I just connected this week my MKR GSM 1400 onto my Edge Control socket 2. Beautifull. Until you push the reset button on either. If you do, Edge Control and MKR freeze forever.
My scenario:
No auxiliary battery connected to MKR
Auxiliary connected at Edge Control
Edge Control is powered via BAT pins
If the MKR is not connected, I can do whatever I want with the Edge Control It wold reset at the press of the button, remain calm when battery is lost thanks to the auxiliary, and everything in between. Then I power off the MKR rail, insert the MKR, power on the MKR rail, and everything runs peachy. Until I want to reset my board, and everything freezes.
It seems the RESET pit at the MKR socket is wired straight to the RESET button on both MKR and EdgeControl boards. If I press one, I am pressing them all. And they stay there because when there is no power the RESET pin at the MKR keeps the lane down. Or so it seems.
I bent the pin before inserting the MKR into the socket again. And voila. Works on resets. The MKR just looses power instead of reacting to the reset button from Edge Control. I can still reset MKR CPU with the reset button on its board (even though it doesn't really reset at all the modem).
So the question is: Am I right? Got lucky? Am I breaking something?
Please, do not suggest to use the backup battery connected to the JST pin on the MKR. I know word is around you MUST use it if you want GSM to work properly. Truth is I only needed it when used the PMIC library. If I did not use the library, my connections is perfectly fine at UMTS all the time. For years. Really, years. It sometimes freezes over, but the reset cause is never low intensity available as it was when the library was used. And, I want to be able to keep the MKR fully powered off. So, no battery. Please.
I expect the mkr to turn on its built in green led and enumerate the usb serial on my computer. Also the edge control to enumerate the serial on my computer too.
But they never do. No matter how much i wait.
Remove mkr to boot from power on and it works. Remove the reset pin between mkr and edge control, and it works too.
Gemini says all resets (mkr1 socket, mkr2 and edge control button) share the same lane. So when i push the button, the lack of power on the mkr rail keeps it forever pressed, neither board ever reaching bootloader.
All mt tests proved safe to avoid the connection of the reset pin. But i am an amateur.
Feels like a hardware defect. Design or local. It seems to affect mkr nb 1500 too acording to the IA. But the source is obviously not trustworthy, hence my question here.
Feels like you are upset because I did not do something expected.
I did not provide code from the begining because there was none. No drawing because it was one board plugged to the other. I consulted Gemini because I did not understand why it happened and tried my best before coming here to bother anyone. Just like I used the search tool just in case someone had the same issue. I am an amateur and help is always welcome even if I have to fall back to unreliable sources.
I am sorry if I did anything wrong. Thank you for your time anyway. I hope the extra data you requested helps someone else spend some time reading about this weird issue.
I do not understand much about electronics, but, if the schematic from the MKR at the top and the one from edgeControl at the bottom interconnect those lanes, why would they stay DOWN if +3V3 at MKR is zero because the rail does not provide power? Could it be because the capacitor C2 at MKR discharges through R6 into GND at EdgeControl and never rises? Shouldn't it rise from the EdgeControl +3V3 and go though to the MKR board isolated from GND by C2 (real GND) and R5 (+3V3 without power, perhaps GNDd) at MKR?
My mistske then. I thought the power led was lit before anything else come into working order, including an sketch that makes the BUILTIN_LED blink (orange) or start the Serial object to connect to console. My tests were all run with scketches that open Serial first thing so I can see the traces and the USB never enumerated, so I assumed it never run setup().
Will load the blink test only (MKR only for now since I cannot find an integrated led on EdgeControl to use) and come back. Thanks!
Test done. I uploaded the blink to MKR and left EdgeControl with my main scketch that powers on MKR very soon, right after Serial console is initialized.
The MKR board never blinked after a 5 minute wait. The ON green led by MKR usb port never lit either. Consistent with previous tests. As soon as I removed the MKR from the socket, EdgeControl run setup and I could see the traces on the serial interface. I turn off the power rail on MKR, insert it, power on and MKR boots happily into blinking forever (with or withour RST connected to the socket).
Then I reapeated without the RST pin connected and both booted inmediately after I pushed the reset button at EdgeControl.
I am going to attach the debug probe to the Edge Control debug interface I just discovered and see what it tells when stuck. Hopefully is not the pull-down effect and the board is actually running something it doesn't like.
