Electronic – Arduino Pro Micro clone: connect USB and RAW power at the same time

I have fried two Pro Micro clone boards that are the same as the ones discussed in this post: Are these two Arduino Pro Micros jumpered correctly?

They are 5V/16MHz models which came with SJ1 open so I naively shorted out the SJ1 jumper with a solder bridge to match what the SparkFun open hardware schematic notes state – that SJ1 should be closed for 5V and open for 3.3V.

They worked fine on USB power but when I connected +12VDC to RAW while leaving USB connected, the USB host computer shut down the USB port and the Pro Micro boards no longer work: when connecting just USB, the green power LED turns on for a few seconds then fades out (since the polyfuse is in overcurrent mode and shuts down). One reads 4.4 ohms from VCC to GND and the other 6.6 ohms so it seems that the Atmel chip is blown.

This doesn't make sense to me. There is a protection diode between RAW and UVCC which prevents over-voltage going out the USB. Why are the boards damaged now? Even with SJ1 bridged, the worst case is feeding back the 5V from the regulator output through the polyfuse into the USB power pin which shouldn't damage either the USB port nor the Pro Micro which is still isolated from RAW by the diode.

Ultimately it needs to run on external power alone without USB connected at all. During code development I would like to connect and disconnect USB as required, leaving external RAW power connected otherwise I won't be able to debug with the IDE's serial port monitor over USB. RAW external 12VDC is used to power the Arduino as well as other peripherals using the Arduino's onboard regulator. RAW power is DC and not AC since it's a wall-wart style PSU. Polarity was correct since RAW was already used to power peripheral and just the + jumpered to the RAW pin.

Since these are the 16MHz version, they need at least 4.5V for guaranteed operation. Therefore, logically, the regulator is a MIC5219-5.0V to generate 5VDC. It needs about 5.5V on input to work, plus diode drop of 0.2V (Schottky). I don't see how it would work on USB power alone. I measured 5.1V on UVCC and 4.9V on the other side of the diode as expected. Since I had solder-bridged SJ1, UVCC was bypassing the regulator and powering the whole circuit.

On one board I get 0.22V across D2 using diode mode on my DMM and 0.97V in the reverse. On the other board I also get 0.22V forward drop on D2 but 1.2V in the reverse. I unsoldered the bridges across SJ1 on both boards.

It seems the regulators may be blown as I get nothing on VCC with 12VDC applied to RAW input.

I've got a few more on order and don't want to fry them all as well. Plan of attack for new ones is to NOT solder-bridge SJ1 and try with just USB. I suspect they won't work as the regulator will not have enough to work with. Then I'll try with 12VDC on RAW and see if one of them survives. Maybe having SJ1 bridged caused the problem.

Additionally, reading the datasheet for the MIC5219 reveals that these are lightweight power handlers. In fact, I don't think they can handle 12VDC input at all as the power dissipation requirement goes beyond what the SOT-23-5 (M5) package is rated for. Most of the calculations on the datasheet show a paltry 6V input for >150mA operation! I think I should reconsider my design approach and feed the whimpy Arduino Pro Micro with an external regulated 5V VCC directly.

What have I missed? Any other thoughts?