gcode: Fix issue on unloaded filament temperature being higher temperature than loaded filament (lower temperature)

[bkerler]

This PR adds a configuration store to remember the past filament and does compare if the temperature is lower than the filament that has been unloaded before, and if, does set the last filament temperature only for loading until purging has finished, thus preventing the nozzle to clog.

Fix is related to issue #3585, #3083 and #1821

[nazar-pc]

Not sure where to comment, but these tags on GitHub are extremely annoying for those who subscribe to releases via Atom feed. Please consider an alternative approach there. I'm talking about the likes of https://github.com/prusa3d/Prusa-Firmware-Buddy/releases/tag/build%2F6.0.0-BETA%2B7062.PR3681.B4

[danopernis]

@nazar-pc thanks for bringing that up. The CI settings that got merged to the master branch after 6.0.0 release seems to contain some changes causing builds to fail and CI system to generate those tags. I will try to fix that tommorow.

Meanwhile I manually deleted those tags.

FYI @bkerler

[spooknik]

I appreciate the work done on this PR and I just wanted to chime in with my input.

I think it would be better to apply a purge temp offset derived from the difference in filament load temperatures. So if loading PLA (215) after PETG (240) the purge temp would be 230. I explained my line of thinking here.

Basically the formula is:

Previous Filament Temp - Current Filament Temp = Temp Delta 

Temp Delta * 0.6 = Purge Temp Offset

Current Filament Temp + Purge Temp Offset = Filament Purge Temp

[bkerler]

@spooknik I don't think a delta of 0.6 will be good as for some filament types this will clearly be too low. Guess some testing will be needed to find a reasonable factor. But I will consider your recommendation :)

[spooknik]

@spooknik I don't think a delta of 0.6 will be good as for some filament types this will clearly be too low. Guess some testing will be needed to find a reasonable factor. But I will consider your recommendation :)

Yea I could see a problem going from PC to PLA for example. I guess you can apply an offset multiplier based on the delta. Bigger delta = bigger multiplier.