3D Printer Filament is not round!

[old man emu]

I noticed that people were saying that you should measure the diameter of your filament and enter that value into your slicer software. I had seen various values for the diameter of the filament, so I decided that I should measure mine, which is nominally 3mm.

As I started to measure the diameter of my filament with a digital vernier caliper, I noticed that I was not getting regular values. They ranged from about 2.92 to 2.98. At first I thought it was my bad technique, then it dawned on me that filament is an extruded product. I began to take my measurements with more care and attention to where I was measuring. I found that if I measured the diameter from topside to bottom side, I consistently got values around 2.98mm. When I measured from side to side, the values were around 2.92mm. Then it dawned on me! The filament is not round. It is oval. I was measuring the major and minor diameters of an ellipse.

I took two sets of ten measurements from both directions and averaged them. I found that from top to bottom, the average diameter was 2.98mm, and from side to side it was 2.91mm.

Using these values, I calculated the area of the ellipse (pi * r1 * r2) to be 6.81 sq mm. A circular filament of 3mm has a cross-sectional area of 7.07 sq mm.
Then I calculated the radius of a circular filament with a cross-sectional area of 6.81 sq mm and found the diameter to be 2.94mm.

Why is this important?
When you load an object's *.stl file into a slicer program, one of the items of data you give the software is the diameter of the filament. From that diameter, the program calculates the amount of filament required to make the object. If you are doing printing for profit, then this figure is important to you for costing the job.

Using the nominal 3mm diameter under-estimates the amount of material required by approximately 4% compared a diameter of 2.94mm.

I do not have any nominal 1.75mm filament to test whether it is oval, too.

Old Man Emu

[MysteryAlabaster]

Moved to relevant category.

[old man emu]

Sometimes I wonder if people aren't becoming too anal about the accuracy of their machine's output. Afterall, these machines are supposed to be for prototyping, not for producing museum quality one-offs.

Old Man Emu

[Vanguard]

Hurrrmf ! ! Almost choked ! Ha ha !

[Vanguard]

Good job Old Man Emu !

[Vanguard]

I finally realized why your filament is oval. It is laying on a solid surface before it has cooled enough. I have not checked our filament yet, but I doubt this is a consistent thing.
If the filament extruder is at an elevation high enough, or the filament is cooled rapidly enough this will not happen. A lot of commercial extrusion companies run the extrusion through a water bath directly after it exits the extruder nozzle, this is not a good thing for filament that will be extruded again later. So most just air cool it. I suspect Old Man Emu's filament very well may be home made, or come from some one with their own filament extruder.

[old man emu]

I purchased the filament from a well known supplier here in Australia (BilbyCNC). There is no place of manufacture on the reel. Unless there are manufacturers of inexpensive filament in the Americas or Europe I guess that this filament comes from China. Bear in mind that the differences in dimensions were of the order of 0.05mm. An extruder nozzle hole could be that much out of round.

The point that I was making was that it is important to know the diameter of your filament if you are involved in commercial production, and that you should not use the nominal diameter in your slicer software configuration. I would suggest that the figure of 2.94 be used for calculations.

Old Man Emu

[Vanguard]

There are variables that cannot be accounted for, and that small of difference is very small indeed. Even air flow can account for that variation.

As filament is extruded it expands to a larger diameter than the size of the orifice it is being pushed through, this happens with all plastics, at all stages, whether it is during the making of the filament, or the printing of a part. That makes calculating actual usage very difficult at best. Your discovery is still something that is helpful, as it lets us be aware of possible hang ups from distorted filament, and other problems that may arise. I would have never even thought to look for out of round filament. I still thank you for your discovery !

Your filament is probably of high quality, I did not mean to imply otherwise. Home made filament can be as good, possibly better than factory filament, depending on the reliability of the supplier. You definitely know what you are doing, keep up the good work. Your contributions are greatly appreciated !

[Kingoddball]

Sorry to Hi-Jack, but.. What do you think of the quality of BilbyCNC??
I have not used their product yet and was going to grab a sample.
With the update of oval filament..

Last few rolls I purchased have been TERRIBLE (NOT from Bilby).

One roll was actually even larger than 3mm!!
Another roll (two actually) did not bond to itself...

Out of 6 rolls, 2 were good.

[PrinterNinja]

..... Unless there are manufacturers of inexpensive filament in the Americas or Europe I guess that this filament comes from China. ...

Old Man Emu

Thanks for the insight on the cross section calculation. I was well aware that filament could be oval, but had not thought much about the volume of material and how that is affected. I would have assumed that if the filament left the extruder nozzle round and sagged into a different shape, like an oval, that the resulting cross sectional area would remain constant. Is it safe to assume that you have a perfect oval? I'm not sure. Maybe it's egg shaped. Who knows. I guess you have to use an oval to get an estimate of what the theoretical round diameter would have been though.

But I wanted to comment about Chinese filament. People are pretty quick to associate it with poor quality, but there are many producers of filament there with a variety of ability. I sell and use filament from a supplier in China who has manufactured engineered plastics of many types for nearly 25 years. Sheets, bars, billets, filament, etc. I personally have 32 spools of it because I wanted to see and test everything that I offer on my site. The roundness and average diameters have all been within their stated tolerances (which are identical to nearly all other manufacturers worldwide). I have had no issues with contamination and the PLA and ABS print with very low odor. In fact I can barely smell the PLA at all. I have very good consistency between colors.

I sampled some filament from a US producer and the 3mm filament was way out of round and oversize. I really wanted to buy American for several reasons - freight costs being a big one - but the Chinese filament I found was great quality and reasonably priced. (Not the cheapest mind you - there are some REALLY cheap offerings from Taiwan and China. I suspect these are the ones selling on Amazon for well under $30.) Not everything made in China is poor quality.