High Impedance, High Voltage measuring circuit

Status
Not open for further replies.
J

Jester

Full Member level 6
Joined
Aug 18, 2012
Messages
377
Helped
7
Reputation
14
Reaction score
7
Trophy points
1,298
Location
.
Visit site
Activity points
4,754
I was working on a circuit the other day that measures ac/dc voltage up to 1kV. The input impedance of the circuit is ~2MΩ , so simply connecting any of my DMM's (10MΩ input impedance) loads the circuit considerably. I do have a 34401A DMM that has a GΩ input mode however this high impedance mode can only be enabled for low voltage ranges.

So I'm thinking about putting together a circuit to place between the circuit to be measured and a DMM for the next time I need to make a high voltage (up to 1kV) measurement in an impedance sensitive circuit.

Thoughts on the circuit below?

R1,R2 are 1%, rated for 3kV working voltage
R5, R6 are to match the gain, I would hand pick them


OP07, is not great with lower input voltages, perhaps an alternate with lower input current OPA191?, suggestions?


 
Last edited:

Should work, but no quantitative specification for offset, drift, CMRR given so we can't decide about component requirements.

In the present circuit, OP7 input current induced offset and -drift is in the same order of magnitude as PGA103 caused errors, switching to higher precision OPA191 doesn't raise the overall performance so much. You would need a better PGA stage to take full advantage of OPA191 quality.
….
Sorry, wrongly calculated. OP07 input current is an issue, go for OPA191 or a similar high precision FET OP.
 
Reactions: Jester

    J

    Jester

    Points: 2
    Helpful Answer Positive Rating
Hi,

The input impedance of the circuit is ~2MΩ
Click to expand...
I rather think this is the "source impedance" or "output impedance" of your circuit.

If your source impedance is about 2MOhms and you want 1% error (for example) you need at least 200MOhms of input impedance.
I see you have 2 x 1GOhms now plus the TBD values.
I recommend to try to go as low ohmic as possible.

OP07 say max input offset/bias current is +/-4nA.
Multiply it with the feedback resistor of 10M and get +/- 40mV output referenced error.
Multiply it with the input resistor of 1G and 4V input referenced error.

This sounds good. But take care: every soldering, soldering and flux residuals, PCB leakage current, bad ESD treatment may immediatly increase the error to multiples of it.

****
Now you have two high impedance inputs, you need carful resistor selection to get good CMRR.
Maybe you can connect both GND of HV signal and measurement circuit, ... then you need just a single inverted or noninverted circuit....without the CMRR problem.

****
1kV input will result in 10V OP07_ouptut signal. I asume the OP07 is at it´s limit to go that close to the supply rail.

***
I recommend to add equal capacitors across R3 and R4 as LPF and to stabilize OP07 (in case of higher input and wiring capacitance).

***
Btw: you may use passive 1G/10M voltage divider combined with HP34401A high impedance input mode. At least you can try it.

Klaus
 
Reactions: Jester

    J

    Jester

    Points: 2
    Helpful Answer Positive Rating


All good points, thank you.

If I build it I will use 15V rails for some headroom for 1kV

The simple voltage divider is worth a try, eliminates bias and offset errors.

I think I will layout both and send in with next PCB order and then compare.

Suggestions for capacitors for R3, R4?

I have no way of measuring 1GΩ resistors directly (for matching) so I guess I will just apply 10V and measure the midpoint while tweaking R5/R6 for exactly 5.00000 volts.

I will post proposed layout when I get around to it to see if it looks reasonable.
 

Hi,

Suggestions for capacitors for R3, R4?
Click to expand...
Depends on signal frequency (how fast you expect the HV to change), or measurement rate.

I assume a 10ms time constant could be useful. --> 1nF. Use a foil capacitor to get low leakage.

****
You need to carefully select the HV resistors to keep safety. I recommend to use at least two HV resistors (each rated at least for full expectable voltage) in series, in a way that in case one resistor accidentllay gets short circuit the other limits the current to a safe value.

I have no way of measuring 1GΩ resistors directly (for matching) so I guess I will just apply 10V and measure the midpoint while tweaking R5/R6 for exactly 5.00000 volts.
Click to expand...
Here you have the same measurement problem. Thus I recommend to measure the midpoint twice: From GND side as well as from 10V side.

Another aproach:
Use one fixed resistor and one pot (pot value = 2 x fixed value)
Then use the setup like in schematic of post#1. Then use alcohol to clean the whole PCB, soldering....

Connect both inputs. Apply AC voltage to both inputs (common mode) referenced to measurement_GND.
Now adjust the pot to get zero AC output at OP07.

Klaus
 
Reactions: Jester

    J

    Jester

    Points: 2
    Helpful Answer Positive Rating
Why not just build a High Impedance Voltage divider to go directly into your Multi-Meter?
Where is Southside?
I have lots of 500 Meg and 2 Gig, HV resistors.
Connected in Series with a Pot, you can calibrate them to read accurately on your meter.
 
Reactions: Jester

    J

    Jester

    Points: 2
    Helpful Answer Positive Rating
Hi,

newbie here on practical circuits, I'd like to ask out of curiosity, The circuit has 2 Mohms output impedance at 1kV. The measuring devices are (a) a low voltage, high impedance input and (b) high voltage, relatively low impedance input. (b) causes loading so a circuit is added in between in order to used (a).

My question is, can we use a voltage follower (high to low impedance) so that we can use (b)?
 

Hi,

The circuit has 2 Mohms output impedance at 1kV.
Click to expand...
I'm not sure which circuit you are talking about.

1G/10M voltage divider has
* 1G + 10M = 1010M input resistance
* 1G || 10M = 9.901M output resistance
Not only at 1kV... it is independent of voltage.

****
You are correct: You may use a voltage follower to use a low impedance measurement tool.

****

The difference amplifier circuit has
* On the non inverting input 1010Mohmsimpedance referenced to GND
* on the inverting input 1000Mohms referenced to 1/100 of the non inverting input voltage.

This is unsymmetric and may cause problems with some applications.
The improved circuit without this drawback is called "instrumentation amplifier circuit".

Klaus
 

Status
Not open for further replies.

Similar threads

Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…