Voltage divider for high volt measurement

[EdStraker]

Trying to get a measurement from a 20kV Negative Ion Generator that is adjustable from 15kV- 20kV. My digital meter is 600v Max. and my analog is 10kV Max. Want to set it to 17k - 18kV

Was thinking the only solution was a Voltage divider. With that said, I need some advise on resistor type 1/2 Watt Carbon / Film? Ceramic power resistors? (I would imagine 1/4 Watt's are out) so I can get the the voltage down to a measurable range.

Or should I just not make the attempt? I figured if the was SAFELY possible, was thinking of making a divider bank similar to a selectable resistor bank.

The only specs I have for the unit is:
Input voltage: DC 12V
Output voltage: 15000V-20000V adjustable
Rated power: 5W

 

[Tony Stewart]

Lexmark laser printers use 15 kV on the drums and use a sine wave stepup signal transformer to create this then rectify and regulate with a long string of 1/4W 10M~100M resistors back to an LM324 to regulate the output like AGC. You can string these inside a DIY plastic wand with a reinforced tip with a total R of 1 GOhm and make this suitable for measuring 50 kV using 1kV/mm for a clean wand with air at 3kV/mm. Surface contamination reduces the withstanding voltage.

Now you can get 1Gohm HV resistors at low cost in stock so you may use fewer in a string https://www.digikey.ca/en/products/detail/te-connectivity-passive-product/RGP0207CHK1G0/2389890

When we used these in the 80's outside a clean room to generate ions with needle inside a 1cm half cup, they created so much RF that it interfered with our magnetic servo writers inside the clean room (for HDD Mfg) , so I asked maintenance to remove it.

 

[EdStraker]

Thanks for the reply, what would this give me as a reduction factor? 10:1, 20:1, 50:1 ?
So I assume a standard 2 resistor VD is not a safe idea?

 

[Tony Stewart]

Not unless you have a long expensive resistor...

If you measure your DMM impedance it may be 1 Meg or 10 Meg then 1Gig is 100:1 ratio and < 1W dissipation (V^2/R)

 

[EdStraker]

the spec sheet says:
DCv = 20k ohm/V input resistance

 

[Tony Stewart]

then you choose max scale and use that V * 20k = Rin for the analog meter but use that to measure the DMM on 600 V scale so that 200V = 20 kV using 99:1 R ratios for 100:1 or whatever you prefer.

 

[EdStraker]

Wait, you lost me.....
use what voltage?

those specs were for my analog meter. Digital meter is 600V max / 10M impedance

 

[Tony Stewart]

Yes Whatever voltage scale your analog meter is multiple by 20k is it's input resistance while the DMM is 10M

I forget what Analog Vdc scales are used in Simpson meters and you didn't specify.. So if you used a series string and a shunt in parallel with meter, you have to compute the new ratio or tell me the voltage and thus the resistance to know what you need

 

[EdStraker]

Ok, so sticking with the 600v / 10M DMM........
The generator has a range of 15k - 20k output. Since I have no idea if its already set at Max or not. I need to protect the meter while dialing it down, or up as the case may be.

 

[MrAl]

Trying to get a measurement from a 20kV Negative Ion Generator that is adjustable from 15kV- 20kV. My digital meter is 600v Max. and my analog is 10kV Max. Want to set it to 17k - 18kV

Was thinking the only solution was a Voltage divider. With that said, I need some advise on resistor type 1/2 Watt Carbon / Film? Ceramic power resistors? (I would imagine 1/4 Watt's are out) so I can get the the voltage down to a measurable range.

Or should I just not make the attempt? I figured if the was SAFELY possible, was thinking of making a divider bank similar to a selectable resistor bank.

The only specs I have for the unit is:
Input voltage: DC 12V
Output voltage: 15000V-20000V adjustable
Rated power: 5W

Hi there,

Using a voltage divider in theory works because it lowers the voltage at the test point down to a voltage that differs by a constant scaling factor. If the divider ratio turns out to be say 10, then if you measure 100 volts with the meter then you actually have 1000 volts at the test point. That is of course very simple to understand.

What is not that straightforward is that the impedance (or in this case resistance) of the meter also plays a part in the calculation of the voltage divider resistor values. For a straight up 10 to 1 ratio with a pure voltage divider, we might use a resistor of 9k and 1k to get that 10 to 1 factor, but with the resistance of the meter itself this changes. If the meter resistance happened to be 1k (not likely though) then our measurement would be way off.
If we measured 100 volts with the raw voltage divider we would know we actually had 1000 volts at the test point, but considering the resistance of the meter itself we would only measure about 52.6 volts, and since we would still be using a factor of 10, that would make it look like we only had about 526 volts at the test probe point. That would be a large error.
So how do we correct that? Well, if we had a second meter that we could use to calibrate our voltage divider plus original meter, we would find we could use a factor of 19 instead of 10. When we multiply 52.6 times 19 we get close to 1000 volts as the result which tells us we have about 1000 volts at the test point, which is correct.

But since the meter is 1k, why bother with that second 1k resistor for the voltage divider? In fact, if we use just one single 9k resistor in series with the meter, we already have a voltage divider with one of the resistors (1k) being the meter itself. Now when we measure that 1000 volts we see 100 volts on the meter, and the factor of 10 then applies just like with the raw voltage divider.

That's the traditional way of increasing the range of a voltmeter. You add an additional series resistor in order to form a voltage divider with the meter. You do have to know the resistance of the meter though.
This would be best when you cannot load the voltage source much. If you want to draw the minimum current from the voltage source, then that's the way to do it. That's unless you want to use an amplifier and voltage divider, which is an even better way to do it.

Using an amplifier and resistor voltage divider, you can set the resistance values much higher and therefore load the voltage source even less. Sometimes you will need to load the voltage source though to get the loaded voltage measurement, and in that case you load it with a separate resistor.

The way the larger resistor values are usually made is a bunch of smaller value resistor are connected in series. That's partly because each resistor can only take so much voltage across its two leads.
You also have to be careful about how you mount these resistors. If you use a zig zag pattern you have to make sure that no node in the series string comes too close to another node or else it could cause an arc over across those two points.

 

[Nigel Goodwin]

Using an amplifier and resistor voltage divider, you can set the resistance values much higher and therefore load the voltage source even less.

Not an issue in this case though, as his analogue meter already has a higher input impedance (20Mohm on a 1000V range) than a digital one (10Mohm on all ranges). Something to bear in mind when calculating the resistor values. Perhaps it would be a good idea to make it for a 10Mohm digital meter, then add a 20Mohm resistor across the 20Mohm analogue meter to make it 10Mohms - making it more versatile.

He also needs to bear in mind the voltage ratings of the resistors, you often need to use higher wattage ones as they generally have higher voltage ratings.

 

[ronsimpson]

You need one of these. Years ago TV repair men all had one. If you were close I would let you use one for a while.

I need some advise on resistor type

I do have some 40kV resistors, but you can't afford them.
Look at the data sheet for resistors. Many cannot handle more than 100s of volts. It is typal to use many resistors in series.

 

[Nigel Goodwin]

You need one of these. Years ago TV repair men all had one.

Not all

We used to have an Eagle EHT meter - basically it was a probe, like above, but it had an analogue meter built-in the handle, and obviously a lead and crocodile clip to connect to chassis.

Apparently, there was once an EHT probe, but it got dropped, and broke the internal resistor - before my time though.

Generally no degree of accuracy was required - if it didn't read about 25KV for a colour CRT then there something wrong, anywhere near 25KV was fine. If one of the tripler stages was faulty (and you couldn't tell by the smell ), then it would be only 16KV or so.

It was rarely ever used, more as a demonstration than anything else.

 

[EdStraker]

Thanks for the extensive replies. I did find an amplified circuit solution as MrAI suggested, but since I was already into a construction project I didn't want another one on top of it. I was kind of fishing for a quick and dirty solution.
I actually did find the high voltage probe and thought "there you go" until I saw the price. It costs more that my meters and I'll probably never use it again since most of my projects never exceed 24v, most are below 12v actually so that's out right off.

The amplified method seems to be the better alternative.......just didn't want to have to build it for a one off.
I do have some LM358's hanging around, maybe they will work.

 

[Nigel Goodwin]

Thanks for the extensive replies. I did find an amplified circuit solution as MrAI suggested, but since I was already into a construction project I didn't want another one on top of it. I was kind of fishing for a quick and dirty solution.
I actually did find the high voltage probe and thought "there you go" until I saw the price. It costs more that my meters and I'll probably never use it again since most of my projects never exceed 24v, most are below 12v actually so that's out right off.

The amplified method seems to be the better alternative.......just didn't want to have to build it for a one off.
I do have some LM358's hanging around, maybe they will work.

I suspect you're a little confused?, what use would an amplifier be? - you need to reduce the voltage, not amplify it.

As suggested throughout the thread, you just need a high voltage, high resistance, probe - which along with the impedance of your meter will make a suitable attenuator.

Assuming your analogue meter has a 1000V range, then to make it read 10,000V you need a total of 180Mohms in the probe, which along with the 20Mohms of the meter, will drop the voltage down to 1000V (divide by ten). To read 20,000V use 380Mohms, which will give you an FSD of 20,000V - BUT you will have to manually (mentally) double the reading on the meter and times it by ten, so 350V on the meter will actually be 7000V.

 

[Tony Stewart]

For a low voltage out of the probe, use 1000:1. You can make or buy easily, ($1 vs $100)
I linked where you can already buy cheap 10kV resistors.

Do you care about accuracy 1% 10% 20%? or DIY calibrated 1%,
Location, availability?
Budget ?
Purpose ? Experiment? research?
Portable or permanent... passive meter or battery operated DMM,
Arc detection required? (PD) etc.

Buy specification parameters:

Can be connected to any digital multimeter
Measuring up to 40KvDC or 28KvAC(RMS)
Accuracy:
DC: ± 1% (1 kv - 20 kv)
± 2% (20 kv to 40 kv)
AC: ±5% typical value at 60Hz
Distribution ratio: 1000:1
Operating temperature: 0℃ -- 50℃
Storage temperature: -20℃ -- 70℃
Input impedance: approximately 1000 megohm

 

[EdStraker]

Do you care about accuracy 1% 10% 20%? or DIY calibrated 1%,
Location, availability?
Budget ?
Purpose ? Experiment? research?
Portable or permanent... passive meter or battery operated DMM,
Arc detection required? (PD) etc.

This kind of accuracy wasn't important, it was just to get an idea where the unit is set from the factory and kind of get it in the middle of the adjustable range. I was hoping for a quick and dirty (safe) method of doing it was all. I really didn't want to get into a construction project that in all probability will never have the need for ever again.

This was more of what I had in mind, but again outside my budget. BK Precision HV44A

 

[Nigel Goodwin]

This kind of accuracy wasn't important, it was just to get an idea where the unit is set from the factory and kind of get it in the middle of the adjustable range. I was hoping for a quick and dirty (safe) method of doing it was all. I really didn't want to get into a construction project that in all probability will never have the need for ever again.

This was more of what I had in mind, but again outside my budget. BK Precision HV44A
View attachment 148885

That looks quite similar to the Eagle one we had at work, as you (and I earlier) said, accuracy wasn't particularly important.

 

[ronsimpson]

accuracy wasn't particularly important.

Here is a silly little high voltage tester. It has a series of resistors and Neon bulbs. Electric Fence Tester.

 

[Nigel Goodwin]

Here is a silly little high voltage tester. It has a series of resistors and Neon bulbs. Electric Fence Tester.
View attachment 148888

Again, accuracy isn't important - and is even less so in this case

As a young teenager I used to go beating (for pheasants) on a local landowners estate, and one place we used to have to beat through was a field of kale - now cows love kale (it's only suitable for cattle feed! ) - so they use to tun an electric fence across the field, so the cows only ate a strip at a a time, them move it back another strip.

However, cows go crazy for kale, and a normal electric fence didn't hold them back - so the farmer ran a 240V mains extension to the field to power the fence. As kids we were VERY wary of this, as you could see the overhead mains wire running to the fence, and we had to cross it, usually soaked through and in the pouring rain!.

I don't have a picture, but they used a sell a spark gap EHT meter, which had a threaded rod with a knob that you adjusted until it started to spark across the gap - I've never used one, or even dared to think of it