High Voltage Resistors

Well, um, high voltage resistors, actually. Check out Ohmite (acquired and assimilated Victoreen, a pioneer of high voltage/high value resistors) at (for example)

and Caddock at (many different kinds that can handle what you want).

But be aware that with a 1ns rise time to 8kV, making a circuit that actually looks resistive will be practically impossible. Note that at

250MHz (1/4 cycle = 1ns), 470kohms reactance is about 1.35fF of capacitance. I suppose that even the Caddock MG745 series, at 6mm diameter and 76mm long, will have well over twice that capacitance from end cap to end cap, and that doesn't count any of the rest of your circuit. So how you actually build things will matter a lot: you can perhaps put the resistor through some sort of shield plate with a hole in it if you need to reduce capacitance. Be sure the hole is large enough to not allow arcing! It's up to you to get all that right, to do whatever it is that you need to do. If you really understand what it is you are trying to accomplish, you might do well to contact the resistor manufacturers directly, for their recommendations.

Cheers, Tom

Ohmite Slim-Mox series (thick film planar); Digikey carries 400K and

500K 1% 5KV, 10KV and 20KV in the $5 region each. Good power rating choices. Mouser lists a wider variety, but for some strange reason has decided to not stock them anymore (gack!). 1nSec is slow enough that you should not have major problems using a planar as compared to an axial which could be used in a coax configuration for better pulse fidelity.

On Mon, 12 Jun 2006 18:41:29 +0200, "Juan" Gave us:

You want a minimum 2 Watt Carbon composition resistor (not sure of your stored energy level), which is what is used as an arc suppressor in HV power supply designs. The bulk resistance medium is the key. That is what will make it last a good long time. Try between 100k and

500K Ohms.

If you buy an HV resistor, they are made for dropping a specific voltage, not arc suppression. In fact, arcs will cause them to fail... miserably.

That is if you plan to use the resistor to quell an arc or discharge the set.

If you are making a dropping circuit so that you can monitor voltage, do a search on "Flatso" resistors. You want at least a 10kV, and I would recommend at least a 15kV resistor The two or three inch form factor variety are suitable for those voltages typically.

On Mon, 12 Jun 2006 20:49:55 -0700, Jamie Gave us:

HV resistors are metal film or thick metal film (whether they are on a rod or flat silica plate). Carbon composition resistors are used as arc suppressors in HV designs as they can handle several arc strokes without appreciable internal fusing or damage.

carbon rod resistors. they look like large fuses.

--
Real Programmers Do things like this.
http://webpages.charter.net/jamie_5

Avoid typing the same text again and again Stop wasting your time on mouse movements Open favorite web pages with a single hotkey press Record keystrokes and play them back with a single hotkey press

------------------------------

------------------------------ EnergyKey Save yourself from repetitive tasks

On 12 Jun 2006 22:46:04 -0700, snipped-for-privacy@yahoo.com Gave us:

ALL SPAMTARDS!

you're getting into the region where you might need to read "J.C.Martin on pulsed power", Plenum Press, 1996

The answer, of course, as I've often stated, is to use a capacitive divider. One can struggle to match the high and low-side resistive time constants if they wish, but for pulse work it's easier to just accurately measure the high-voltage steps. For example, consider a 10kV measurement cap with 10pF of capacitance working into 10nF for a 1000:1 ac divider. With a 10-Meg bias resistor to ground, it will nicely pass all signals above a 1.6Hz high-pass 3dB frequency. Note, low-value high-voltage capacitors are quite easy to make or buy.

--
 Thanks,
    - Win

Hi Win,

do they suffer from the same sorts of evil, nonlinear behaviour as HV resistors? I suspect not.

Cheers Terry

An old 1B3 type HV rectifier tube makes an excellent 30KV vacuum cap.

John

I saw a "vacuum variable" once - it was about 4" (10 cm) diameter and about 6" (15 cm) long, and had a bellows, and the electrodes looked like copper pistons.

This was in the military - I'm sure such a thing would be horrendously expensive, unless you have your own glass-blowing rig and metal shop. :-)

But, if we're still talking about capacitive HV probes, what about a piece of ordinary glass tube, plated inside and out? Just wondering if it's a PITA to plate metal onto glass?

Thanks, Rich

Silvering glass is a simple chemical procedure...

A capacitive divider works fine, of course, if you want to divide, but if you want 470kohms of pure resistance out to high frequencies, that's a horse of a different colour I fear.

C0G caps have a max 30ppm/C tempco, and are a lot cheaper than resistors of that stability. I suppose in a divider, the ratio will generally be even more stable than that. Bob Pease pointed out to me that you can get some rather large C0G's these days--DigiKey has 0.1uF ones, I believe. You could have quite a division ratio between 10pF and 0.1uF. Beware, though, at low values that the capacitance through the air around them becomes significant, and varies with varying barometric pressure and humidity.

Cheers, Tom

Terry Given wrote: ...

as HV

Some do; some do not. C0G (aka NPO) ceramics are pretty good. Vacuum caps per John L. are pretty good (though I suspect his 1B3 suffers some from microphonics). Non-C0G (high-K) ceramics can be awful, plus they are not temperature stable. Micas are usually pretty good, but I've come across micas that are very noisy. You can get C0G ceramics in small values up to many kilovolt working voltage ratings, and rated for high current as well, which could be important in high voltage RF work (though probably not for measurement setups). Metalization on each side of a quartz plate (cheap microscope slide, for example) will give a stable capacitance, but beware of corona off the edges of the metalization! Proper design requires some careful thought. Beware that charges can accumulate and stay on good dielectrics, so the shape of the electrodes and the dielectric become important for precision work...some lessons learned "the hard way."

Cheers, Tom

On 15 Jun 2006 01:09:58 -0700, "Tom Bruhns" Gave us:

We used to embed about 15pF in the circuit board for a bit of pre-emphasis across our 10M Ohm or whatever HV feedback resistor in one circuit we had. A pair of quarter inch by >1 inch "plates" on internal layers of FR4 works great! have to watch your keep outs though!

Resistor feedback is the most stable if one wants an accurate voltage reading. Capacitive feedbacks are quick to respond, but not as accurate.

15pF with 10M seems kinda high, making a 1kHz zero. I'm using about 0.2pF with 150M resistors, for a 5kHz zero. The 0.2pF is made with two 1910 pad sets, one under the other (the bottom two pads are shorted). The spare pad set on the top will take a 3kV cap if needed.

Properly made, with shielding, they can be very stable. I add an adjustable-gain output-buffer stage to allow for calibration.

--
 Thanks,
    - Win

Yes, with about 1.2pF to 1.4pf capacitance to the cap on the top. This is a scheme I've been using for six years or so, after you or someone else here suggested it, and after I grabbed an assortment of such tubes from various eBay sellers. I add a copper shield outside the tube to prevent stray wiring capacitances from supplementing the intended small "calibrated" plate-to-filament capacitance.

--
 Thanks,
    - Win

I have seen 200uF 200V C0G caps. Of course they are huge, and about US$1 per uF. Not for the limited budget.

OTOH I've had difficulty finding a 22nF 0603 C0G. That I've been prepared to pay for.

Cheers Terry