Filament mass estimate
Hi!
I need to produce a print with a specific mass (for ballast, for an experiment), and I'm trying to avoid having to reprint it too many times.
Yesterday I made a first test print to check the geometry of the part, to make sure that it fits into the assembly and I chose to print it with 3 bottom layers, 4 top layers, 2 perimeter layers and 5% lightning infill (with a 0.8 nozzle on the IS setting for 0.4 QUALITY layer height). PrusaSlicer estimated the printed weight to come to 107g. The actual weight turned out to be 100g almost exactly.
I double-checked the density of the material by calculating backwards from the Prusament website details I got from scanning the QR-code and the density comes to 1.227 g / cm^3, so taking cross-section variability into account it is probably close enough to the stated 1.24 g/cm^3. The difference is either way not enough to explain the difference in extruded mass. That would require an actual density of 1.159g/cm^3.
Are there any other error sources that can explain the difference in estimated mass and actual mass?
The actual part needs to be 900g, wich requires an infill level of 76-83% (and a print time of 12 hours) depending on if I trust PrusaSlicer or if I correct according to the printed test part. I don't want to have to reprint unnecessarily...
RE: Filament mass estimate
I made a test print anyways last night. In the slicer it is calculated as 892g, and it came out to 833g. The ratio of sliced to actual mass is 892/833 = 1.07, the same as for the previous test print with less infill.
I realize that it could be my scale that is off. I will check against a calibrated scale on Monday.
RE: Filament mass estimate
If you know the actual density of the filament from the manufacturer you can configure it to match in the Filament settings. I notice the preset profiles for prusament PLA are set at the generic 1.24 still. Makes sense to accurize it to start with.
The used filament calculation includes everything, so any brims, skirts, supports etc. I think it even includes the initial purge line or anything from the start gcode block.
Did you look at the details on the legend shown on the slice preview or did you just look at the totals given after a slice ? The legend breaks the types down so you should be able to see what slicer at least thinks the parts weigh.
RE: Filament mass estimate
Well that's the issue. This is Prusament PLA that has a density of 1.24 g/cm^3 according to the datasheet. This is set in the slicer. As I stated in the first post I calculated the density backwards for my specific roll from the data on the Prusament website that you get from scanning the QR code, and it comes to 1.227 g/cm^3. Too close to 1.24 to explain the difference.
The picture below shows the item (a cylindrical shape). It prints without support or brims so all 892g goes into the part.
The only explanations for the difference I can think of now is that either my scale is off by 7%, or the used filament calculation in PrusaSlicer is off by 7%.
RE: Filament mass estimate
Remember the material is hydroscopic too. It will absorb moisture. Although I would be surprised if that would account for 7%.
Id be running test on smaller test parts and experimenting with drying the rolls before printing to rule things out.
RE: Filament mass estimate
I just checked my spool on tensor beam and I noticed that almost full spool which is 1kg will vary +-4g within time spam of 2 days mainly because of different temperature and humidity. So it's about 0.4%, so it's not even half percent, thus in my case drift caused by temperature/humidity is negligible in comparison to the real weight. This means proper weight calibration is more important than other factors.
See my GitHub and printables.com for some 3d stuff that you may like.
RE: Filament mass estimate
Were you using Arachne or pressure advance? I would guess that calculating the extruded mass with those features enabled would be tricky.
RE: Filament mass estimate
If the weight of the part is that critical I think I would print a hollow version of the part and then fill with some type of granular material such as sand, BBs or lead shot rather than trying to get an exact weight using all plastic filament.
RE:
Were you using Arachne or pressure advance? I would guess that calculating the extruded mass with those features enabled would be tricky.
I was, on both counts. I thought these were interesting questions so I reprinted the hollow version of this part (i.e. sliced @ 107g) in two versions:
- Using Classic perimeters instead of Arachne.
- Using a non-IS setting in PrusaSlicer (this should turn off pressure advance right?)
The first version did not affect the sliced mass. The second version estimated 106.8g instead of 107.2g, so it obviously affects the calculation somehow - but negligible for this problem. Both versions came out to an actual mass of 100g.
I'll check all my prints tomorrow against a calibrated scale at the lab at work.
RE: Filament mass estimate
If the weight of the part is that critical I think I would print a hollow version of the part and then fill with some type of granular material such as sand, BBs or lead shot rather than trying to get an exact weight using all plastic filament.
I'd like the mass to be evenly distributed inside the part, that is why 3d-printing is an interesting solution. That's definitely an option for other applications though. Also, that would not answer the question of if PrusasSlicer has a systematic error in the estimation of used filament - or if my scale is wrong 🙂
RE:
OK, it took a good while longer to actually check the weight of my part. I've measured it on a calibrated scale now and it came out to 833.2g, so identical to my kitchen scale. I'm going to get my part done with the knowledge I have now.
So far I've done the following:
- Very hollow test piece. 2 perimeter layers and 5% lightning infill. Sliced to 107.2g. Actual mass: 100g.
- First try at actual piece. 76% rectilinear infill. Sliced to 892g. Actual mass: 100g.
- Another no 1, with classic perimeters instead of Arachne. Sliced to 107.2g. Actual mass: 100g.
- Another no 1, with pressure advance turned off. Sliced to 106.8g. Actual mass: 100g.
All pieces so far are consistently 6.5% under the target mass. I will make another test piece next week, in a cube shape, and slice that to just over 100g. That should give an indication if the error is shape dependent.