RepRap vs. Marlin Firmware

[spegelius]

I work in IT (software development/DevOps) so I spend quite a lot of time with problems somewhat similar to these issues so my problem resolving abilities (and tolerance of stuff shitting itself) is quite high at this point . But yeah, there are moments now and then when something just doesn't work and it does get discouraging at some point.

I have S3D, but don use it much nowadays. My main slicer is KISSlicer, but occasionally I use PrusaSlicer, mostly for prints that I know KISS has problems with. Also supports are better on PrusaSlicer and of course S3D's custom supports is the best solution. Although I've gotten into habit of modeling my own supports when needed and also generally modeling in a way that the need for supports is minimal. I've also tried soluble supports with one multi-material machine and it does work quite nicely in general.
For multi-material printing I use all those slicers. To add the tool change gcode, I wrote a post-processing tool to add the extra commands around the tool change commands because at the time when I got into multi-material printing, none of the slicers supported a single-nozzle/multi-extruder setups. The post-processing tool supports PrusaSlicer, S3D, KISSlicer and Cura at the moment. There's also MMU2S support in it, although I haven't tested it myself, the profile is from a user. The tool: https://github.com/spegelius/filaswitch

About MMU2S: I assume that to change filament, all you need is to issue the tool command: T0, T1 etc.? I sliced a multi-color print on PrusaSlicer for MMU2S a while ago just to see how the gcode looks and I didn't see any extra commands so I'm assuming that the Marlin handles the MMU2S-stuff when it sees a tool change command. But as I haven't actually seen or used a MMU2S-unit this is just speculation. That's how I would do it anyways... Getting bit impatient waiting for my unit to arrive, luckily there are other stuff like this to keep me occupied
Edit: tool change command won't move the nozzle to purge bucket/purge tower, that usually comes from slicer. Although maybe there's support for that in Marlin? In RRF that could be done with a macro that is executed with tool change command, but RRF doesn't support MMU2S AFAIK.

RRF progress: I tested the fan frequency and with 500Hz, only 100% speed spins properly. After testing few frequencies, it seems that below 50Hz is good, so settled for 25Hz. With that the fans starts to spin at 45%, although it really doesn't move any air at that point. BTW, I did all that on my phone using the web UI to edit config files, control the fan etc. Can't get that with Marlin, even though through Octoprint quite a lot is possible.
I printed a benchy, but it seems I need to get a better part cooling fan and shroud because there were obvious issues with cooling. I'll probably redesign something to better work with the Volcano. Still the print finished and apart from cooling and stringing issues it looks fine.
I calibrated Pressure advance and compared to Marlin Linear advance there values aren't comparable. The PLA I used isn't the same as when I calibrated Marlin, but the difference is quite big. With Marlin, the LA value was 0.28 and with RRF a comparable value is around 0.7. I also calibrated retractions so I'll print another benchy to see if the stringing is better. Also maybe try another filament, current one is cheap PLA from Poland.
RRF in general: the limitation of LPC1768 based boards is that it has less memory than Duet boards which means there can be only only HTTP session at a time. So connecting with my phone to DWC while I was also connected to it on my PC resulted in connection errors and the board even did a reset few times. The connection limit can be probably worked around by setting up a HTTP reverse proxy on some PC/raspi so only thing directly connected to the DWC is the proxy -> only one session. The random board resets just show that the FW is still experimental, as it says on the github page.