Voltage divider fail! / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Voltage divider fail!

Move Thread

LAN_403

PanosB_GR

Sat Apr 25 2009, 07:08PM

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

Well, here's my story of the day:

I decided to make a resistive voltage divider out of 6 x 1MOhm resistors, with a 50Ohm low voltage resistor, for matching purposes, i.e a 120k voltage division ratio.

These HV resistors are cylindrical, 18cm long, 32mm in diameter. According to manufacturer are good up to 70kV.

I stacked these 6 resistors in an acrylic glass tube of 32mm inner diameter, so I got a 1.1m high voltage divider, said to withstand 420kV.
At the lower end i put the low voltage resistor inside a small aluminum case with a bnc terminal and the divider was ready.

I set my marx to 250kV and the output was driving the voltage divider alone. The readout was a beautiful 2V pulse, until after 3-4 shots. Then, there was a large flash but (thank god) no burned instrumentation.

And what i saw afterwards, was actually a burn mark throughout the acrylic tube on the surface of the resistors and on their brass terminals. Wow!

Conclusion: Do not force surface flashovers where you dont want them!

likewhat

Sat Apr 25 2009, 07:16PM

Account deactivated by user request on 6/11/2009.
Registered Member #1071 Joined: Fri Oct 19 2007, 02:13AM
Location:
Posts: 44

You might wanna fill it with oil. also, what did the output look like, got any traces of it?

PanosB_GR

Sat Apr 25 2009, 07:22PM

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

I have the impression that the resistors are damaged. I didn't measure their resistivity yet, but i guess it is not wise to use them again. And i ran out of resistors as well...

The 2 traces that I saw and unfortunately not recorded had a sharp rise time and a veeerrry long fall time since the only damping was this megohm structure.

And btw i didn't get to see the waveform of the flashover, the instrument blacked out, and i was scared to death but then i turned it off and back on and *phew*.

Myke

Sat Apr 25 2009, 09:53PM

Registered Member #540 Joined: Mon Feb 19 2007, 07:49PM
Location: MIT
Posts: 969

It's not a good idea to measure voltages of marx generators because of the very fast rise time. This fast rise time can induce very large voltages in lines that aren't connected to the marx (iirc).

There was a thread about grounding marxes here:

Proud Mary

Sat Apr 25 2009, 11:03PM

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

Were they wire wound types, Panos? These is nothing wrong with wire-wound if they are of the bifilar non-inductive type.

May I suggest you would be much better using copper sulphate solution in your voltage divider? There is not much that can go wrong with it.

Choose a good wide tube, because increasing the cross-sectional area of the CuSO4 solution tube improves the non-inductive nature of an electrolytic resistor. If you make it out of an acrylic tube, then the tube acts as a resistor of near infinite resistance, which draws an infinitesimal current. This will attenuate the effects of coupling capacitance with surrounding structures and effectively makes the CuSO4 resistor into a shielded resistor.

Your measuring resistors would be best to be carbon types connected in parallel inside a screening can, with a coaxial output to your storage oscilloscope, and a 50R tail resistor beneath that, to attenuate reflections of the measurement pulse.

I hope this helps.

Antonio

Sat Apr 25 2009, 11:17PM

Registered Member #834 Joined: Tue Jun 12 2007, 10:57PM
Location: Brazil
Posts: 644

The problem was probably the acrylic tube. With air inside and some unavoidable corona the ionized air accumulates inside, reducing the breakdown voltage of the assembly. It would be safer to not use the tube, leaving the resistor string suspended, or to fill the tube with insulating oil, as already said.

PanosB_GR

Sun Apr 26 2009, 12:31AM

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

Myke wrote ...

It's not a good idea to measure voltages of marx generators because of the very fast rise time. This fast rise time can induce very large voltages in lines that aren't connected to the marx (iirc).

There was a thread about grounding marxes here:

Myke, yes, i started this one haha :)

and my marx scheme is working well so far. 6 stages-50 kV,100nF/stage. :)

except for the bad resistors-in-a-tube divider idea.

The fast rise time indeed produces high frequency shot noise and the challenge is to make the voltage dividers as shielded as possible (and use good cables too) so as to get any useful measurements from them.

Harry wrote ...

Were they wire wound types, Panos? These is nothing wrong with wire-wound if they are of the bifilar non-inductive type.

May I suggest you would be much better using copper sulphate solution in your voltage divider? There is not much that can go wrong with it.

Harry, they say they are non inductive wire wound, so if they are still good, that would be nice. I could try with CuSO4, a low resistance load that i already made seems to perform well.

And btw, its indeed a bit tricky how to make a "good" low voltage side...

Antonio, the acrylic tube was most likely the problem, it probably initiated flashover due to high static charge accumulated or triple joints across the inner surface.

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