Meter for Low Current but High Voltage Advice?

[strokebow]

Hello I'm trying to measure dc current.
Within the given circuit I have high voltage (maximum 5 kV) but typically low current. Maximum: 10s of mA. Typical: 0.1 - 1 mA

Any ideas or suggestions would be most welcome?

Thank you

[MosherIV]

Hi

Welcom to the EEV forum.

Please give us more information:

What is your level of experience/competance?

What equipment do you have access to?

(We need this information to help guage how safe it will be for you to measure this).

[strokebow]

Hi Thank you,

I've been working in the electronics industry for the last 3 years and as a hobbyist for a lot longer with a degree in electronics.

We have a decent budget so I'm looking if there are any off the shelf solutions? Perhaps someone can recommend a product. My searches haven't been successful yet.,
Failing that we will build something.

Thank you

[Rerouter]

Ok, what current can the supply source. and how much series resistance can your device cope with on its supply, you may be better off using a multimeter with a high voltage divider probe (correctly rated), over a high voltage resistance. essentially low side sensing, the correctly rated probe is just to cover all edge cases so people nearby are not showered in exploded parts of multimeter if it shorts a 5KV supply with some current drive when its input protection tries to clamp it.

A high voltage probe is generally 9 or 99 Megaohm to divide down the meters normal 1M to a smaller scale, e.g. the 9 Meg probe would make 5000V measure as 500V on the meter (Over 10 Megaohm) if for whatever reason your current sense resistor went open (it will need to not breakdown under the full supply voltage, but can fail from overheating, the first would appear as a short, the second would become an open circuit, fusing itself.)

so for 10mA lets say, you can probably afford to throw 20V of signal at the problem. which would be a 2000 ohm low side current shunt, 10mA is 20V, and only disspating 0.2W, you would likley build it out of something like 10 x 200 ohm resistors in series to get the voltage rating high enough.

You measure this signal through the divider probe giving your 2V for 10mA, 0.02V for 0.1mA, which should be inside the highest accuracy range for a multimeter. (its counts will generally line up with millivolts, so a 2500 count would roll to the next range at 2.5V.

[capt bullshot]

I don't know of any off-the-shelf multimeter supporting this.
But you can use any multimeter to measure the current at high voltage potential, as long as you provide sufficient isolation externally and take care of your personal safety.
- use a battery operated multimeter
- put it onto an isolating surface
- don't touch the multimeter while your 5kV is switched on
- stay away from the system while running it
- depending on the available energy in your system, the multimeter might fail in a spectacular way if the energy gets dumped into the multimeter (say, a short happens at the wrong place and time)

Modern handheld multimeters should have an all side isolated enclosure, making the task of isolating the multimeter simpler. But always keep in mind, this isolation isn't designed to stand 5kV in regular operation.
There exist multimeters with a detachable wirelessly operated display unit, this can ease the task of reading the measurement.

[idpromnut]

A hobbyist here, so take what I say with a large grain of salt.

In theory, if you isolate the meter in the manner described above, and place the meter in the current path, then you will only need a meter that can deal with the relative voltage drop between the inputs. In the case of the mA range, it will probably be in the ball park of 10ohms (-ish, measure your meter!!!). Assuming that, then you would be seeing at worst (assuming 10s of mA is 50mA max) 10ohms x 0.05A = 0.5V.  Most (any?) DMM should be able to handle this without any issues I would think.

Again, this all goes out the window any part of the DMM (or test setup really) has the potential (no pun intended) for getting grounded. So with a careful set up, this should be reasonably safe to measure.

EDIT: I just realized I should probably spell it out: you are "floating" your DMM here, so assume the ENTIRE DMM (case, leads, etc) could be at the 5kV potential!

[alm]

Also something to consider: what happens if the fuse opens for whatever reason?

My suggestion is a (contact less) current clamp (AC or DC). Just make sure the jaw is wide enough to fit the wire with sufficient insulation. Getting one that goes low enough may be a challenge. You can wind the wire around the clamp for increased sensitivity.

[duak]

G'day,

How about a LEM magnetic current sensor?  www.lem.com

These are non-contact sensors that measure the magnetic field surrounding the conductor and provide a current or voltage signal proportional to it.  The conductor has to pass through the sensor and to increase the sensitivity to the mA range requires a number of turns.

This model has the primary winding already in place and is rated to 20 mA and 6 KV isolation: http://www.lem.com/hq/en/component/option,com_catalog/task,displaymodel/id,90.20.34.000.0/

Cheers,

Duane

[Elf]

My relatively non-exotic suggestion:

• Wireless multimeter left at a distance (e.g. Mooshimeter or something fancier from Fluke)
• Two (redundant) inexpensive spark gaps (GDTs) in parallel with the meter, to save the meter if the fuse goes

The voltage across the current shunt in the meter should be low, as long as the current is low. Of course if the fuse opens that will not be the case, but the GDTs should handle that situation and clamp the voltage to 90V or less which should be very tolerable for the open fuse. The only thing I can think of is that GDTs can be a little slow, so you may look into other protection devices (e.g. double zener TVSS diodes) in a similar configuration.

[alm]

This model has the primary winding already in place and is rated to 20 mA and 6 KV isolation: http://www.lem.com/hq/en/component/option,com_catalog/task,displaymodel/id,90.20.34.000.0/

Note that the isolation rating is for bare wire. You can go beyond that if you use insulated wire of the correct rating (probably a good idea if you loop the wire multiple times ).

[Someone]

This model has the primary winding already in place and is rated to 20 mA and 6 KV isolation: http://www.lem.com/hq/en/component/option,com_catalog/task,displaymodel/id,90.20.34.000.0/

Note that the isolation rating is for bare wire. You can go beyond that if you use insulated wire of the correct rating (probably a good idea if you loop the wire multiple times ).

Exactly, a current clamp/transformer/Rogowski coil can be uprated for arbitrary voltage ranges with suitable insulation on the conductor passing through it. The more important specifications tend to be bandwidth and peak current reading/withstand.

[Zero999]

Also something to consider: what happens if the fuse opens for whatever reason?

Whilst that's a valid question, we don't know what the circuit is. The original post says the maximum current is 10s of mA, so it could be that the circuit has such a high impedance, that it's impossible for the current to be high enough to blow the meter's fuse, making it inherently safe, as long as the meter isn't touched whilst the power is applied.

My suggestion is a (contact less) current clamp (AC or DC). Just make sure the jaw is wide enough to fit the wire with sufficient insulation. Getting one that goes low enough may be a challenge. You can wind the wire around the clamp for increased sensitivity.

Contactless current measurement is difficult to do at low currents, because the H-field is tiny. I suppose you could use a large number of turns and a hall effect sensor but then insulating the wire won't be easy. I suppose you could use ordinary magnet wire and dip the coil in several coatings of epoxy.

[alm]

The Tektronix A6302 + AM503 is specified at approximately 0.15 mA RMS noise and goes down to 10 mV/mA. The jaw is only 0.15", so fitting an insulated wire may be tricky, never mind multiple turns. That might just be usable for these kind of measurements. I realize that that is an expensive combination that is overkill, and I am not sure if the modern probes with integrated amplifiers have the same performance. If Tek could do it from DC to 50 MHz in 1980, however, then I expected it to be achievable with an off-the-shelf Hall effect sensor from the usual suspects (e.g. LEM) from DC to a few kHz.

In my opinion the contactless method is the safer option. Putting a meter not rated for the voltage behind a blast shield can work, but is a more complicated setup with more things to go wrong.

If you go for the DMM, would this be the one occasion where it might be safer to bridge the internal fuses (obviously the meter should be marked as such) and provide a properly rated fuse outside the DMM? In my opinion no fuse is better than an arcing fuse in this case (arcing fuse = high voltage across the input terminals).

[duak]

Hello,

I just remembered these guys: http://appliedphysics.com/products/other-products/  They make a couple of non-contact milliammeters using fluxgate sensors.

If low noise, precise measurements are needed I'd consider some sort of current shunt  with a digitizer floated at the HV and a fiber data link back to earth.

Just a note on double insulating conductors.  The voltage across each insulator is a function of both its thickness and of its dielectric constant.  If the insulators are the same thickness but have radically different dielectric constants, the voltage across each will be different.  The better insulator will take the lion's share of the voltage and if it's above its breakdown voltage it can fail.  Bottom line: the insulators' breakdown voltages do not sum nicely and unless there are other considerations each insulator has to be rated for the peak voltage.  Two insulators do buy a safety factor where there are possible defects or damage.

About high voltage fuses. HV arcs and sparks are fast phenomena with edge times on the order of a microsecond or less.  I can't think of any fuse that would open faster than a semiconductor junction will short circuit.

BTW, I have a Tek P6042 current probe from the mid 60's.  It does mA level DC to 50 MHz using a current nulling hall effect sensor for low frequencies and the nulling coil itself for the high frequencies.  Probably fun to calibrate.

Cheers,

Duane

[strokebow]

Wow thanks for all the responses.

Im blown away by how many good ideas there are and all of the helpful suggestions!