PC Blend, MK4, Obsidian Nozzle
I am preparing to make a print for a client of a scanned object that will be used as a mold for Cast Bronze. It must be dimensionally accurate. It will be used a number of times, therefore must be robust, and in a high temp environment. So, I selected Prusament PC Blend.
Step 1: Print Precision Method Cube and check Extrusion Multiplier. Default of 1 in Prusa Slicer, but mine measures at 1.1 (which is high). The PMC has a lot of buckling at the sides when it cools, which confirms case of deformation by filament contraction on cooling. Its important to make measurements of wall thickness only when the cube has cooled down and removed from the bed.
Step 2, when XM is dialled in, print an item of known dimensions, and measure it when done. To get "precise" dimensional accuracy, measure in all directions, L, H, W (X, Y ,Z) and get the amount of expansion/contraction by (a) calculate percentage difference in each of the 3 dimensions (b) calculate the average, or take the most important dimension for your production model. Scale up or down the model to compensate for expansion/contraction.
What I found, was that PC Blend Black prints smaller than expected. The Extrusion Multiplier using the precise method works out at 1.1 (for me, anyway) on a MK4 having standard 0.4 ObXidian nozzle, printed on a satin bed with glue applied to aid print removal.
Step 3. If the item is to be cast, then you need to account for expansion of the liquid metal being poured into the mould, then contracting later when it cools, the goal is to get the right dimensions when cold. So, determine by Google search, the coefficient of expansion for the exact metal alloy being used. Calculate the 3D expansion using online calculator (that has inbuilt applicable laws of physics formula) Then scale up the model by whatever percentage it needs to be when the metal is in liquid state. Then when the metal cools, it should contract to the correct size. Or at least, that's the theory.
Or, perhaps someone with a lot of experience doing prints for casting will have better ideas, if so, please share them.
Best Answer by Thejiral:
Here is the stl file (zipped)
shrinkage+exposure_calibration
On one side measure the distances of the inner sides with a caliper, on the other side measure the distance between the outer sides. Try to measure the distances without the influence of the first few layers to avoid bias due to elephant foot.
You can use the calculator here to calculate the results: https://yaqwsx.github.io/printer-calculator/#/shrinkage
The calculator is intended for resin printers but shrinkage is still shrinkage and the "exposure bleeding" indicates overextrusion, even if negative values appear to mean overextrusion rather than underextrusion.
RE: PC Blend, MK4, Obsidian Nozzle
PC-Blend is not a good choice if you need high dimensional accuracy without printing it in a printer with very hot enclosure (70-90°C I would assume). Otherwise it has a high tendency to warp, even if you have perfect bed adhesion to the end, and let it cool down entirely in the enclosure. There is of course shrinkage, to quantify that I would recommend not using the cube method, you need a calibration print with several distance markings so that you can do linear regression and caclulate shrinkage error vs over/underextrusion error.
If you need the high temperature resistence and robustness of PC-Blend, maybe you want to have a look into PCCF from Prusa. It has the strength of PC-Blend by far and large but is much less prone to warping and should be suitable for dimensionally accurate printing.
Mk3s MMU2s, Voron 0.1, Voron 2.4
RE: PC Blend, MK4, Obsidian Nozzle
Warping is not one of the issues I face at this time. I already produced a test print specifically to establish that (and check the dimensional accuracy).
Can you suggest a model that I can use for the "calibration print with several distance markings" ?
I acknowledge that PC Blend CF is an option, but one I'd prefer not to exercise at this time. But ...
At this point I have scaled the object for (a) dimensional issues caused by shrinkage of the PC (found during testing)
and (b) allowance for thermal expansion then shrinkage of the molten metal.
and about to run another test print. If dimensional accuracy is not tamed when that's done, I'll pivot to PCBlendCF as you suggest.
Another option is PolyCast, of which I have a roll on order for another part of the same project. PC represents the safe zone.
I just do not know much about PolyCast, whether its robust or not. It is temp tolerant however. I'll be testing it....
However PC Blnd CF has a generally rough-ish surface finish, while being smooth at the same time, with reduced layer lines.
This particular target object unfortunately has a curvy bulgy top surface that does not look great, when a shiny finish is the desired outcome, so a lot of post-print processing work will be needed. (Did I forget to mention that this object is to be a re-usable object model for casting (of phosphor bronze), the reason it needs to be robust and temperature tolerant.) And before you ask, I did suggest to the client that he use SLA for this project, which he declined.
RE: PC Blend, MK4, Obsidian Nozzle
Ok, I totally get that you prefer PC without CF, given that there might be a lot of sanding involved in post processing.
I personally am using a modified calibration print originally designed for SLA printers. Maybe not perfect but for my purposes good enough (for me printing is just a hobby) and it worked much better for me than test cubes of a single size. I can upload my test file in the evening and you can see if it is any help to you.
The nice thing is you can fine-tune the extrusion factor and the shrinking compensation with the same test print.
Regarding casting, I fear I am not much of a help. Never attempted it myself.
Mk3s MMU2s, Voron 0.1, Voron 2.4
RE:
Here is the stl file (zipped)
shrinkage+exposure_calibration
On one side measure the distances of the inner sides with a caliper, on the other side measure the distance between the outer sides. Try to measure the distances without the influence of the first few layers to avoid bias due to elephant foot.
You can use the calculator here to calculate the results: https://yaqwsx.github.io/printer-calculator/#/shrinkage
The calculator is intended for resin printers but shrinkage is still shrinkage and the "exposure bleeding" indicates overextrusion, even if negative values appear to mean overextrusion rather than underextrusion.
Mk3s MMU2s, Voron 0.1, Voron 2.4