GSM Modem Interference

[abuelmagd]

Posted on June 23, 2014 at 10:13

A GSM modem is in proximity to the uC. When the current bursts of the modem occur I get false logical 1 on my port pins. They are configured as analog inputs. I am using the STM32L100-DISCO board. I was wondering if the same behavior will happen for the microcontroller alone or if the evaluation board is the problem.

[Tesla DeLorean]

Posted on June 23, 2014 at 14:03

You can't read logic levels when the pins are in analogue input mode, there is no schmitt trigger enabled.

Is your power supply adequate? A USB based supply probably isn't.

[abuelmagd]

Posted on June 23, 2014 at 15:46

You are absolutely right clive. I do not read logical 1 using firmware. However, when I monitor the pin with an oscilloscope I can see it toggling between 0V and 3V (supply level). It seems as if I can read the frames sent wirelessly by the modem.

As a side note:In my application I have comparator 2 hooked up to one of the pins and comparing to Vref. I also have the comparator output connected to a timer to measure the pulse width. All this works flawlessly when the modem is off. During transmission I get random results and that's what drove me to look with a scope. All my input pins have this issue during transmission. At the moment I'm able to filter out the noise at that point in time but I prefer to clean up the noise from the source. I know mobile phones have arm processors in them so this noise doesn't make sense.

[Tesla DeLorean]

Posted on June 23, 2014 at 16:15

I'm using STM32 processors with GSM and CDMA modems without problems. This is with multi-layer boards with adequate power and ground planes, and without random wires floating around.

I'm pretty sure someone could build a mezzanine PCB to connect DISCO boards, and modem modules. You could partition the supplies. As I recall some modems have 2A instantaneous current requirements, so if the supply is inadequate it's going to impact ALL the attached circuitry when the modem transmits and the supply distorts.

[Andrew Neil]

Posted on June 23, 2014 at 20:11

''As I recall some modems have 2A instantaneous current requirements''

Indeed: that would be commonplace - not exceptional!

[Tesla DeLorean]

Posted on June 23, 2014 at 21:41

Quite probably, but it seems to catch enough people all the same. Usually complaining the POR circuit in one of my devices is resetting their whole board. Yeah, sorry about that, try keeping the supplies within the documented range, and the part where I say that NRESET is bidirectional.