water loads as high voltage resistive dividers / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

water loads as high voltage resistive dividers

Move Thread

LAN_403

PanosB_GR

Thu Sept 11 2008, 08:16AM

Registered Member #1680 Joined: Fri Sept 05 2008, 04:19PM
Location: Greece
Posts: 43

This is my first post here, and so to begin with I say hello to everyone.

I would like to design a resistive voltage divider for a marx bank (under design also!).
I was told that resistive dividers are better for fast rise time waveforms, for example the driving of a marx bank onto a resistive load with no waveshaping resistors in the marx bank. I understand that all parasitic inductances might worsen the risetime, but, anyways, let's leave that for now and assume that a fast waveform will be achieved.

So, a cheap way to build a resistive divider is that of using water loads, which is pipes filled with a salty solution that provides some conductivity.

I was having some questions about this practice:

The voltage output should need some 100.000 times reduction, so a multi-MegOhm water load might be needed.
Apparently, that should be a pure water column, or tap water column because i guess any salt might raise the conductance?
I know they say a lot about bubbles etc concentrating on the surface of the electrodes, but a copper tube inserted in the water on top would provide enough contact surface with negligible air-bubble effects imho.
The water column will have any significant unknown parasitic inductances or capacitances that may alter the waveform?
As the low-voltage side, a simple resistor would work well?

Not a big deal of questions, but water loads seem an attractive technique, so any ideas welcome!

Regards

Dr. Slack

Thu Sept 11 2008, 01:39PM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

Two things - resistive divider for good waveshapes, and water resistor for divider.

The outupt impedance of a resistive divider is going to be pretty much the bottom resistor, this will be driving the parasitic capacitiance of your scope. It therefore has to be very low to maintaina good bandwidth. If you have a scope with a 50R input, then you'll get good waveforms, but need a very low Z divider. Note that the top big resistor will have a capacitiance across it, which into a ressitive load will peak, so you really need to design the whole divider so that both the resistive AND the capacitive pot-down are equal. Inductances will have an effect at Marx speeds, but unless you have a really tasty scope, you probably won't be able to see those effects.

Water would be a lousy matieral for making the top resistor of a resistive divider. If you try to go for a high value using pure water, the value will be extremely unstable as the water contaminates itself from the electrodes and the air, it could increase conductivity an order of magnitude or two just from that. If you use a reasonable concentration of solute to get a stable conductivity (barring temperature effects of course), then you will need a very long thin pipe to get the resistiance you need.

You could (and I'm ignoring the capacitive aspects here, so this is really a DC and low freuqency solution) build a nice resisitive divider by making BOTH top and bottom resistors from water, and circulating the water continuously between them so that any variation in conductivity occurs in both at the same time. But that could get messy.

High voltage probes, especially high frequency ones, are rather more difficult than they seem at first. Commercial companies charge a lot of money for good ones.

Proud Mary

Thu Sept 11 2008, 02:06PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

You might find this helpful:

"For measuring high voltages, a shielded copper sulphate (CuSO4) resistive voltage divider having divider ratio 1000:1 is used. The construction of this is as follows: The divider consists
of two resistors R1 and R2 connected in series. High voltage is applied to one end of R1, while the other end of R2 is grounded. Voltage is measured across R2. To avoid reflections
of measured voltage pulse, we introduce a 50 ohm resistor with one end of R2. The resistor R1, having value of 1 k ohm, is made up of dilute solution of CuSO4 in de-ionized water filled in a hollow perspex cylinder 300mm long and 24Â·1mm in diameter. Increased crosssectional area enhances the non-inductive feature of this resistor. The hollow perspex cylinder acts as a resistor of infinite resistance, which draws infinitesimal current. This suppresses the effect of coupling capacitance with the surroundings and thus makes the resistor R1 a shielded resistor. The other resistance R2 is made up of 10 non-inductive carbon film resistors, each of value 10 ohm, connected in parallel. The effective value is one ohm. The R2 resistor assembly is enclosed in an aluminum cylinder having outer diameter 75 mm, inner diameter 66mm and length 100 mm. This metallic enclosure for the R2 resistor gives an EMI shielding to resistor R2. Attenuated signals from this divider are fed to a digital storage oscilloscope."

Sadhana Vol. 30, Part 6, December 2005, pp. 757â€“764

Development of a 300-kV Marx generator and its
application to drive a relativistic electron beam
Y CHOYAL, LALIT GUPTA, PREETI VYAS,
PRASAD DESHPANDE, ANAMIKA CHATURVEDI,
K C MITTAL+ and K P MAHESHWARI

PanosB_GR

Fri Sept 12 2008, 05:47PM

Registered Member #1680 Joined: Fri Sept 05 2008, 04:19PM
Location: Greece
Posts: 43

Well, i might build a divider like the one described by those indian guys. A 1000Ohm:1Ohm divider will draw some current from the bank, so I believe a 10.000Ohm/1Ohm divider would do just fine. Since I will drive that to a 50Ohm Oscilloscope, it's impedance parallel to the 1Ohm low voltage resistor would keep the low voltage side close to 1Ohm. I will add some attenuators after that to keep the instument safe :)

10kOhm load is not big, and i reckon it will need a moderately salty water column, and probably the value of the resistor won't be changing a lot from shot to shot. CuSO4 with copper electrodes seems nice choice.

Circulating the same water solution into two water loads will keep the ratio definately stable, but to make a 1000:1 divider ratio, someone would need a really fat low voltage side and a tall and thin high voltage side... The ratio just doesn't seem easy to achieve with commercial water pipes.

Thanks for the tips! :)

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob