Op Amp? Help -/+75V to 0-5V ADC

Need some help here.. I'm lost... Total Op Amp (or whatever I need..) newbie..

The attached image shows the main power supply to my circuit as being 2 x -/+75VDC sources (voltage on either supply at any time could be from -75V to +75V)
I'm tying those supplies together to create ORD_Vout to feed a 12V SMPS to power the rest of my circuit. And planned to feed the 12V into another SMPS to generate 5V if needed by this op amp or whatever circuit.. I could also use the 5V into a 2.5V series reference if needed by the op amps or whatever.

So basically..
I have 2 ADC's (part of a 2 input digital voltmeter) that are being used to monitor/display the voltage of each main -/+75V supplies.
So I need to generate a 0-5V signal for each of the 2 ADC's where
approx. 2.5V input to the ADC = 0V supply voltage
approx. 5V input to the ADC = +75V supply voltage
approx. 0V input to the ADC = -75V supply voltage

 

You just need a 1:30 voltage divider, since you need to squeeze 150 volt swing into a 5v package. So that's the span.

The other issue is ground, the offset. Do you need the 0V output for a -75v supply to be at the same "ground" as your power supply ground?

 

I don't believe so.. That's sort of where I was getting confused..
It should be 0V with respect to the ground off the full wave bridge.

Right now I have the "ground" for the output of the ADC/12V SMPS, etc.. will be the "ground" point taken off the full wave bridge.
I basically think of it as having 2 power supplies that get diode or'd to create this "new" supply that will always be positive and is created by ORD V_out and the 2 ground points at the output of the full wave bridge rectifier part. But then I need to measure the 2 "old" supplies to determine their voltage.

 

See I started with this circuit..(Which as the 1:30 divider) Which worked just fine.. But then once I threw the second supply and the full wave bridge for OR'ing and reverse polarity protection and needing to measure each of the supplies I started to get lost..
I wasn't sure what my ground became.. or which ground was what

 

Basically I have this dual input digital voltmeter (actually a prebuilt unit composed of a digital display and a PIC micro..) Now this voltmeter needs a 12V supply.. So I started by creating the supply as shown in my first post and the output of that (ORD_Vout and ground off FWB) feeds a SMPS (LM5008) to generate the 12V needed to power the voltmeter. And now I need to measure the voltage of each of the -/+75V supplies and feed that signal to the 2 x ADC ports on the PIC.. The PIC ADC's are expecting 0-5V which through the software is scaled to -/+75VDC.

In actuality my -/+75V supplies will actually never go below -/+19VDC but the software is setup to expect -/+75V scaled to 0-5V

 

The grounds all need to be the same (connected together) to measure the voltage.

Your circuit shows an attenuation of 3/103 = 1:34.3 not 1:30.

Edit: Remove incorrect statement about gain...

100kΩ on the output of the op amp may be a little high. What's the input impedance of the ADC?

 

The PIC on the voltmeter module is a PIC16F88 (datasheet says max impedance of 10K ohm)

 

So here.. This is where I'm stuck.. V(adc1) seems correct (+19 to 75V battery output = approx 3-4.8V op amp output)

But its when its a negative input voltage (V4 and ADC2 section) that it all goes wrong.. Just a 2.5V straight line.. (The 2.5V reference I suppose)

 

The power supplies need to be connected to ground, as I noted in my previous post.

 

Oh wait........ Did I just figure it out? I think so.. Just did a dance in my chair.. when I saw that aqua line drop/mirror the red line.
See attached.. Is this correct?
Are the resistor values ok? (100K/3.3K/10K)
Any need for additional capacitors,etc.. or anything else.
Voltages (PULSE setting in LTSpice) start at 19V..1/2 second delay..then ramp up to 75V and delay again.. Sure looks ok..
Please tell me its perfect.... And again please let me know if there are any other things I should add to make this a "professional" circuit.

 

If the PIC has a 10K input impedance than the 10K at the op amp output would reduce the voltage by 1/2. I don't think that's what you want.

You should add a 0.1μF ceramic decoupling capacitor directly across the plus and minus power pins on the op amp.

Other than that it looks good. (I misspoke in my previous post about needing to remove R2 and R3. That was incorrect.)