LC Reactive elements cancel cause huge V drop. High voltage

[Jonathan Mc]

Hello all,

I am a senior electrical engineering student at Purdue University and I am working on my senior capstone project. One of the aspects of my project is to measure voltage and current waveforms as well as the relative phase shift between them. I need to get these waveforms into a data acquisition cards interfaced with LabVIEW. The DAQ card that I am using have the following parameters:

Range: +\- 10V
Resolution: 16 bit
Sampling Rate: 50kS/s

Because of this, I have designed a programmable amplification current front end circuit to take advantage of the full resolution of the DAQ (measuring in the mA range all the way up to 20A_RMS within 1% accuracy). The current is converted into a voltage via a burden resistor connected to a current transformer.

Everything is going fine but I recently stumbled upon a huge problem, and I need some guidance.

I initially made the assumption that the highest voltage that I would measure is 120V_RMS, but after playing with some different RLC load configurations with the on campus resistive, capacitive and inductive load boxed, I ran into this situation:

Hooking up a simple series circuit that has a source voltage of 120V_RMS at angle 0 degrees @60Hz with the following impedance:
Z_R = 171 ohms
Z_L = j1200 ohms
Z_C = -j1200 ohms

Because of the reactive elements having equal and opposite impedance, they cancel each other out on an impedance scale, which means that the source only sees the 171 ohm resister. Using ohms law, the current drawn from the source is:

I = V/R = 120V/171ohms = ~700mA_RMS at angle 0 degrees

Since this is a series circuit, the voltage drop across each element is simply the source current multiplied by each impedance.

V_Z_R = 700mA(171ohms) = 120V at angle 0 degrees

V_Z_L = 700mA(j1200ohms) = 840V at angle 90 degrees

V_Z_C = 700mA(-j1200ohms) = 840V at angle -90 degrees

As you can see, the 840V drop across the reactive elements is WAY above my initial design that can only handle 120V max.
(Now if you were to imagine that I have 3 reactive elements in series, and 3 resistors in parallel, each of the reactive elements will have a reactance of 3600 ohms, and the resistor will have a resistance of 52 ohms, which would produce a voltage drop across each reactive element of almost 7579V!)

So my question is, what do I do?

I was thinking that I could simply make a programmable attenuation voltage front end circuit similar to my current circuit, but what do I assume the maximum voltage will be? Is the above circuit a realistic representation of some load that may be used/measured in power systems?

We have a system on campus right now that assumes a source voltage of 120V_RMS, yet their measurement system has a maximum input voltage of 800V_RMS, so obviously this problem has been considered, but I don't know where to go to justify their value of 800V.

I apologize if this topic has been addressed. I'm on a tight schedule and didn't have time to research the forum. Any input would be appreciated!

Thanks,
Jonathan McCrory

 

[jim hardy]

Well you have discovered Q of a resonant circuit. Ratio of energy stored to energy dissipated per cycle, also conveniently X/R ..

Observe Q = X / R = 1200 /171 = 7.02
and voltage gain = 840 / 120 = 7

Basic freshman EE: "The voltage gain of a series resonant circuit can kill you."

So my question is, what do I do?

Lower your input voltage at least tenfold .

And NEVER play with line voltage resonant circuits until you've thought your experiment through.

Glad to see you're in a school lab setting where hopefully there's safety conscious oversight. You don't want 740 volts at almost an amp just laying around on a benchtop.

old jim

 

[Jonathan Mc]

Ahhh. That makes a lot of sense now. 60Hz IS the resonant frequency. duh!

Thanks for your response. The only issue is, I cannot lower the voltage for this application. The project is to observer voltage and current magnitude and phase relationships at a source voltage of 120V rms. I suppose I am going to have to go with the attenuation circuit, even though this is a worst case scenario.

Thanks again!

 

[donpacino]

Is this in a prebuilt power lab?
I know in my undergrad we had setups with plexiglass between us and the power systems.

 

[Jonathan Mc]

Is this in a prebuilt power lab?
I know in my undergrad we had setups with plexiglass between us and the power systems.

Our power lab has the old Hampden Benches from the 60s. But we have one module from LabVolt. My project is to build a data acquisition system that is cheaper then the LabVolt system that we have. (Not as good though).
https://www.labvolt.com/solutions/6.../59-8010-A0_power_electronics_training_system

 

[jim hardy]

I suppose I am going to have to go with the attenuation circuit, even though this is a worst case scenario.

Be aware standard resistors have a maximum voltage related to their physical body size. 400 volts is about max for 1/4 watt , check your resistor datasheets. It's a volts-per-meter stress limit not a thermal power dissipation limit .
You will likely have to series a few resistors to stay safe from overvoltage breakdown even though you're well under the thermal power limit for high ohm values..

Cover that thing and do not leave it unattended. What you have described is a lethal contraption.

old jim

 

[donpacino]

was thinking that I could simply make a programmable attenuation voltage front end circuit similar to my current circuit, but what do I assume the maximum voltage will be? Is the above circuit a realistic representation of some load that may be used/measured in power systems?

To answer your original question yes this could work. something like the link below, but higher power could help you, hint hint nudge nudge

https://www.grainger.com/product/40...ice^c-plaid^96863074677-sku^40JZ72-adType^PLA

 

[donpacino]

note: I don't know if they sell high power versions, but you could make one custom for your needs (Just not as many options). Another idea is to make a second transformer with a different windings ratio

 

[donpacino]

Talk to your advisor or professor, there are a lot of solutions, finding the best one depends on the resources you have available to you, and your constraints!

 

[Jonathan Mc]

To answer your original question yes this could work. something like the link below, but higher power could help you, hint hint nudge nudge

https://www.grainger.com/product/40...ice^c-plaid^96863074677-sku^40JZ72-adType^PLA

The circuit that I am going to build will be an instrumentation amplifier with several attenuation stages to scale 800Vrms down to 10Vp. The gain will be auto adjusted in LabVIEW. The students (this project is for upgrading the lab) will not be expected to set the gain manually. I have a prototype built but It is scaled to 120Vrms.
Thanks for your responce!

 

[Jonathan Mc]

It's a volts-per-meter stress limit not a thermal power dissipation limit .

Can you show me some documentation or guide me on this volts per meter thing? This will definitely be important when I choose my resistors. I cannot seem to find this parameter on a datasheet.

 

[donpacino]

The circuit that I am going to build will be an instrumentation amplifier with several attenuation stages to scale 800Vrms down to 10Vp. The gain will be auto adjusted in LabVIEW. The students (this project is for upgrading the lab) will not be expected to set the gain manually. I have a prototype built but It is scaled to 120Vrms.
Thanks for your responce!

Ahh ok that should work and I like it.

With a project like that, adding hooks into the labview program for the students the manually adjust the gain (if they choose) might be beneficial as well. As in change the gain via constants or selection in the program. Allowing students to see what happens if you reduce the gain too much (wasted resolution and decrease accuracy) would be a good lesson. With a system like this flexibility is key. Labview is easy to allow for that sort of thing.

Do you know about creepage and clearances? https://resources.altium.com/pcb-design-blog/high-voltage-pcb-design-creepage-and-clearance-distance

 

[jim hardy]

Can you show me some documentation or guide me on this volts per meter thing?

There's an old PF thread on that same question https://www.physicsforums.com/threa...sistor-power-rating-and-rated-voltage.735180/

and here's a snip from a resistor datasheet linked in that thread

as you see, the 0.25 watt resistor has a maximum voltage of 250.
Now, 250 volts across one megohm = 250^² /10^⁶ = 0.0625 watts, ¼ its rated power.
That's a volts per unit length limit on that resistor body .

and here's what i wrote in post # 9 of that thread:

Place yourself for a moment in the shoes of manufacturer.
You have a line of resistor bodies in various body sizes, each capable of dissipating Xsize watts (or milliwatts)
Each size offers resistance values ranging from probably one ohm (or 0.1 ohm) to megohms. And by size is how they're sorted in your catalog.

What is the limiting factor for each body size - watts or volts per inch of the internals ? observe dlgoff's arcing, that's a 1000 word picture if ever there was one. But only in high ohm values will you hit the volts-per-inch limit before the watts limit.

It's unusual in garden-variety circuit design to bump against the volts-per-inch limit because most consumer electronics doesn't have high enough voltage.
So it's not emphasized much in our education .

But the prudent designer is aware of it as just another limit to be observed.

see also http://www.vishay.com/docs/49873/49873_sg2113.pdf

Now you know what to look for in your resistor datasheets, voltage limit for the body size. .
Don't push high ohm resistors all the way to their power limit you'll let their smoke out.old jim