On Mon, 30 Jul 2001, Sanjay Punjab wrote:

> 1. Use a 1:1 transformer to isolate the voltage across
> the low-ohm resistor (current sensor), from the
> interfacing analog circuitry.

For me this have not too much sense if you could built yourself the
transformer. How many pieces do you need ( more than 100 ? )
Why a resistor and a transformer when you could just a transformer ?


 On the secondary side,
> the voltage can be referenced to ground, making it
> quite easy to get reasonable accuracy out of an
> inverting op-amp circuit with .1% resistors.
>  This will work, since I am only concerned with AC
> current. The problem is finding a 1:1 transformer with
> truly accurate winding ratios (within 1%).

Why you need 1% accuracy if you will use an operational amplifier which
gain could be slighty ajusted ? Using a trimmer you could use 5%...10%
accuracy transformers quite well.


 Or else, I
> will have to perform some type of testing in each unit
> during manufacturing and come up with a normalization
> look-up table in software.

No, a simple fixed signal generation ( by pressing two buttons known only
by yourself in testing mode ) and one potentiometer adjustment is cheapest
than a variable table read

>
> 2. Use some type of optocoupler. Again the same
> problems as solution #1. I am wondering how difficult
> it would be to get an monolithic optocoupler
> that has consistant voltage/current transfer ratios
> between device samples.

  Will be expensive. Search at Burr-Brown for example and you'll like to
kill yourself. You'll need 2 common optocouplers ( photodiode +
phototransistor ) with at least 20 KHz bandwith and 2 common
operational amplifiers. The problem will be another floating power supply.
This version is not doable for large production bacause is much expensive
like the first one.

Cheers, Vasile

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads