| ??? 07/11/09 14:05 Read: times |
#167084 - You should, to find out what's wrong! Responding to: ???'s previous message |
Saman said:
the shunts are already 0.1% metal film resistors that are differing by a factor of 10. Finding exactly the same resistors need that I select them among 100-500 pieces. 100-500 0.1% can be very expensive, is this really necessary ? They were 0.1%. That's what's stamped on, but are you sure they are still 0.1%? This selecting is very simple: Take 1% metal film resistors from the same production (they are sitting on the same belt). Most of them are nearly identical. By fabricating these two reference shunts which differ by factor of ten you can eliminate errors from the original shunts and the switchbox. Then you see what error the ADC makes and the current source with different load. You can go even a step further and build a refernce circuit, by amplifying the signal running through the lower reference shunt by a factor of ten. Take another set of reference resistors in combination with a low offset precision OPamp to accomplish this. Then you see directly without the need of ADC (by the help of DMM) what the error of current source is when it sees a different load. The reading should be identical, if the current source is error free. Kai |
