Quote:
Originally Posted by unregistered
Groth
I understand the need/benefit of the optical isolation (std in other things),
but how does that resolve the floating ground issue, or is it then moot ?
|
Spoon fed, eh?
The isolation is of course only an issue if you want to measure the voltages of a target system from another PC, the temperature reading circuitry is internally isolated.
Powering the reader IC from the target computer eliminates the ground float related to current return though the PSU power cords, plug, and some wall wiring (if you don't have your two computers plugged into the the same socket). You still have the potential for float/noise from things like the ATX connector, long strings of molex connectors, wimpy power planes, etc.
My prime concern was the CPU power. Accurate measurement of Vcore was more important that accuracy of 12V, or Vdimm, or any others. So I grounded the reader IC at the back of the CPU socket. It's ground will float relative to others' due to the 60+ amps the CPU can dump into the ground plane, but it's ground will match the CPU's ground. Not allowing for that float overestimates Vcore and CPU power.
So, yeah, the bus optoisolator resolves between-computer ground float/noise. And it moots local float/noise if you ground it to match your measurement of interest.
MMZ_TimeLord, I used Digikey part number PS8701-ND (I like SMD), but TLP550-ND would work just as well. I'm sure there are others, they just have to be faster than around a microsecond, to keep the bus from timing out.