Bridge issues

Hello everyone,
after much experimenting, I have drawn some conclusions about bridge management by Slic3r, that I would like to share to gather other users' feedback.
My setup: Prusa Mendel, K extruder, Melzi board with Sprinter firmware (the kit assembled by Heacent). Printing in ABS, 0.15 mm layer height with 0.3 mm nozzle, 240 degrees for all layers (use less and bridges will be damaged by subsequent nozzle passes, use more and the filament will soften, the extruder gear will dig a notch in it, and printing will stop). Using Slic3r 1.0RC2.
First thing I noted is that when drawing bridges Slic3r seems not to take into account that the layer height is not the same as in other conditions, since there is no material to spread it upon. Maybe I noticed this because I am using layers much thinner than the nozzle opening. In practice, bridges will snap unless I set an insane flow rate multiplier (at least 2.5, 3 works best). Anybody noticed this?
Secondly, it looks like that, when covering an infill area (in my case honeycomb), Slic3r correctly recognizes the area as a bridge area and sets the correct flow rate and activates the fan. But fails to use the (in my case slower) bridge speed, extruding instead at infill rate. This means that, due to the huge multiplier, lots of plastic are extruded, requiring a great amount of energy and potentially leading to hot end cooldown because of slow response of the PID controller, and thence filament block, notching and failed print. Or, if the printing goes through, at the end of the bridge layer (also due to PID inertia) temperature rises, exceeds 300 degrees, the idiot firmware shuts down the hot end and does not turn it back on when temperature drops, so again failed print.
Anybody can confirm this?

Edited 1 time(s). Last edit at 02/18/2014 04:27PM by jameson.

Sounds like you need to tune your PID If it is behaving that badly. It shouldn't significantly overshoot in this scenario unless something is wrong. I agree that it is a very bad idea to use your bridge flow rate but not your bridge speed. If your printer is tuned for speed, the limiting factor is often the extrusion rate the extruded is capable of. If what you describe is true, you would have to slow theprinting of the majority of your model down by a factor of 3 just so it can keep up when it does the top layer bridging.

Thank you for your help. I am already printing quite slow, 20 mm/sec for perimeters, 10 for bridges and 35 for everything else. Tuning the PID is one of the next steps, together with switching to Marlin, but I am waiting for a spare Melzi to experiment upon (being a professional digital hardware and firmware designer, I am afraid of abandoning a working condition if I am not sure I can go back to it and restart from there - I got the board preprogrammed and no source). Looks like Sprinter does not support PID tuning except from recompiling and reflashing (I have not yet looked at the source). Actually I forgot to tell that the hot end thermometer behaves wildly above about 250 degrees: I think the conversion function (or the voltage divider for reading the thermistor value) has not been tuned to work above that temperature. The effect is that, at high temperatures, the reading runs above 300 degrees in a couple of steps. I do not think the reading is accurate, but nevertheless it stops printing.
Going back to the main topic, have you experienced too that the speed for the first layer above infill is not taken from bridge settings, while everything else is? It looks like an oversight, not easy to spot unless you are using bridges and thin layer height, and have an inaccurate temperature reading.