I’m in the early design stages of a project that will use the Telit GM862 cellular module for remote data acquisition and control. The remote unit will receive commands via DTMF tones, and respond via synthesized voice. The audio input and output ports on the GM862 are differential, and I’m wondering if anyone has a couple of good circuits they’ve used to convert these ports to single-ended (ie. Gnd referenced) signals? Conceptually, I think I’ve got a pretty good handle on using an op-amp to convert the differential speaker output of the GM862 to a single-ended signal, but I’m scratching my head about how to go in the other direction with the microphone input, where a single-ended signal becomes a differential signal. Can I simply use a pair of op-amps for this - one a follower, and the other an inverter? Has anyone here done this yet?
This is a very good idea. There is a considerable amount of noise hooking up audio directly to the pins.
In addition to leon’s document, there are some pretty good audio interface examples (a bit over-engineered) in the telit hardware design guide. If you haven’t looked at this, its an absolute must read.
I have very good experiences with differencial signals in case of using speaker or mike together with GSM Module. The GSM signal is quite agresive and it can be demodulated on each PN junction quite easily. That is why we can often see 33pF capacitors behind transistors or protecting diodes in audio paths. I have tested several electret microphones and the right choice is also very important (there are some with embebded small capacitors also). The next significant factor is the PCB layout design. It doesnt have to be 6 layer PCB but there is needed to use ground around and below the whole audion circuits and PCB vias where possible. This ground shall be connected to the common ground on one point if possible. I have somewhere a nice document about all this from Siemens. In case anybody ask me I can find it and post.