Hi all,
we are using the ADC connected with strain gauges configured as a full weathstone bridge; ADC and the bridge are powered and connected to 3.3V and GND, as the reference pins, output signal of the bridge are connected to AIN0 and AIN1.
We tested our boards for several months, on a couple of prototypes; now we are starting mass production and we are meeting a strange behavior on about half of them.
ADC is programmed for continous conversion and cpu software execute an offset calibration every 10 minutes; occasionally we have a instantaneous drift of thousands counting on output data value, not directly connected with changes on the bridge, apparently at offset execution but not only.
There are always a couple of drifts of reverse sign, so what we can see graphically is a square wave on 2 stable levels, as follows.
on X axis the time, in this case 30 minutes, on Y axis ADC output value
A single drift can happen in various way, after a couple of minutes or a couple of hours.
We had a doubt about offset management, so we started reading offset register after offset calibration; values are quite stable and on a normal noise range of the ADC.
Please, can you give us an answer ASAP; we are scheduling a production up to 10000 pieces per year and now we are stalling.
thanks in advice
Andrea Gasparini
