G120 SOM based board not being recognized by Windows

I’ve got a custom mainboard based on the G120 SOM that I baked this weekend. The board powers up but is never recognized by Windows. I used the Cobra II schematic as a basis (pertinent section of mine is shown below). What would keep it from being recognized by Windows? This is my first board using the G120, so I must be doing something wrong. Any help is appreciated…



Hi Mike

In your drawing, what is header 4 (USB) connected to?

On a USB header, pin 1 is USB VCC, pin 2 is D- and pin 3 is D+ and 4 is GND. You seem to have this reversed but could be valid depending on what that part is or is connected to?

Dave, I’ve got the header connected to a USB connector (I fabricated it by simply cutting off one end of a USB cable). I have tried reversing pins 2 and 3 to no avail. I have also tried 2 separate cables

Your connector should be fine. I’ve done this myself without any issues.

How are the tracks laid out on the board? Short or long and differential pair?

It sounds like the G120 is not coming out of RESET. How does the reset line look when you probe it?

Dave, Tracks are probably 2 to 3 inches. I’ve also soldered directly to the SOM and it makes no difference. The RESET line is normally ~1.4v and goes to 0v when I press the RESET button.

I just realized something, shouldn’t the 1.4v on the RESET line be greater than 2.0v?

Should be close to 3.3V. Your processor is not coming out of reset.

So in future, it would be nice if the modules came preprogrammed with a heartbeat signal (like flashing an led) so that you know the module is working after it’s baked. Simply putting an led between heart beat pin and gnd would let one know.

@ Dave, did anyone ever make a bed of nails for the G120?

1 Like

I would assume that GHI do this sort of testing off the production line?

I’ve never had any issue with a G120 other than with a slow power up supply and had to hold the device in reset longer. I now use the MCP130 as a belt and braces reset for it.

@ Dave - Thanks, will keep this in mind if I still have issues after getting the voltage right.

@ Dave - I am assuming it is this part, MCP130T-315I/TTCT-ND, right?

Hm, you are using SPI2 to interface a LCD… SPI2 is the interface that is connected to the SoM-RAM, maybe there is a problem?

So, the problem turned out to be the WiFi module. Since it was also wired to the RESET line, and the chip was not powered (I had not installed the jumper to power it) I guess the RESET line was floating a keeping the MCU from coming out of reset mode.

Thanks for everyone’s help!



@ munderhill - So this was a Design error at the schematic capture phase?

@ Mr. John Smith - No, it was a fabrication error. I did not populate all of the headers on the board during the initial fabrication so that I would not waste parts should the boards/design prove to be non-viable. In hindsight, even if I had populated the header I am not sure that I would have installed the jumper on it.