ThePicMan wrote:

>>> (some devices I will monitor may be at a different ground potential
>>> than the PC).
>> 
>> I think this is the crucial point that you need to address before
>> deciding anything else. How much isolation do you want? And how do you
>> want to achieve this?

> What I decided then was to isolate the serial port, via an optocoupler
> (4N25?). 

> Now I'm still not sure if I should use a transformer or if I could use
> the DTR/RTS also to power an isolated DC-DC converter, like those nice
> ones made by Texas Instruments (e.g. DCR010505), but they may be too
> noisy, perhaps.

I don't think that you will see noise problems in the precision range you
need. (You wrote somewhere 1%.) As others wrote, you may have trouble with
different PCs. Depends on your requirements of inter-exchangeability
between PCs. 

So now that you have decided that you want your circuit on the ground level
of the input, you obviously don't want to connect your case to the PC
ground. Besides the serial connections and possibly the DC-DC converter
input, nothing in your circuit should be connected to anything coming from
the PC.

I think a common ground point near the input would be good. You connect the
input, the DC-DC converter ground and the circuit ground to it. Possibly
not the case -- I'm not sure you want a case that can be touched to be on
the same potential as a circuit under test. Depends on the potentials...

Gerhard
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist