| ??? 07/09/09 13:58 Read: times |
#166930 - More details needed... Responding to: ???'s previous message |
Saman said:
So for example if current changes linearly from 100uA toward 1A then I need to see a perfect line in collected data not a line having bumps when range changes. What "bumps" can you tolerate? Less than 0.1%? What is the maximum allowed shunt impedance for each range? Are you allowed to have different shunts for different ranges? Are you allowed to switch the shunts during the measurement when changing the ranges? (You know that this can result in total desaster in combination with large inductances when doing the switchings wrong...) You say the current source is a constant current source. Does it produce glitches when changing the setting? Is there any need for protecting your circuit? (I once had to fight with a kV power supply which produced narrow kV spikes each time I changed the output voltage, even when I only wanted about 100V from it...) Saman said:
My ADC is fast 16 bits with S/H but in practice I can't go better than 14 bits.
... Ok, data rate is about 10K sample/sec Why so many samples, when you have a constant current source? Why not using the dual-slope conversion technique like used in the ICL7106, e.g.? Kai |
