alranel
commented
10 years ago I can't say I never saw this myself, but using M109 commands is probably overkill and largely suboptimal solution. The usual approach about this includes two things:
- using distinct PID settings when fan is on - I think there's an open issue here about sending an M-command upon fan activation, or maybe I saw someone doing that with a post-processing script;
- protecting nozzle from airflow - fan should only cool the print, and NOT the nozzle. In theory, you shouldn't see any temperature drop.
Thingiverse contains several mounts and ducts to achieve the second goal. Have you done something about this?
cgapeart
I ran into an interesting edge case last night. Just as my printer transitioned to the 2nd layer, and the cooling fan turned on (for PLA), I noticed in the log from Pronterface that the cold extrusion prevention warning came on, and a couple of extrudes were skipped while the PID control caught up. This caused a few voids in the layer, and it took a few layers for the printer to build up enough plastic to fill in the holes or weaknesses.
In my case, the default cold extrusion warning came on at 170 degC - the Marlin default. I had a setpoint of 185 (down from 195 for the first layer), so I experience at least a 15 degree drop while the PID control loop was adjusting for the change in cooling rate. 20-30 seconds later, the PID loop caught up with the change, and there were no further problems with the print
The skipped cold extrudes can be avoided using the existing heating and cooling options: -ramping up the fan speed slowly -starting at a lower speed -printing at a hotter temperature to begin with.
However, I suspect there are other cases where the fan speed is changed suddenly, e.g. when starting a bridge, that could cause a similar problem. I also suspect the sudden drop of head temperature could create adhesion inconsistencies in the printed objects, even if the temperature doesn't drop low enough to prevent extrusion.
In order to deal with that case, I was thinking of coming up with a couple of tweaks to the final gcode to improve performance in these cases: -Issue M106's before M104's and M109's when happening during the layer change -print head temperature changes use M109 instead of M104 following a fan change (causing the printer to wait a bit for the temp to settle again before continuing) -Add an M109 command following some/all M106's resetting the same temperature and causing the printer to wait -If the printer is going to stop and settle the temperature again, an option to retract or move the print head off the model might be necessary as well - sitting on the model with a hot or dripping head will likely do more damage than could be fixed even in the worst case where some extrudes were skipped.
I suspect this is not worth the effort to add these options. I am going to retry the print I had problems on later tonight to generate sample files/logs and gcode, Unless somebody else has a me-too or a another take on the issue, it should probably just be closed off as a non-issue. I am just posting this as a discussion topic at this point, to see where it goes.
I suspect it started when I was cutting down my print temperature and adjusting cooling because the whole part was not firming up very well, and I set my minimum fan speed too high to begin with.