Poor implementations of new features in 6.4.0
As usual, when I saw the announcement of the 6.4.0 firmware update with new features, I was eager to install it on both of my MK4S printers.
But after running a bunch of prints through both printers, I'm very disappointed with the print process and have turned off the new features.
Auto-retract (MK4 family, XL)
About 90% of my prints use the same filament for the last print as the next print. Due to this feature, I now have to deal with the strand of filament ejected at the end of the last print and the strand of filament ejected while reloading it.
I imagine that this feature has value in printer farms where each printer has a queue, and an automatic filament change could be initiated between queued prints with different filament. But for my use case, it's nothing but a nuisance.
Print sheet detection (MK4 family)
This feature triggered false positives on about 25% of my prints! Sometimes it occurred when the print sheet being even very slightly askew relative to the locator pins at the back of the bed, even if the print would have occurred just fine. But I can't even say that that's the cause of false positives, because sometimes it occurred when the print sheet was perfectly aligned. This feature seems to be poorly tested, and since I can't recall the last time I forgot to remount the print sheet, I just turned it off.
Lazier homing, more cautious calibration
Since installing 6.4.0, I've noticed that the time between issuing a print command and the start of actual first layer deposition feels noticeably increased or delayed. I haven't sat and watched it for the full five minutes to identify what's changed, but I presume that this is the issue. This isn't a big deal - I am not running a printer farm to maximize throughput, so an additional 2+ minutes of startup time is fine - but I wish that we could get some additional information as to what "more cautious calibration" involves and what it achieves in exchange for the delay.
Reworked filament sensor evaluation
I'm hoping that this fixes the UI hang issue that I raised a while back, where it was impossible to escape the "insert filament" dialog without resetting the printer and losing your print.
But the only change that I've noticed regarding filament sensing is that filament runout sensing failed twice in a row with one of my printers. The spool ran out and the printer didn't detect it and just printed in the air without filament; not only did it ruin the print, but I needed to pop open the side cover and pull out the remaining length of filament with pliers. Feels like a regression.
It's possible that this new "evaluation" process requires filament sensor recalibration, but if that's the case, then the new firmware should have prompted and required recalibration.
Overall, these updates really make it feel like the MK4S line is no longer a priority for Prusa Research, which is full-steam-ahead on the CORE One lineup. This is a disappointing turn of events and I hope that future updates prove me wrong.