I have just ran @Alex_vG shrinkage tests (which is great btw!) and come across a bit of a hurdle and wondered if anyone has noticed it before?
Basically I have a H2S and H2C… I ran the initial test piece in ASA on the H2C, completed the spreadsheet and then ran the adjusted for shrinkage settings on the H2S expecting the shrinkage to be a lot better. Both using the same brand/type of material just not the same reel.
But it’s still massively out compared to the results I’ve got with other tests i’ve done so far, ABS, TPU, PLA etc.
Bottom line - It has properly thrown me out as this now completely negates accounting for shrinkage if the shrinkage calculation cannot be cross referenced with the SAME brand/type of material. Just on two different printers?!
H2C on left, then shrinkage accounted for on the right (H2S)
I must stress, this shrinkage calculator has been absolutely spot on when running the material on the same printer, back to back. Just not for two different printers, two different reels?
Any pointers? Just makes me want to sack shrinkage off if its not compatible across printers/materials of the same type but diff reels.
The problem when you don’t use exactly the same filament in terms of the same spool:
It cannot be confirmed whether the spool comes from the same manufacturer or batch.
That’s a significant uncertainty in your comparison. Take Bambu, for example. Even if we could guess where the filament comes from, we wouldn’t know how many factories or manufacturers are involved behind the scenes. I think it’s great how much work you’ve put into this. However, you should check this aspect.
Even if it comes from the same manufacturer, it is possible that one is more moist than the other due to storage conditions. If I understand correctly, we are talking about a difference of 0.152%? That is so small that we are already in the range where measurement errors come into play. Not from your side, but in the overall picture. For example, if one of the printers is calibrated 0.15% off.
So now the question arises: Is it really shrinkage that’s the cause… or is it because a different spool was used, or the printer calibration is a little off, or even a combination of these factors?
As I said, I don’t know, you measured it, I’m just asking!
Incidentally, this topic has already been discussed, albeit with a focus on the profile settings that Bambu provides, but shrinking is also a major aspect of the topic. It’s worth reading.
Basically the filament is not homogenous through the roll, and 2 printers of same make and model both claim to have 50C on the heat bed don’t, and have uneven bed heaters, and etc etc
I ran into problems like this and spent a long time learning “engineering tolerances” and “DFM” (design for manufacturing).
Basically this means things like making larger holes for pins, putting “dogbone fillets” on inside corners, making slots for screw holes instead of round holes so you can move the pieces about sideways a bit, etc etc. Plenty of resources using the keywords mentioned.
Plastic shrinkage is a range and is dependent on many many factors. In plastic injection molding, we aim for the center of the range and for most materials, that gives good enough results. ABS is .4-.8 percent. We generally use .6. Your results show on the low end. .36 and ,49. That sounds like .4 to me.
In 3D printing, I’ve never fussed with shrink as most of these materials are generally low shrink resins, But this is good info to know. I’ll be using .4 on any ABS parts I make that require any dimensional reliability across a long part.
Cheers, its a tricky one as now I don’t want to set a standard for both printers to use as one will be over and one under. I guess that’s tolerance for you.
My next plan of action is to do the same calibration with the same reel on the two different printers to prove out if its the material and one being dryer than the other etc.
Purely for interests sake really.
Keen to know what others do for tighter toleranced parts in terms of workflow.
I just thought about something based on the ABS parts I’ve made that may be helpful. How we let the parts cool will have an effect on shrink rates. I notice that when I take a freshly printed plate with an ABS part on it still hot from printing and place it on a cool surface, it will begin cracking and eventually break free from the plate that is holding it. When I let the part cool inside the printer as the printer slowly returns to room temp, it doesn’t do that, and I bet that has an effect on the overall shrink rate. I believe this might be an important factor too, as we often change the expected shrink by letting the parts cool longer for less shrink, or less cooling for more shrink. For consistent numbers, consistency in how long the part is allowed be restrained to the plate might have to be considered as well.
Good reference IMO. Total dimensional accuracy is a result of the material’s shrink properties, printer gantry mechanical accuracy, plus extrusion related constant errors. Only a couple of those things are guaranteed to travel between printers.
@dann2707 Have you run a calibration with the vision encoder on both printers?
This can possibly rule out a part of the deviations. The tension of belts and the production accuracy of the belts also influences the accuracy of the motion system.
I ran the H2S by itself, no shrinkage settings changed with material B I will call it (original h2s material)
Material A is material that was in H2C.
And my findings were, pretty wild! The part had shrunk 99.1% ish. So significantly more than the h2c, with all the settings exactly the same. 99.5% to 99.1% is pretty mad!
So then I did another test and brought material A into the H2S to trial that. And the results were around 99.2% or so.
So that’s indicative that it’s mainly the printer that’s causing it. Pretty conclusive, not determining the root cause but at least it’s narrowed it down
That’s a brilliant shout and being honest I don’t actually have an encoder plate. I saw a few posts suggesting it’s caused issues so swerved it but now I’m tempted to you know.
I run the calibration approximately once per month and directly after a belt tensioning procedure or a firmware update.
The printer doesn’t give a lot of feedback on the corrections besides a visual map. But the result is a printer with zero skew. Still it requires some flow and shrinkage calibration to get parts super accurate, but I only do that when the parts require it.
Cheers for that. Just gone to look at it and it’s £86 now I thought it was about £60 ish but maybe that was a sale.
Is skew the issues that I am seeing? I’m not familiar with it.
I’m almost tempted to run a full calibration on both and see how it effects it and keep this thread up to date.
I know i’m probably chasing my tail a bit with it, but i’m hoping I can at least get the parts a lot closer. Especially when i’m printing batches of materials on both printers that need to be similar in size.
Have you calibrated the H2C and H2S using the vision encoder? This should be your first step in ruling out any mechanical discrepancies between the machines.
Additionally, did you swap filaments between the two machines and observe the same results? It’s also important to ensure that you dried the filament before testing—this applies even if you’re using a brand new spool, as ASA/ABS is pretty sensitive to moisture.
In my opinion, these factors can significantly impact the results, far more than any minor variations in ASA from the same manufacturer. While there may be slight differences between batches, the issues mentioned above are likely to have a much greater effect on your outcomes.
Thanks for that, i’ve literally just come back from purchasing the vision encoder so will report back once it arrives and keep this thread as a log - even if it might help someone in the future.
Yes I had a go printing with original material A on the H2S and the results were similar to H2S printing with material B. So to me that has ruled out the material. I will post the results once I get back to my computer later.
I can confirm now after doing those tests above, that the parts coming off the H2S appear to be shrunken a lot more than the H2C.
Both printers are in the same room.
A few other variables as highlighted above:
Are the chambers getting to the same temp?
Am I measuring them after the same “post printing” amount of time ie letting them cool off the bed or removing them straight away.
I’m actually enjoying the process of trying to get to the bottom of it. Will see what the vision encoder entails once I get it.
I thought it was about £60 ish but maybe that was a sale.