HV Resistors for HV measurement.... / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

HV Resistors for HV measurement....

previous 1 2 3 4 next

Move Thread

LAN_403

radhoo

Mon Dec 06 2010, 10:28AM

Registered Member #1938 Joined: Sun Jan 25 2009, 12:44PM
Location: Romania
Posts: 701

Probably they can take much more than 11KV in air.

Proud Mary

Mon Dec 06 2010, 10:57AM

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

We all know that resistance can vary with voltage - especially high voltage - but quantifying dR:dV empirically might involve access to HV calibration standards not easily obtained.

I do not know the magnitude of the changes to be expected in resistors of differing compositions, surface treatments, temperature, humidity etc, and how significant those variations might be for given uses.

If, as I suspect, the resistance changes are a matter of a few percent or less, then they can be disregarded for most practical purposes, but if it is likely they might be larger, then they might need further investigation. There is also the question of whether such resistance changes are permanent, or exist only when the high voltage is applied.

Patrick

Mon Dec 06 2010, 11:06AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Proud Mary wrote ...

.....If, as I suspect, the resistance changes are a matter of a few percent or less, then they can be disregarded for most practical purposes, but if it is likely they might be larger, then they might need further investigation. There is also the question of whether such resistance changes are permanent, or exist only when the high voltage is applied.

Yes, this is what Xicon and Caddock warned me about, and since these are measuring circuits tolerances, accuracy, precision all matter more then usaull.

Patrick

Sat Dec 18 2010, 04:00AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

new pics:

This is again the basic design, all-thread rod, soldered to leads.

Here's the lesson learned from the previous attempt:
The Problem:
-- When filled with oil, then pressing the cap on caused oil pressure to rise, eventually bursting through the all-thread/epoxy endcap.
The solution:
--I put the whole thing together with epoxy, and without oil. Then after the epoxy setup I drilled and tapped a 1/4-20 screw hole, then filled the body tube with oil. With proper headspace the hole was sealed with a nylon threaded set screw.

Next, the area was cleaned with Acetone. More epoxy prepared and a curved PVC "cap-patch" was put over that nylon screw. Presto! liquid tight!

The labels are temporary. I will make better ones with my laser printer, then laminate them.

Steve Conner

Sat Dec 18 2010, 07:58PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Patrick, I don't know what the problem is with those resistors, but I bet it's not inductance.

The reason I say this is, their resistance is so high, it would take a huge amount of inductance to be significant compared with it. For a 5 degree phase error you'd need about 3M ohms of reactance, which is something like 50 Henries at 10kHz. To mess up a 60Hz measurement you need even more still.

I don't see how they can possibly be that inductive. Maybe if they were wirewound with millions of turns of microscopic wire, but nobody makes a 200M resistor that way.

Stray capacitance is FAR more likely to be your problem. I don't know how to calculate that, but I bet less than 1pF would screw up your measurement at 10kHz.

What are you doing with a bottle of Nukie Brown in California?

Hazmatt_(The Underdog)

Sat Dec 18 2010, 10:28PM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

I also consulted an engineer about directly measuring very high voltages, e.g. a divider to measure a TC output and scale it for an O-scope.
What I was told is that the stray capacitence would be so great with the resistor chain that I was thinking of ( couple of gig-ohm) that the capacitence would load the source, and that there would be no appreicable output at the end of the divider.

Also understand that at high voltages you are adding capacitence with the container (PVC), the dielectric, and whatever else the resistor comes in contact with.

If you go to Caddock's (think that's how its spelled) web page you can see how they make the resistors. They are non-inductive carbon track that winds back and forth over a ceramic core.

I had pretty good success with a capacitive divider for low freq stuff, and they're not as expensive as precision high resistance devices. So you might want to give that a try if you have a bunch of 1kv caps lying around, it does work.

Patrick

Sat Dec 18 2010, 11:24PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Steve Mcconner:

Steve McConner wrote ...

Patrick, I don't know what the problem is with those resistors, but I bet it's not inductance.

They actually, from o-scope data, seem fine.

Steve McConner wrote ...

The reason I say this is, their resistance is so high, it would take a huge amount of inductance to be significant compared with it. For a 5 degree phase error you'd need about 3M ohms of reactance, which is something like 50 Henries at 10kHz. To mess up a 60Hz measurement you need even more still.

Yes, after a quick scribble of math, I see you are right, for the 75 Meg and 200 Meg.

Steve McConner wrote ...

I don't see how they can possibly be that inductive. Maybe if they were wirewound with millions of turns of microscopic wire, but nobody makes a 200M resistor that way.

No, i can see through one of them, the epoxy/glass is kinda thin, but it looks like a spiral of ruthenium oxide, 50-100 turns of it across about 1.5 inches of length.

Steve McConner wrote ...

Stray capacitance is FAR more likely to be your problem. I don't know how to calculate that, but I bet less than 1pF would screw up your measurement at 10kHz.

Yes, the capacitence keeps me awake at night ! worry - worry - worry !

Steve McConner wrote ...

What are you doing with a bottle of Nukie Brown in California?

Get buzzed, you brits must be pretty lame if you cant figure that out. Pass the BuckFast ive got a football game to attend...

************************************************ ***********************************

HazzMatt:

Hazmatt_(The Underdog) wrote ...

I also consulted an engineer about directly measuring very high voltages, e.g. a divider to measure a TC output and scale it for an O-scope.
What I was told is that the stray capacitence would be so great with the resistor chain that I was thinking of ( couple of gig-ohm) that the capacitence would load the source, and that there would be no appreicable output at the end of the divider.

you need proper spacing, shielding and counter-capacitence at the low end of your divider. (and RC-RC divider like here:

)

Hazmatt_(The Underdog) wrote ...

Also understand that at high voltages you are adding capacitence with the container (PVC), the dielectric, and whatever else the resistor comes in contact with.

Yes, the two you see above have about 2-7pF if wrapped in foil.

Hazmatt_(The Underdog) wrote ...

If you go to Caddock's (think that's how its spelled) web page you can see how they make the resistors. They are non-inductive carbon track that winds back and forth over a ceramic core.

The Caddock ones have the same inductance as a striahgt wire of the same dimensions. i was on the phone with Richard Drawz (Caddock's app engr) for several hours. Caddock makes the primo stuff, if you can afford it. Its not the Caddock ones that are the problem, the EGB, and Vishay / Dale ones are more iffy.

Hazmatt_(The Underdog) wrote ...

I had pretty good success with a capacitive divider for low freq stuff, and they're not as expensive as precision high resistance devices. So you might want to give that a try if you have a bunch of 1kv caps lying around, it does work.

Yes, please post here how your C dividers are made, or start another thread. I'd like to see it. The resistors are expensive/hard to find... so I like the cap dividers better, when possible.

Pinky's Brain

Sun Dec 19 2010, 02:17AM

Registered Member #2901 Joined: Thu Jun 03 2010, 01:25PM
Location:
Posts: 837

Since you are already putting so much work in per resistor why not just create the capacitance from the shielding? By packaging the resistor like this (black insulation, grey metal, brown resistor) :

Patrick

Sun Dec 19 2010, 02:44AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

that might be too much capacitence, most shields are far away, and the body/lead makes one plate with the shield being the other.

Pinky's Brain

Sun Dec 19 2010, 02:59AM

Registered Member #2901 Joined: Thu Jun 03 2010, 01:25PM
Location:
Posts: 837

How can that be too much capacitance? If you don't put any metal horizontally along the horizontal divider it has exactly the same capacitance as your present shield (from the point of view of the source). It just puts a much better external field distribution across the resistor for AC measurement than your present shield.

previous 1 2 3 4 next

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