Capacative measurement on high voltage powerline.

[7kasper]

Hello fellow engineers!

I am trying to do a capacitive measurement on a conductor.
In simple form I basically apply force a small voltage on the conductor and measure the same output voltage. Based on the change in rise time I can detect a (significant) change in capacitance loading the conductor. This seems to work just fine.

Now comes the problem: the conductor I do my measurements on will sometimes have around 6kV on it (with respect to ground).
The measurement device doesn't have to measure when the high voltage is on, but the key point here is that it just musn't break.


I have thought of multiple solutions. These are the main ones I came up with:
1. Completely seperate the measurement circuit.
Basically put one or two relais after the measurment circuit. Only make electricty to be able to flow back if the HV line is off.
Problem with this is that you need to have a relais that seperates >5kV and I haven't found something that isn't super expensive.

2. Operate measurement at relative or floating point.
So the voltage doesn't want to go through the measurement circuit. This is quite impossible though. You need your capacitive measurement with respect to something right? Also there will not always be 5kV on the line. It can also be 0v. So I cannot have the plus of the measurement circuit be at 6kV + some test voltage.

3. Diode seperation.
I believe there are some diodes with capable reverse blocking voltages. I basically put them in line with the capacity measurement circuit.
If the HV is applied the diode should stop the reverse current into the measurement circuit.

Questions:
Method 3 seems most promising to me. But can it work? I guess the capacity will change somewhat because of the added diode that may or may not have wierd effects. But perhaps the relative measurment should still work, do you aggree?

Second of all, is this a sensible approach for this problem.
And lasly: Can this method be adapted if the 6kV signal was AC instead of DC?

Of course I am open for all (other) ideas to perform such measurement.

Thanks a lot for your effort and creativity!

- Kasper

[bob91343]

I think the most sensible approach is to disconnect the measuring apparatus when not actually measuring.  The diode approach might work if you provide a path for its leakage current and can tolerate the transient that occurs when the 6 kV is applied.  If application of the high voltage is sudden, capacitive coupling around the diode may destroy your measurement device.

[EPAIII]

What about capacitive coupling the pulse into and out of the conductor. Use some form of over-Voltage protection on the test circuit side of the capacitors to protect that circuit when the HV is switched on or off. Keep your test pulse below the level where that protection becomes active. Your test circuits would only need to withstand that level where the over-Voltage kicks in.

You did not mention any frequency so I do not know how this would work when the HV is AC. Power line frequencies (50 or 60 Hz) would probably be OK but higher frequencies may not.

What over-Voltage devices? Zeners, ordinary diodes in series, or any of many other circuits would be candidates.

[PartialDischarge]

I'm curious, is this for a university assignment or a project for a company you are working with?

[asdf336]

Yes AC coupling is the most promising method.   The known capacitance in your device can be backed out with calibration/calculation.

A remaining problem is the inrush event if you accidentally connect your device to such a high voltage.  Hard to know the scope of that problem without knowing more about the C you’re trying to measure (which impacts the series C you’d choose). 

There are few other series components rated for such voltages though you could reference high voltage multimeters which have 1kv type protections on their leads. 

[741]

I do not know the fastest rise-time of the 6kV source - is the right-hand side (RHS) of the attached image correct?

But if things are capacitively coupled (as suggested in #2) and also current is a worry, maybe add an inductor as shown in the image?

This is just a vague idea, likely it has problems...


Out of interest, why is it useful to measure the capacitance - how is this information to be used?

[7kasper]

Thanks for all your answers!

I can perform the rise time measurements in the future. I think they are quite quick.
Frequency would be just a couple of hertz though. (<20Hz) Its more like there are a few HV pulses on the line.

The capacitive measurement is a safety check. If anything weird is touching the conductor the HV musn't be applied.
(of course this is not the only safety measure but Its something I want to try out).
Note that in my setup its not really possible to completely remove the measurement setup or HV source.
The circuit by 741 seems to be roughly what I am talking about.



Measuring.
Of course the rough capacitance to be measured and the rise time of the HV lines are useful details. Its just that they are a bit hard to measure and I don't want to blow up my scope for instance. Could I just rig a voltage-devider on my probe and be fine measuring 6kVs?

Through the (known) capacitive and inductive coupling I can still measure capacitance of the line while protecting against inrush current and high voltage peaks. The diodes are only required when the voltage stays on (DC). Is this correct? @741

[CaptDon]

Measuring the capacitance to indicate a potential safety issue of a 6kv line is a total fail in so many ways. More likely way is like the HiPot testers in common use to test leakage for U.L./C.E. type certifications. Slowly ramp up the voltage to the desired max or even more to measure a safety margin above your expected operating voltage and measure leakage current. Your capacitive time delay charge constant measurement is too inaccurate. Is this line going to be exposed to weather? Is it an outdoor transmission line? How many KW of power are we talking here?? What do you expect to compromise your High Voltage line?? Thieves tapping in to it?? Is there a load on the other end??? Whole thing sounds like a bogus science project some EE professor dreamed up to see how his/her students would solve the issue. How about tell us what this H.V. line does or is used for, maybe (probably) there are some off the shelf solutions???
 

[7kasper]

Hey CaptDon.
Thanks for your enthuasiatic comment. I could totally see this being the case 
The real story is a bit different, but perhaps just as dumb?
I don't know. I'm just trying to make this thing work with my limited knowledge.

Basically the project is more like a smart electric fence. Here shocks of serveral kVs (but with low current) are given off to deter for instance cattle. In this case I want to see if I can perform measurements on a small sort of electric fence and see if something is off or perhaps even measure the sort of wildlife touching the fence.

There might or might not be a load at the end. And I don't think there will be kWs of power. Just short HV bursts.

[geggi1]

I beleve the most common way to do this is by making a capacative voltage divider.
On HV/MV switchgear the insulated conductor is running through a metallic sleeve.
The conductor acts as one of the plates in the capacitor and the metallic sleve as the other.
Then there is a resistor or a capacitor that works as the other side of the voltage divider.
The primary voltage is normally from 10kV and up.
Normally the design is so that your secondary (output) voltage is 100V.
You have to include the internal resistance of your voltmeter when designing your measuring setup.
Remember this design is also frequency dependant.

You can also do this with resistors and you will have almost no influence of the frequency up in the several 100kHz range.

For lower voltages 10kV and under its also common to just use transformers. These will normally give good readings form about 40Hz and up to somewhere close to 70Hz.

[CaptDon]

I had an F15T8 flourescent lamp, the body of which was electrical taped to a ReBar rod in the ground. The bulb stuck about 4 inches higher than the rebar. I had a wire soldered to one of the bi-pins on the end of the bulb sticking above the rebar. That wire was tied to the hot wire of the electric fence. Every night I could look out the kitchen window and see the light flashing. It didn't load down the fence charger and worked even in rain and snow. If the light didn't flash I knew the weeds had shorted it or the horse broke part of it down. The light flashed due to the capacitive coupling of the body to the rebar. DO NOT ground the lower end, that will pull the fence charger output down!! B.T.W., even a dead lamp will flash!!

[Vovk_Z]

Capacitive dividers are used on 50/60 Hz power lines. But they are not good is frequency goes from DC, as tansformers and inductors too.
As for me 6 kV isn't very high voltage and we have almost DC so usual resistive divider has to suite this purpose fine.

[Terry Bites]

Measure the transient current in the ground return to your psu. You dont need to connect to HV