AC Potential divider problem

Dear all,
I want to step down a 500V(MAX)/50Hz AC waveform into 3.3V AC signal. In my circuit i can't use a transformer for this process. After that i will rectify the AC into DC by using a precision rectifier.
While using a resistive potential divider i can able to step down the voltage as required. But if i use a precision rectifier along with the potential divider, the negative output of the potential divider is getting reduced & the positive remains unchanged. The same thing happens for simulation & real hardware also...
For testing purpose I'm using a (0-15)V/1A transformer with R1=47K & R2=10K. But the potential divider output & precision rectifier output is giving negative voltage with reduced amplitude.

(For precision rectifier, I'm using the reference circuit from "burr brown" single supply circuit)

Please help me regarding this...

You missed the simple fact, that the burr brown circuit is expecting a low (<< 10k) source impedance. You can however increase R3 to compensate for the voltage drop in inverting mode.

Short answer: You need to increase R3 to about 27K.

Long answer: When Vpd is negative, the input pins of U1 are at ground potential so Vpd is reduced due to the current flowing through R1 and R2.

If Vpd were provided by a perfect voltage source, there would be no problem. However in your circuit the source impedance is R5||R4. That forms a voltage divider with R1||R2 for negative inputs.

For positive inputs, the voltage at U1's input pins is the same as Vpd, so there is no voltage divider action due to R1 and R2.

Thanks for the solution. Yes after i increase the R3 value to 27K the amplitude of the waveform is almost equal to the positive.

Can i solve this problem by using a split power supply? I need to measure true RMS voltage & zero crossing of the AC waveform by using a high speed ADC (10 MSpS at the least).
1. Can anyone recommend a very good precision rectifier.
2. May i use the same circuit or with the split supply version?
3. Is it enough to take only a positive half cycle for measurement (precision half wave rectifier)?

Regards
Udhay

1. Here's an article that may help you pick the best one.

2. Whether you need a split supply depends upon the circuit you use.

3. You will get accurate results using a half-wave precision rectifier only if the positive and negative halves of the waveform are perfectly symmetrical.