Howto Measure High Voltages with a Multimeter

[joeqsmith]

What is the normal way to measure high voltages with a multimeter?   I have been thinking about the problems with some meters and how to harden them.   Plugging the probes into the wrong connectors of the meter or having the meter on the wrong setting may cause a problem, not to mention not all meters may actually meet the ratings marked on their cases.    It seems with a little circuit, we could address these problems.   

Obviously, I have made my own HV probes in the past for an oscilloscope.   There I was looking for wide bandwidth in the several MHz.   For a multimeter,  looking at mains, I would think even 1000 Hz is more than enough.  This should really ease the design.

The following article talks about high voltage probe designs:

http://repairfaq.cis.upenn.edu/sam/hvprobe.htm

In here they show a simple attenuator:

To design the voltage divider, the input impedance of the meter must be taken into account. There is a minor but significant difference between DMMs and VOMs.

    DMM: Z-in is usually constant, often 10M ohms.
    VOM: Z-in is the voltage range (full scale) times the ohms/volt rating of the meter.

Here is the basic circuit:


  High Voltage <------/\/\/\/\/\---------+-------------> + to DMM/VOM
                           R1            |                      |
                                         \                      \
                                      R2 /                   R3 /
                                         \                      \
                                         /                      /
                                         |                      |
  Ground Clip  <-------------------------+-------------> - to DMM/VOM



What I don't like about this approach is if R1 arcs, bad things may happen.   

This article is more about using a meter on mains.   He considers an arc condition and uses fuses and GDTs to control it.   He warns about the need for a good return path.   

http://www.marcspages.co.uk/tech/5103.htm

What I again don't like about both of these approaches is the you could swap the in and out of the attenuator, just like you could swap the leads of the meter.  In both of these circuits, this may cause problems.

I have two thoughts about this.  Using the second article, the design could follow a standard T attenuator.   This would allow you to swap the in and out with the same performance.    What I don't like about this is we still have a low impedance path to the meter.   Maybe this is a good thing, but I would rather have the meter isolated.  It seems like using a balanced T attenuator would be a good solution.   Voltage ratings could be set high enough to handle a single point failure.    You would be giving up noise and resolution but I am not sure it would be important when looking at high voltages.   

The center could be made from multiple R's to handle an open condition (current to meter could still be very limited) and clamped to limit the voltage to the multimeter.   

I scribbled some calcs and checked the parts.  Seems like it could be fairly inexpensive to make.     I don't like that the resistors are not isolated, but the multimeter itself is not isolated. 

[Lightages]

What I did to get a reliable and safe way to measure high voltages is I purchased a complete Fluke 27/FM kit from ebay. The price of the complete package was less than buying the probe itself. An example:
http://www.ebay.com/itm/Fluke-27-FM-Multimeter-w-case-HV-probe-and-leads-/151710145261?pt=LH_DefaultDomain_0&hash=item2352a11aed

If I didn't already have mine I would not hesitate to purchase that one. The appearance also implies it is the TRMS model. The bonus is getting the multimeter essentially for free.

[joeqsmith]

Thanks for the response.   I am not at all familiar with this kit but from the link you provided, it comes with the  80k-6 HV Probe.  From Fluke's website:

The 80K-6 is a high voltage probe designed to extend the voltage measuring capability of an AC/DC voltmeter to 6000 volts peak AC or DC Overvoltage Category I. This means the probe can only be used to make measurements on energy limited circuits within equipment. Examples include high voltage within televisions or photo copy machines. DO NOT use this probe to measure high voltages on power distribution systems. The probe provides high accuracy when used with a voltmeter having 10 Megohm input impedance.

    A high voltage probe that allows a multimeter to measure up to 6,000V. Intended for low energy applications only
    One year warranty


We can also see the Fluke 27 is not CAT rated which makes sense as the probe is CAT I only.    This is actually not a bad deal at all but what I am more interested in CAT II and up. 

[joeqsmith]

To maybe help clear it up, starting with 20dB attenuator or divide by 10.  Assuming multimeter is 10M input and rated for 1KV.   Basic bob CAT 1 multimeter, like the Fluke 27 above....

Upper left, basic T attenuator formed out of 2 X 20M and 1 12M resistors.    If meter was shorted (say you connected the outputs of the attenuator to the current inputs) the power on the input R is 2.7 W.  Voltage on 12M is 3KV with 10KV applied which is what we need to clamp to.    So you can flip the in to out but I don't like the low impedance path to the meter.   

Lower left, same circuit but the 20Ms are split into two 10s making it balanced.  Power requirements for the 10s is much less.  Now flipping the in and out and the red and black makes no difference and there are now no low impedance links to the multimeter.

On the right, the 12M is now made up of a single 30M in parallel with 2 X 10M in series.  These 2 10M are in parallel with the two GDTs and equalize them.  Same as the one article linked.   

If the Caddock MX431's were used, there is a crap load of head room.   Even if one arc'ed the second would not cause the fuses to blow.    Fuses would come into play on a double fault.   

I would pot the whole setup.  If it was ever hit hard enough to take out a fuse, I would never trust it again anyway.



Again, just thinking about it.  Curious what others have done.

[Lightages]

The probe is rated CAT1 but the grey 27/FM were not CAT rated as the standard came out after their production. The latest yellow with grey front are CAT rated at CATIII 1000V. I would trust the the old 27s too.

So you want to make a test probe for some unspecified voltage at CATII. How are you going to determine it meets that requirement?


Edit:
Also, where are you going to be measuring CATII at over 1000V?

[joeqsmith]

The probe is rated CAT1 but the grey 27/FM were not CAT rated as the standard came out after their production. The latest yellow with grey front are CAT rated at CATIII 1000V. I would trust the the old 27s too.

So you want to make a test probe for some unspecified voltage at CATII. How are you going to determine it meets that requirement?

I have a lot of respect for high energy.   If I ever needed to measure 440, I wouldn't trust anything!  I would have a 20' pole with the little probe taped to the end!      

Sort of.  Again, I am just thinking about the whole CAT ratings and how to make a more robust setup.   The attenuator is not so much a probe.  It would go between the probes and the meter, limiting the current and voltage that could be supplied to the meter.     Obviously, the probes would still need to meet what ever you were wanting to measure.   My thought is with an attenuator like this, with the CAT III 1000 surge test (8/20 2ohm 8KV) the output would never exceed the meters limits.    If I build it, I plan to allow it to sit on an 8KV DC source. 

If I build it, I plan to test it along with the low cost meters I am looking at.   The easiest way I can see to test it is to use a 10M load and scope to monitor the output during the surge and hook it to a 10KV DC source for a day or so.   

Just thinking about it for now.   Obviously  from the links, others have also gone down this path.    I see some things I am a bit concerned with about what they are doing.   I like the idea that the leads could be flipped and there is no low impedance path to the meter.  I also like that the currents are sub mA at the meter, even under a surge condition.  I also like that even if any one of the resistors broke down or opened, the voltage at the output would still  remain within the meters limits.    There may be a down side but I am not seeing it.   

[joeqsmith]

Edit:
Also, where are you going to be measuring CATII at over 1000V?

I am not sure I understand what you are asking.  CAT II 1000V is tested to 6KV with a 12 ohm source using the same waveform.   Chances of me looking at the house wires in the main circuit, during a lightning storm and then getting hit, slim to none.   

Or are you asking if I would ever actually use the meter at this level?   If so, maybe.   The problem with what I am showing is the source impedance is very low.    With the output shorted, the input would be 27.5M   So using it for HV power supply work would be limited.   

That said, there have been times where I have cobbled attenuators to increase the limits of my bench meters.   Actually, as many times as I have done this, I have to say I never made anything that I saved. 

[tautech]

If you're going to design/build a circuit for members to copy, what is the accepted max V most would need?
Are CAT ratings for "in circuit"(not mains) testing? Or are you wisely playing it safe. 

There are many members fixing CRO's that don't have the equipment to measure/set the CRT HV of normally 2-3 KV and voltages in that range are not too hard to find cables and probes for.
5 KV max would be a good limit IMHO, still offering reasonable display resolution.

Good idea though. 

[Lightages]

joeqsmith

So you seem to be trying to design an additional safety device to be used in between a multimeter and the voltage being measured. This instead of a way to measure higher voltages than the rating on the meter?

If this is true, then I think you are beating a dead horse with another dead horse. If you buy a meter rated to to what you are measuring, and that meter is actually third party tested to the CAT rating needed, it is hard to imagine that you could be safer using a home made device in the circuit. I would certainly trust my Fluke or Brymen meter over anything I could design and build. I cannot verify a home made device to be as safe or safer than those brands.

There are considerations in design, components, and construction that are beyond a hobbyist to consider and test properly. Sure you can design and construct a device that from all appearances looks like it will do the job, but without testing you will never know if your device actually is as safe as you intended.

If you are trying to measure high voltages inside equipment that are plugged into the wall, then CAT1 is all you need. In this case you need to know what you are doing and not touch the wrong thing. It would very unfortunate if you touched a 3kV source with one hand whilst touching the chassis with the other. There really needs to be a video done on how to minimize the risks of working in such an environment.

If you really want to be working on equipment that is truly CATII over 1000V, then you better take a training course on dealing with these kinds of energy and forget using home made devices. I am not sure that CATII 1000V sources are even in use, or almost not at all.

[w2aew]

Another thing to keep in mind is that the input resistance of a DMM may change on different ranges, as discussed in this video.

[joeqsmith]

joeqsmith

So you seem to be trying to design an additional safety device to be used in between a multimeter and the voltage being measured. This instead of a way to measure higher voltages than the rating on the meter?

It's an attenuator so yes it would allow a meter to read higher voltages.   Granted it has added consideration for single point failures.   It is common practice to use attenuators to change the working limits.   This is not a new idea,  nor is considering single point failures.    Yes, it would clamp the voltage and may prevent damage to the meter in some cases.   Yes I plan to use CC type fuses if I make it.     

If you are trying to measure high voltages inside equipment that are plugged into the wall, then CAT1 is all you need.

Dang, I would hate to open up a CAT IV surge tester that is plugged in and start probing it with my CAT I meter!       But yea, if I decide to fix my toaster I'm sure I will survive.   

I would certainly trust my Fluke or Brymen meter over anything I could design and build. I cannot verify a home made device to be as safe or safer than those brands.

That's fine and you make a good point but don't assume we all have the same skills or equipment available to us.   When it comes to trust, my rule of thumb is don't. 

If you're going to design/build a circuit for members to copy, what is the accepted max V most would need?
Are CAT ratings for "in circuit"(not mains) testing? Or are you wisely playing it safe. 

There are many members fixing CRO's that don't have the equipment to measure/set the CRT HV of normally 2-3 KV and voltages in that range are not too hard to find cables and probes for.
5 KV max would be a good limit IMHO, still offering reasonable display resolution.

Good idea though. 

I doubt many would want to copy it as is, but you never know.    I suspect simple attenuators like the ones shown in the links or just HV probes like the one shown in the kit are far more common.   The main reason I posted was to see what others were doing.   

5KV is no problem, however even at 3000V it would be over 100uA of draw.  This seems like a lot.  Of course you could make an attenuator that raised the impedance higher.    Maybe something like 4 X 25M with the center clamped would work better.   Noise and coupling may be a problem which is why I kept the impedance so low.       

Of course there is always just putting a single 100M in line a praying it does not break down.    My last attempt of a HV wideband probe uses the OHMITE MOX series.   I just did not care about the accuracy and drift.    These were only about $7/ea.    I only have used them to 4KV in the probes.   I think if I make a good attenuator for the meter, I will stay with Caddock.     

[joeqsmith]

Another thing to keep in mind is that the input resistance of a DMM may change on different ranges, as discussed in this video.

   

[Lightages]

joeqsmith

If you are trying to measure high voltages inside equipment that are plugged into the wall, then CAT1 is all you need.

Dang, I would hate to open up a CAT IV surge tester that is plugged in and start probing it with my CAT I meter!       But yea, if I decide to fix my toaster I'm sure I will survive.   

Obviously a device made to increase the energy available is no longer CAT1

[Fungus]

What is the normal way to measure high voltages with a multimeter?

Use a high voltage probe: http://en-us.fluke.com/products/all-accessories/fluke-80k-40.html

[HighVoltage]

If you look around, you can find this HP 34300A high voltage probe for very little money.
Interestingly this one has a label: "Made in England" on it.
The tip can be unscrewed from an M4 tread.

[joeqsmith]

joeqsmith

If you are trying to measure high voltages inside equipment that are plugged into the wall, then CAT1 is all you need.

Dang, I would hate to open up a CAT IV surge tester that is plugged in and start probing it with my CAT I meter!       But yea, if I decide to fix my toaster I'm sure I will survive.   

Obviously a device made to increase the energy available is no longer CAT1

Exactly.  You made a blanket statement which is not true for all cases.    If you play with HV, you better know what your working on, know what your equipment can do and most of all know what you are doing!      That includes working on the toaster...    In w2aew's video, he made another good point beside the meters input impedance.  If your not comfortable working with HV, don't. 

If you look around, you can find this HP 34300A high voltage probe for very little money.
Interestingly this one has a label: "Made in England" on it.
The tip can be unscrewed from an M4 tread.

It's a nice looking probe.   Seeing that CAT IV is only 12KV, you would think that both the Fluke and HP probes would be just fine.    Would be interesting to know more about why Fluke considers their probe CAT I.     

Anything I have worked on at home in that 40KV voltage range has been very low energy stuff (< 200mJ) but much higher frequency than the 150Hz.     I have made my own capacitive coupled probes for this.   

[Lightages]

Exactly.  You made a blanket statement which is not true for all cases.    If you play with HV, you better know what your working on, know what your equipment can do and most of all know what you are doing!

I also said the same thing in my previous post. Context is everything, usually conveniently left out to make a straw man.

If you are trying to measure high voltages inside equipment that are plugged into the wall, then CAT1 is all you need. In this case you need to know what you are doing and not touch the wrong thing. It would very unfortunate if you touched a 3kV source with one hand whilst touching the chassis with the other. There really needs to be a video done on how to minimize the risks of working in such an environment.

If you really want to be working on equipment that is truly CATII over 1000V, then you better take a training course on dealing with these kinds of energy and forget using home made devices. I am not sure that CATII 1000V sources are even in use, or almost not at all.

[joeqsmith]

Attempted to build a balanced 100M attenuator but without using some sort of shield, seems a bit too susceptible to surroundings.   I was thinking a shield inside of another plastic box but I still would need to attach it.   Seems like it would make it more useful if it were done this way.

Tried the 27.5M as well using Teflon to hold the parts in place.   Much less sensitive.   BW is around 800Hz.   I am using Bussmann CC fuses and EPCOS EF GDTs.     Plastic box and a few other parts are on order including some decent probes and extensions.  Once done, I'll take a picture of it.   

[joeqsmith]

Sorry for the Electronics 101 stuff.   Thought it would be good to explain it. 

https://www.youtube.com/watch?v=bj_YndfA4Qs&feature=youtu.be

[Lightages]

Interesting video. I understand what you are trying to do, but I have some questions.

You said that you have never spent more than $50 on a multimeter. But then you spend $150 making a protector for them?

The output can be just normal shrouded banana plugs, but shouldn't the input jacks be something rated for higher voltages?

Shouldn't the input impedance be much higher, like maybe 1 gigohm so as to not have to carry so much current and also not load the device under test?

The Fluke 80K-6 is not just a resistor, it seems to be a resistor in series, 85 megohm, and one in parallel, 85kohm,  across the output to the multimeter. Yes, this is only rated at CAT1, but for most people how does your $150 box benefit them?

I like your project, don't get me wrong as it does solve a problem with high energy and high voltage. It just seems much more practical for most people to just buy a high voltage probe if they want to work on B+ supplies and other high voltages inside tube amps and TVs. I purchased the FM/27, plus high voltage probe, plus RF probe, plus the case, all for well under $150.

[Syncopator]

What is the normal way to measure high voltages with a multimeter?.

The normal way is surely to use either a meter which has a suitable range, or a meter plus a high voltage probe designed to be used with it. 

I once had a customer who wanted to measure about 1.2kV, but the highest range on the (identical) meters he had was 1kV. 

I suggested he use two of them, running on internal batteries, connected in series, thus making a 2kV meter with half the total voltage measured being displayed on each one. 

It worked, of course.

[joeqsmith]

Interesting video. I understand what you are trying to do, but I have some questions.
You said that you have never spent more than $50 on a multimeter. But then you spend $150 making a protector for them?

It's not really a question.  Possibly just confusion on your part.  Maybe I can help clear things up.   If you watched any of my videos, you would be aware that I have spent far more than $50 on a single multimeter.   My first Fluke was way over that.     I have not spent more than $50 on a handheld multimeter.      Am I willing to spend $150 ish on a device that will extend the upper range of a voltage meter and prevent it from being damaged, well, I did show it didn't I??   

The output can be just normal shrouded banana plugs, but shouldn't the input jacks be something rated for higher voltages?

I wanted to use the standard test leads.  So, no.   

Shouldn't the input impedance be much higher, like maybe 1 gigohm so as to not have to carry so much current and also not load the device under test?

If you paid attention, I was explaining how to design one, not how to copy one.   There's a difference. 

The Fluke 80K-6 is not just a resistor, it seems to be a resistor in series, 85 megohm, and one in parallel, 85kohm,  across the output to the multimeter. Yes, this is only rated at CAT1, but for most people how does your $150 box benefit them?

Hopefully the benefit does not come  from coping what I presented but from learning how to make your own.   If you listened, I provided a part number for a resistor series that would be in the $3 range and how fuses can be obtained for a fair price.   

[Lightages]

I think you are taking my questions as being adversarial. It was not my intention even if my lack of tact gave the questions that tone. They were genuine questions. Yes I was paying attention and I understood that you were teaching an idea, not the final object you produced. Your video was very good and I enjoyed it. The question about the input impedance was a genuine one and a higher impedance would have made your device more useful to you too, IMHO.

[joeqsmith]

Another demonstration of my homemade attenuator is about half way through the clip.   

[vinlove]

From what volts does it regarded as the high voltage which needs extra care and probes and devices in measurements?