> Hi Sanjay, sorry I don't know where you were trained,
> but I was taught that to get good reliable measurements
> of "active" signals you need to allow 2% to 5% of the
> total for the measurement equipment.

I don't believe this, care to justify it?  Frankly this sounds like one of
those conclusions reached in a particular case where the answer was
remembered but the conditions governing the particular case were forgotten.

Suppose I showed you a wire and asked you to design a circuit to measure the
current through it up to +-100A with 8 bits accuracy and minimal
intrusiveness, and I wanted the measurement circuitry isolated with the
current information also passed back thru some sort of isolation.  Other
than the dielectric strenght of the voltage isolation, would the circuit
change any if I told you the wire could be up to 10V from ground?  What
about 100V, or 1KV?  Would you take 2-5% of the voltage in each case?  Of
course not since the voltage doesn't really enter into the design except for
the isolation strength.  You'd use the minimum sense resistor so that under
worst case conditions you still have a good 8 bits, which has nothing at all
to do with the voltage on the wire.

> You mention a
> sense resistor of 1 milliohm, this is ridiculous.

I wouldn't be so fast to say that.  With 1mohm you get +-100mV max signal.
That is enough to maintain a signal to noise ratio of 50dB (8 bits) after
amplification to a convenient level for an A/D, like 0-5V.  That only
requires a gain of 25.  It would be tough to maintain this accuracy for DC
due to offset voltages, but the problem specifically said audio going to a
speaker.  In that case things can be AC coupled with rolloff around 10Hz or
so, thereby eliminating offset voltages as a problem.

> I have
> a good milliohm meter, and any electrical connnection will
> give you a few milliohms resistance. Your sense resistor
> will have two of these connections.

True, but there won't be any current folowing thru the sense leads, or what =89=tle current there is will still be proportional to the current you are
trying to measure.  I agree that the "resistor" is just a length of wire,
preferably one that needs to be there anyway.  And at 100A you need to be
extra squeaky clean about good connections.


********************************************************************
Olin Lathrop, embedded systems consultant in Littleton Massachusetts
(978) 742-9014, olin@embedinc.com, http://www.embedinc.com

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body