Bed temperature shoots up when printing

I have a strange issue with my Prusa i3 - the bed heats up just fine before I start printing, and stays at the desired temperature accurately, but once I start printing, the bed temperature rises by some 20-30 degrees Celsius.

I've done the PID autotune and set the values to Marlin's configuration.h. The problem is not caused by heat transfering from the hot end or print itself, but rather it seems that there's voltage on the bed connectors even when there's not supposed to be.

My setup:
Prusa i3, e3d v6 hot end
Slic3r - Repetier-Host - Marlin
PCB heatbed MK2b dual power, 12V power supply

I can't find any descriptions of similar problems anywhere, and I'm clueless as to whether this is an electronics or software/firmware configuration issue. Any ideas?

Edited 1 time(s). Last edit at 01/28/2015 08:52AM by Waiski.

I have had those same problems.
Turned out it was due to using the fan output of the Ramps board (D9)
Are you printing PLA and using the D9 fan output?
Try using a separate constant voltage for the fan to troubleshoot and disable the FAN_PIN in marlin
When i print PLA and have my fan mounted, i switch it on manually after the first layer via a switch.
In pins.h (marlin) i have set the FAN_PIN to -1 so it will not be used.

Still, it fixed my issue, but not sure why it showed that behaviour (maybe some inference of the PWM signals?)

Can it be that in your slicer settings you have something like a different bed temp for the first layers and the rest of the layers?

---

http://www.marinusdebeer.nl/

The bed is connected to the correct place and it's not the slic3r temperature setting - I use temperatures like 55-70 C, often 5 degrees higher for the first layer than the rest. The measured temperatures often go up to 85-95. Also when this happens, Repetier-Host temperature curve might show 0 at the bed temperature output but the bed temperature still goes up, and I measure 12 volts between the bed connectors.

However it might be that winand is onto something - I did some debugging and I suspect there might be something strange about the wiring between my RAMPS's D08 and D09 outputs. I need to investigate that further.

Edited 1 time(s). Last edit at 01/31/2015 11:45AM by Waiski.

Setting FAN_PIN to -1 seems to have solved the problem - I can't say for absolute certainty yet since the problem appeared erratically to begin with. If this is indeed the case, wouldn't it mean that I have a short somewhere between the D8 and D9 outputs? I can't find it with my cheap multimeter and puny electronics understanding though - can someone point out a sure-fire way to test if there is a problem? What should the measured resistances between D8 +/- and D9 +/- be?