heatbed heats up when off

I have a Huxley reprap that I just built with a Melzi v2.0 controller. I connected the heatbed thermistor to the "btemp" connector and the heatbed MOSFET input to the pin labeled "A4" on the Melzi. When I connect the MOSFET, the heatbed slowly heats up, even though I have the heatbed off in the prointerface controller. Any ideas why the Melzi is driving the MOSFET? Why do the instructions say to connect it to "A4" and not the "bed" connector?

Try disconnecting the input pin from the Melzi and plug in yhour printer and see if it still heats up. If it does, use your multimeter to check the resistance between the right pin of the MOSFET and the tab of the MOSFET. If there's little to no resistance, the MOSFET's gate is blown and the bed is receiving full power while the printer is plugged in.

The bed doesn't heat up when the pin is disconnected from the "A4" output pin of the Melzi. The MOSFET checks out fine. This leads me to believe that the controller is outputting a PWM on "A4" when it's not supposed to. I'll check the pin with a scope to verify that assumption when I get a chance later today at work. Is it possible that this is a firmware issue?

I just probed "A4" with a scope and found a 50% PWM on that line. This makes me think it's a firmware issue.

I've attempted to reflash the firmware, but I'm having a slight issue with it. When I try to install the 64 bit 0023 version of the arduino IDE, the .tgz file contains a "arduino" file without an extension where an executable should be. Adding the .exe extension to the file doesn't make it run. I can download and run the 1.03 version of the IDE perfectly well, but when I try to compile the firmware, it complains that a "conflicting return type specified for 'virtual size_t SdFile::write(uint8_t)'". I'd rather get the 0023 version up and running, if possible, since 1.03 doesn't seem compatible. Any ideas?

*EDIT*: My bad. I downloaded the 64 bit Linux 0023 IDE by mistake.

Edited 1 time(s). Last edit at 02/22/2013 12:17PM by nicholasreiser.

I was able to reflash the Melzi using the 0023 Arduino IDE. However, I still have the same issue. There is still a square wave with a 50% duty cycle on pin "A4", which drives the MOSFET on heatbed, even when the prointerface software isn't running.

I've solved the issue. It was a stupid mistake. I misread the instructions and connected the bed temperature input to "A4" instead of "D10", the pin above it. Thanks for the help.