Neutral to Ground Voltage on ADC

[electronerd213]

Not at all, it's just got protection diodes like a PIC - so, presumably, (just like a PIC) you simply feed the pin via a current limiting series resistor (value depending on the maximum expected voltage) - and this could actually be part of the divider that feeds the input.

So I think Nigel is right in that I COULD apply voltages outside of that range to the signal inputs, but even if my differential on those inputs is, say 1 volt, but Vin- referenced to supply- is 10 volts (making Vin+ 11 volts), then the reading will hit the rail. Therefore, there is no difference between the outputs where the inputs are 10 & 11 volts or 10 & 13 volts. They are both railed due to the difference in "reference" between the signal inputs and the supply V-. I think we are all agreeing on that concept, right?

What exactly are you stuck on?
Theory or real world [NOT WORKING]?

Please see schematic below. At first, I did NOT include the pink circled ground reference. I had difficulty reading anything as the output looked like a bunch of differently amplified dirac delta functions (pulses). with the main level at 0. So I added the pink circled ground reference, and now it mostly works (though I am having trouble matching the datasheet gains, and I have contacted TI for help on that one). The problem is, I feel adding the pink ground works great under lab conditions, but under the 1000V Vng design spec, V(R41) is approximately 500V, giving my theoretically isolated dc supply (I have since found that it may not actually be isolated after all) a 500V potential from earth ground. I would fear blowing up a trace or accidentally passing current from that 500V potential through a chip's ground pins. ??? Is that valid? I just see red flags.

BTW, this design is basically what Strantor outlined but for an amp differential input of 3.3V at 1000V oops condition, in case you are curious.

So I think I am back to wishing there was a chip that could do this function but without really referencing or limiting reference to the supply. I am nearing having to add another isolated DC to DC converter. uggg...

Do you think maybe a DMM actually connects battery(-) to the input similar to the pink circle?

 

[ronsimpson]

Question;
I can see that voltage applied to the wire at the top left side of the schematic, passes through a voltage divider R40, R11.
Voltage is also applied to the wire labeled "GND". GND does not equal DGND. I don't see what R41 does!

I used a diff amp ahead of the ADC. So "N" could have any voltage on it.

 

[electronerd213]

Question;
I can see that voltage applied to the wire at the top left side of the schematic, passes through a voltage divider R40, R11.
Voltage is also applied to the wire labeled "GND". GND does not equal DGND. I don't see what R41 does!

Well, the original purpose of this is to measure the potential between N and GND. DGND should not even be connected under the pink circle. The problem is that I need to measure N to GND, but the ADC is referenced to DGND. So R41 is supposed to balance the voltage divider in a scenario where my measurement circuit (again, remove the pink circle DGND) is not referenced to N or GND but has to measure the difference between them. So you end up measuring a small voltage across the middle resistor that is 9Mohms away from either N or GND. Again, the trick is that the pink DGND was never intended to be there in the pink circle.

I used a diff amp ahead of the ADC. So "N" could have any voltage on it.

Do you know of a diff amp that can measure the difference between N and GND knowing that DGND (powering only the ADC side of the amp, not the N-GND measurement side) could be well above or well below the potential at Vin-? Thanks!

Edit: I guess in similar words, I am looking for the case where "N" could have any voltage on it (with respect to DGND) and "GND" could also have any voltage on it (with respect to DGND), and I do not connect DGND to the inputs of the amp/chip. The voltage divider takes care of the case where the difference between N and GND is very large, but what this circuit does not do yet is reject a very large or potentially negative common mode voltage with respect to DGND so that I am truly measuring a difference of the inputs regardless of how far those inputs are away from DGND.