PWM fan ports

Finally took board out and scoped it while running. M106 S128 does send a nice 50% PWM signal to the gate of the mosfets controlling the two unused connectors: CN18 "FAN" and CN19 "PWM". So PWM fan controls through M106 are ~ok in the firmware.

But then it gets weird in the hardware:
* The gates of both mosfets Q7 (CN17) and Q8 (CN18) are connected together to the same I/O pin - if the connectors worked both fans would be at the same speed.
* As pointed out in other threads the PCB layout is wrong: specifically the drain and source leads are swapped for the transistors of the 3 connectors in a row: CN9 "LED", CN18 "FAN" and CN19 "PWM".
* Here's a link to a vaguely similar mosfet that shows the pinout of the SOT-23 package. I hope the mosfet on the board is rated for more current though! http://www.vishay.com/docs/70226/70226.pdf
* Because drain and source are swapped the two fan connectors and the LED interior light strip are always on - current flows to ground through the internal mosfet diode. That's how those connectors are always on despite the wrong layout.
* Fortunately the gate connection is correct. If the gate connections were wrong too the +12 would could have flowed back to the I/O pins and blown up the board, but weren't, so didn't :-)

So...I don't care about the LEDs being controlled and only need one PWM fan, so I only flipped and rotated Q7 for the PWM connector. Less work and damage.

Attached picture shows flipped and rotated Q7 (and some accidental soldering iron melting of the connectors. Whatever). The source and the drain could be soldered down to the pads with some ugly blobs, but it needed a small piece of wire to connect the gate.

Haven't run the wires through, reassembled the printer, and printed with a fan, but M106 now controls a fan plugged into PWM connector. The Pn option to select fans does nothing, any M106 controls the connector.

If you're not comfortable soldering surface mount parts...ask a friend who is...don't want to ruin a board made of irreplacium.