the diameter of holes. If you watch your printer, you can actually see it go slower on circles than straight segments and even vibrate when printing above 20 mm/s feed rate.
Re: All objects ever so slightly too small… e_steps_per_mm value to blame?
October 16, 2011 09:06PM
I printed that exact gear with the same white ultimachine PLA, and I think the model file is to blame. I just hit it with a lighter though, it did crack a bit, but seem to work fine.

The white PLA also seem to "flow" more than the other PLA I have, and tended to fill smaller holes which were fine before. My printer is now calibrated to compensate for it though, so I have not had many problems with it.


--
-Nudel
Blog with RepRap Comic
Sorry, only registered users may post in this forum.

Click here to login