# High voltage DC parameters measurement. How?

Hello All! Have a nice day and stay healthy!

I need to measure relatively high voltage - DC (constant current) in range from something like 25 volts (low level floating and unknown now) up to 1200 volts and current in same line from 0 up to 30 Amps. Its possible over voltages spike like up to 2500 volts and current "flashes" up to 100 Amps at very shorts times durations - something like 1/5 millisecond - and if this current spike occurs - independent protection circuit shut down all.

So i need idea how i can do measurement.

Good luck to all!

It is possible with a voltage divider. Just a few resistors is safe enough. You might want to add a few extra clamping diodes, but that is not needed. Don't use a zener-diode, its leakage current makes measuring the voltage unreliable.

Perhaps it is safer to use a battery powered Arduino board and transmit the data wireless to an other Arduino which is connected to a computer.
When you have peaks of 2.5 kV, then there can be high voltage peaks in the GND as well.

Use larger metal film resistors.
For example R1 = four resistors in series of 10 MΩ each.
and R2 = 33 kΩ. Because it is higher than 10k, add a 10 nF or 100 nF capacitor in series.
With an Arduino Uno with an internal reference of 1.1V.

If four resistors is not enough, then perhaps 10 resistors in series of 1 MΩ each for R1 and 10 kΩ for R2. Then it is safe for peaks up to 15kV, since it is allowed to push 1mA into a pin.

Can you show a photo of it before you use it ?
I think it is best to put the resistors in series in one long row and nothing near the input point. Use a board without copper (not prototype board).

The 1.1V internal reference should be measured for each individual Arduino board.
Set the analogReference() to internal 1.1V and measure AREF. Use that value in the sketch.
Measure also the resistors.
Perhaps you have to tune it a little more in the code.

I'm not sure about measuring the DC current. Most things can not deal with 2.5kV voltage peaks. Is there already a wire in the circuit that you can use to measure the voltage drop ?

You can’t do this with resistors in a divider - they may well arc over .

This is a specialist task and you need to use proper commercial equipment designed and installed by a professional. Such voltages with that current capacity are a killer .

It’s not a hobby activity .

hammy:
You can’t do this with resistors in a divider - they may well arc over .

This is a specialist task and you need to use proper commercial equipment designed and installed by a professional. Such voltages with that current capacity are a killer .

It’s not a hobby activity .

Thanks for your reply. Yes, i understand - power electronics its really danger, but im really careful.
I work with high power systems before but i newer use any MCU to control it.
But now i need MCU to control output parameters because all process very unstable and i don't have to fast reaction to control whole system. We design vacuum system for titanium nitrite coating and we move from resistance coating (metal vaporization by direct heat) to magnetron sputtering. And in this process we must "stay on edge" - we need working voltage something like 600 V (depending on pressure in chamber) and high current - something like 20 Amp. And if we have to much ionization in chamber - current jumping up and we see arc discharge - this is abnormal mode for sputtering. So my new power supply must have voltage stabilization and current control. Now i have ready to use solid state protection circuit, but this circuit is "black box" for me - she don't have external connection with current and voltage output - only error output - i can see by my system - arc is happened, but i don't know critical point by voltage and current.

Koepel:
It is possible with a voltage divider. Just a few resistors is safe enough. You might want to add a few extra clamping diodes, but that is not needed. Don't use a zener-diode, its leakage current makes measuring the voltage unreliable.

Perhaps it is safer to use a battery powered Arduino board and transmit the data wireless to an other Arduino which is connected to a computer.
When you have peaks of 2.5 kV, then there can be high voltage peaks in the GND as well.

Use larger metal film resistors.
For example R1 = four resistors in series of 10 MΩ each.
and R2 = 33 kΩ. Because it is higher than 10k, add a 10 nF or 100 nF capacitor in series.
With an Arduino Uno with an internal reference of 1.1V.

If four resistors is not enough, then perhaps 10 resistors in series of 1 MΩ each for R1 and 10 kΩ for R2. Then it is safe for peaks up to 15kV, since it is allowed to push 1mA into a pin.

Can you show a photo of it before you use it ?
I think it is best to put the resistors in series in one long row and nothing near the input point. Use a board without copper (not prototype board).

The 1.1V internal reference should be measured for each individual Arduino board.
Set the analogReference() to internal 1.1V and measure AREF. Use that value in the sketch.
Measure also the resistors.
Perhaps you have to tune it a little more in the code.

I'm not sure about measuring the DC current. Most things can not deal with 2.5kV voltage peaks. Is there already a wire in the circuit that you can use to measure the voltage drop ?

Please check my circuit - i understood you right?

Yes - i can use shunt but i don't know how to build a amplification circuit to voltage drop (current )measurement.

I check great book The Art of Electronics by Horowitz, Hill but i don't see clear solution for me, because i don't have experience to work with such power systems....

Any kind - thank you very much for reply. Stay healthy!

I don't sure can you see circuit.

C1 should be parallel with R5.
You don't have to quote my complete post, what I wrote is already there.

Now we have a conflict for the names. In a voltage divider, we are used to call the resistor on the high side "R1" and the resistor to GND "R2". In your schematic, you have other R1 and R2. I will call the virtual resistors of a voltage divider "R1" and "R2".

Let's go for more safety.
I prefer 10 resistors of 1MΩ or even 20 resistors of 1MΩ for "R1" and 10k for "R2".
The capacitor is to reduce the noise and to compensate for the high value of "R2" and to filter peaks for more safety. When you want to detect a peak and "R2" is 10k or less, you can omit that capacitor.

It might be better to not connect the Arduino GND to the ground of that system. I suspect that the ground of that system can have voltage spikes in the ground of hundreds of volts. It is safer to have two voltage dividers, one for the high voltage and one for ground.

When the voltage over the shunt is about 5V, then you could measure that with two seperate voltage dividers and calculate the difference. You have to tune the resistors and the calculation in the Arduino. Maybe you need a newer Arduino board for 12-bit ADC.

When you have an Arduino isolated (battery powered and wireless data transfer) then you can put the Arduino near the high voltage line and measure the voltage over the shunt (with a lot of extra safety). The voltage divider would be used to measure the ground. This increases the danger, because you sometimes want to connect a usb cable to update a sketch. Perhaps two Arduino boards is safer, one for the voltage and one for the current.

5V over a shunt that measures dozens of amps? Doesn't sound like a great idea to me, that's some 100W of heat to dissipate!

The INA series current sensors may work here, provided they can handle the voltages involved. They're designed for such currents.

Koepel:
It might be better to not connect the Arduino GND to the ground of that system. I suspect that the ground of that system can have voltage spikes in the ground of hundreds of volts. It is safer to have two voltage dividers, one for the high voltage and one for ground.

Now capacitor set up in right place? I understood about serial connection of resistors - “more resistors in chain - more safety”, right?

I think to use two MCP3421 ADC with I2C interface to Arduino board. And ADUM1250ARZ I2C isolator chip for safety.

I have precision shunt resistor with 75 mV voltage drop at 30Amp. ADC have 18 bit resolution and with gain set to 8 i can measure up to 0.256 V on 18 bit - theoretically this is very good resolution for measurement.

On second ADC i don’t need to use gain to measure voltage output from devider.

So if this is good solution - i need answer on two questions - how i can separate grounds and how i can separate +5 power supply. Or +5 separation not necessary?

A single resistor can not withstand a lot of voltage. Also dust and dirt and the pcb board might cause troubles. The more resistors you use the lower the voltage over each resistor is and the further the high voltage is physically stretched over a longer distance. For 10kV I'm happy if the distance is 10 cm between the high voltage and low voltage. There must also be some kind of mechanical way to prevent that the wire of the high voltage breaks, and if it breaks then it should do no harm.

When you say that the voltage is 1200V with a spike of 2500V, then I assume that there can be very short voltage spikes that are a lot higher.
Using a very small isolator chip that can withstand 2500V is not good enough in my opinion. The pins of the chip are very close to each other. It can work, but then you have to add a protection circuit as used in CAT III or CAT IV multimeters with MOVs and so.

Here is a test from EEVblog with multimeters with and without protection circuit: https://youtu.be/7j5NtKB0vss?t=900.

I would use something with at least 10kV isolation.
Are there USB cables with only a optical wire inside ?

When using two of those voltage dividers, use one to the high voltage of the system and one to the ground of the system. You can calculate the difference in software or use a differential mode. There will be no direct connection from the system ground to Arduino GND.
Then you can add a resistor from Arduino GND to earth ground, to avoid charging the whole Arduino board with the high voltage.

Suppose that "R2" (R5 in your schematic) gets broken, then the high voltage will enter the Arduino. That is no problem, because the current is lower than the current the Arduino uses. The current is also too low to damage the input pin of a Arduino.
That means you could already do a test with only resistors and a old laptop.

Koepel:
Are there USB cables with only a optical wire inside ?

i don't have any optics or USB in my system.
For visualization i want to use LCD. And USB i need only at debug stage - without switching on high voltage.
When system be completed, i think to use isolated RS485 Modbus RTU to worck with main system controller. Or this power supply system be stand alone device. I don't know now.
I have three phase power controller based on triac (solid state device with 0 - 5 volt control input) at low side of transformer. At high side i have full bridge three phase rectifier (solid state module), filtering circuit, ballast circuit (resistors in parallel connection), and short circuit protection system (solid state). Without load i have 750 volts of output at transformer winding end (one phase). I think spikes only occurs when triac controller starting up. When sputtering magnetron "falling in arc" - current going up and voltage falling down at same time. But at working mode sputtering magnetron work as zener diode - constant voltage and floating current depending of ionization and pressure in chamber.

That makes it easier.

You can start with an Arduino Nano, so you can solder the resistors of the voltage divider to it (assuming that your soldering is better than a connector ).

The LCD displays are very sensitive for electrical noise. Even when you put it in a shielded box, the electrical noise will go into the display via the front and disturb the high impedance circuit inside.
You can start with a LCD display, but you might need to upgrade to 7-segment led displays. Some modules are easy to use: 7-SEGMENT LED DISPLAY MAX7219 & SHIFT REGISTERS - Arduino tutorial #21 - YouTube.

CrazyIgels:
Without load i have 750 volts of output at transformer winding end (one phase).

I thought you’re looking to measure DC. This would be AC - 750V RMS is about 2,000V peak to peak, or about 1,000V DC when rectified & smoothed.

wvmarle:
I thought you're looking to measure DC. This would be AC - 750V RMS is about 2,000V peak to peak, or about 1,000V DC when rectified & smoothed.

Right. Rectifier nominal working voltage is 1200 V. Dc. Working voltage i think be something like 300 - 500 V. So i need range from 0 to 1200. In real - not from zero, because triac start not from zero opening angle and i have start output voltage something like 10 - 30 V, but this is no matter.

Oh, rectified but not smoothed. So you'll be sampling the complete waveform, right?

Note that the current will also fluctuate big time, like the voltage, so that also needs pretty fast sampling.

wvmarle:
Oh, rectified but not smoothed. So you'll be sampling the complete waveform, right?

Note that the current will also fluctuate big time, like the voltage, so that also needs pretty fast sampling.

Smoothed. i have 1 mkF high voltage capacitors 3 pcs in parallel and sputtering magnetron work as zener diode and smooth voltage. And no mater for me if i measure voltage and current at top of wave or at bottom.

If this is for a commercial application I would look for an off the shelf turn key solution. That would include a Voltage Transducer and a Current Transducer. 25 VDC ri 1200 VDC is a very large range. The links are merely examples but I would be looking din rail mounted turn key solutions like DC voltage and current transducers.

Ron

Ron_Blain:
If this is for a commercial application I would look for an off the shelf turn key solution. That would include a Voltage Transducer and a Current Transducer. 25 VDC ri 1200 VDC is a very large range. The links are merely examples but I would be looking din rail mounted turn key solutions like DC voltage and current transducers.

Ron

I think about commercial available sensor, but it cost to much and because all borders are closed (thanks fucking virus) - shipping time is unknown. And also - for only one experimental system - i dont think that is be a good solution...

Anyway - thank you for good link.

Latest shipment from China to me, a week or two ago so with the lockdown of Hubei still pretty much in full force, went perfectly normal - two days. One day for order processing/packing/handing to courier; one day to get it across the border and delivered to me. That even with the 14-day quarantine requirements for arrivals from the mainland (well, transport drivers are exempt from that) and most border crossings completely closed.

Or maybe just get it from something closer by where-ever you live.

CrazyIgels:
I think about commercial available sensor, but it cost to much and because all borders are closed (thanks fucking virus) - shipping time is unknown. And also - for only one experimental system - i dont think that is be a good solution...

Anyway - thank you for good link.

For a one-time test setup then I would be thinking about high voltage thick film resistors. They are designed and made for high voltage divider applications. This is what most commercial high voltage use, like the Fluke 80K40 for example.

Current? I would look at a low side current shunt.

Ron