On 2019-11-17 04:11, Michael Balzer wrote:
I also saw this & can also reproduce this on my workbench.
What is the method to induce the problem?
Mark, it seems we've got a hardware flaw here: EGPIO output port 0 (MDM_EN) is connected to transistor Q2 base via R16 without any pull-up. As the EGPIO ports are open drain, that can only work by chance. Maybe R16 was meant to be the pull-up, i.e. be connected to +3.3V, and MDM_EN should be connected directly to the transistor base?
I've verified this by checking the voltage level at MDM_EN & PWRKEY. When setting the port to 1, MDM_EN only rises for a very short time (too short for my multimeter to tell), and PWRKEY only drops very briefly. The modem needs PWRKEY low for at least 180ms to power up and at least 500ms to power down.
Manually pushing S1 reliably switches the modem on/off.
Maybe a fix for existing boards could be adding a pullup resistor at the transistor base?
I could try this if you suggest a value (and perhaps mark up a schematic?) Craig