With my eyes not being as good as they use to, I’m really liking the “auto” version of the flow dynamics. I can also test multiple filaments at a time which is great. My concern is accuracy.
Do you guys use the Auto version when calibrating, or do you stick with the Pattern Method?
Thanks, Cody
I find the auto tuning very much hit or miss. I think the largest discrepancy I had was for sunlu basic PETG in white, where auto came up with the K factor of around 0.023 while manual tuning resulted in 0.039
Auto is useful for finding a starting point for fine-tuning, but in my experience it is not precise enough.
If I run Auto five times in a row, I get different values each time, with the occasional outlier showing a significant jump up or down.
What is particularly surprising is that this still happens despite the machine’s sensor package, especially since the eddy-current sensors are clearly intended to support accurate automatic calibration.
I just rely on the other auto, the eddy current sensor thing. It appears to be fine enough for me and ChatGPT claimed it’s more accurate than one might think. Then again, ChatGPT itself is less accurate than one might think, lol
And the saddest part is you never get a K value figure from it. That would have been so nice for assessing repeatability and for comparing with the manual/auto things’ results.
I’ve found that the pre-print automatic flow calibration has way too much variability to be useful compared to manual, but the ‘manual’ automatic calibration that prints a line of the selected filament on the buildplate that you access from the calibration tab does a better job than I can do unless I break out some form of magnifier like a loupe or pocket microscope as long as the filament is properly dried first.
Or so you could grab the K when the autocalibration lands on a really good setting, and “lock it in” in the filament parameter dialog thing, and turn off autocalibration = speed up prints a bit, reduce wasted filament a bit.