110V DC Isolation

[Archangel]

Hi,

Thanks for the reply,

Yes I am wanting to use it on RA0. The voltage divider connects to the +110V DC line and ground (I am using 33K and 1K) the voltage measurement of 1-4.335V is the output I am getting from the voltage dvider when I vary the voltage from 20-110VDC. I am just looking at providing isolation between the PIC and 110V DC as I am scared of a nasty suprise. We have also tried using a variac to simulate the 110V DC using the same voltage divider but we have not tried that either on the PIC as we have not isolated it.

Are you trying to isolate for a final installation or for test purposes? <br>I think I would move those divider values up a couple of magnitudes from the "K" area to the "MEG" area, just for safety sake. If this is for a finished purpose I would isolate the switches etc using opto isolators and put the hardware out of human reach, I. E. in a box . . .

[skimask]

Are you trying to isolate for a final installation or for test purposes? <br>I think I would move those divider values up a couple of magnitudes from the "K" area to the "MEG" area, just for safety sake. If this is for a finished purpose I would isolate the switches etc using opto isolators and put the hardware out of human reach, I. E. in a box . . .

If you move into the "MEG" area, you'll have too much source impedance to feed the A/D input on the PIC! Unless of course you throw an op-amp set up as 1:1 unity gain voltage follower thingy...

[ErnieM]

The impedance of the resistor divider is approximately the same as the resistor from analog input to ground; this resistor is in parallel with the other as far as impedance goes (look up Thévenin equivalent for the reason why). Since one resistor is much larger then the other, the smaller one predominates.

Isolation may mean many things. I think a more useful term here would be "Protection." You want your PIC to survive if something goes wrong, something to prevent 100VDC from being applied directly to a PIC pin.

I agree you need this, having done such myself. The PIC will not survive. However, most of the pieces stay on the table for a further failure analysis.

I would do two things: Put a zener to ground or a diode to VCC at the input to work as a clamp. Plus, I would split the resistor from 100V to PIC into 2 or three resistors. Resistors tend to fail open, but probes and such can short them anyway. By putting another resistors in series you can still short one out and have the next one provide some impedance for the protection diode to use. Without such impedance the protection diode will open followed shortly by the PIC.