1. Home
  2. Electronics Design
  3. Problems with carbon composite in a HV string.

Problems with carbon composite in a HV string.

We have this situation where a section of a string of 22Meg@ 2 watt carbon composite resistors may physically split down the sides there for both lead wires of course become detached. This may happen to several in the same section string.

This is a problem that has recently sprang up after a replacement of several sections that showed signs of over heat due to many hours of service and other events, but not splitting.. So I think it has something to do with the way the new strings are being assembled and installed.

To let you get an idea of how this is being used..

This is the resistor string inside of the rectifier stack that generates the 1 to 2 Mev volts for an accelerator. When the oscillator is at full potential on this unit, a max of ~ 2Mev with no loading is achieved and it'll drop down to 1.5Mev loaded. The resistor string connects to the High end and spirals down the center to the back end of the vessel, where it comes to a J-box which has a spark plug there gapped for 5th of an inch. This comes into play only if for some reason the output load is lost in the metering circuit, for example.

Each section has 8 22Meg 2W carbon composite resistors and they wrap

180 degrees around the center. there are delrin stock going down the center where the connecting post are to join these 180 degree sections together for support.

In the past to my knowledge there has not been an issue of this kind, even if the unit is arching out, etc. But recently after the techs replaced a few of these strings due to what they say was signs of aging with heat spots, they now have these sections blowing apart and adjacent sections that were not replaced, when ever the spark gap gets a collapse. We know there is external issues with the metering circuit and that will be fixed later, but it shouldn't be causing this.

Is it possible these halves are suppose to be phased magnetically so that when impulsed current appears in the string they null out physical movements?

Or is there something else?

P.S. There are a total of 288 22M 2W resistors, each half radius having 8. each R gets no more than 1.1 Watts at full voltage output, so this does not look like an over current problem.

Jamie

Reply to
Maynard A. Philbrook Jr.
Loading thread data ...

On a sunny day (Sun, 4 May 2014 12:52:30 -0400) it happened "Maynard A. Philbrook Jr." wrote in :

Carbon composite has this tendency, you should use film if at all possible.

Hundreds if not thousands of TVs failed in the past with carbon composite resistors doing things like shorting (danger) and splitting (going high).

I think here they stopped using those 40 years or longer ago?

Right!

formatting link
scroll down to carbon composite for failure modes.

Reply to
Jan Panteltje

On Sun, 4 May 2014 12:52:30 -0400, "Maynard A. Philbrook Jr." Gave us:

Carbon comp is absolutely the right choice.

The problem is that they age about as fast as a top fuel dragster piston.

Unless you want to parallel up about ten HV film resistors to match each element of your string (many tens of resistors) and be prepared for constant replacements of those which have opened up.

With carbon comp, each impulse gets properly absorbed / passed/ dampened/current limited/whateveryouwanttocallit by the string. But each impulse ALSO fuses some of those loosely bound lattice elements as it bounces/fires through the 'comp' medium. So it should slowly show an adulteration toward lower resistance with use.

Prep the assembly for a method of changing out the strings and consistent operation should return when you incorporate a regimen of constantly swapping it out every few events. Or every ten... or hundred. The quality of the resistor matters too. Declaring 2W and actually being able to take that level and remain efficacious in operation could elude some modern schlock house 'makers'. Carbon comp is hard enough to find as it is. High values of good quality likely even harder.

Individual 'bad' elements of the 'old' strings could still be culled out and the string refurbished into acceptable operation.

I would start by swapping in a new string and examining the old.

Document the new one before installation, and then again after a certain number of operational cycles.

Reply to
DecadentLinuxUserNumeroUno

Possible issue: carbon comps can bake themselves, which makes the resistance go down, which can be a runaway failure. I think that's a greater hazard for high ohm values.

You might consider a different type of resistor, spiral-cut thick or thin film maybe. Ideally, something designed for high voltage apps. Some air flow would help cool them, too.

Not many people still make carbon comp resistors, and maybe it's a lost art.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

Comps typically have a negative TC, which can make a series string unstable.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

It sounds like you are dealing with a familiar problem of high voltage circuitry, that any kind of corona discharge causes accelerated corrosion. The usual treatment is to add large-diameter electrodes so that the surface field (volts per inch) around the resistor string is lessened. Another possibility is to get larger-diameter resistors. It may help to coat the string with 'corona dope' (really, polystyrene varnish), or a modern silicone equivalent. Or, even to seal the resistors inside glass tubes with potting compound (or an insulating oil).

If it really IS corona, you might be able to see a glow when you dim the lights. That glow includes ionized gas, and ozone.

Reply to
whit3rd

You know, we'll find out tomorrow. I do think they are going to get me involved. I told them to make sure they did a physical shake test on the strings and then test them afterwards. I still think they are not installing the new strings correctly and are not accounting for thermal movement along with magnetic movement.

P.S. Looking around it seems that 22M 2Watt RC resistors are not to plentiful. I looked a Chinese site where they can make them but of course it's like a 5k piece min, which maybe just fine because we have 4 units, all using the same value R's and the small one uses

288 of them. Never hurts to have some spares!

Jamie

Reply to
Maynard A. Philbrook Jr.

On Sun, 4 May 2014 12:52:30 -0400, "Maynard A. Philbrook Jr." wrote:

--
 

http://www.globar.com/ec/bulk-ceramic-resistors 

John Fields
Reply to
John Fields

Caddock makes great HV resistors.

And some amazing 450 and 950 ohm axial resistors. Stick one in the end of an SMA female and you have a 6 GHz passive scope probe.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

I vaguly remember moisture being a problem in those resistors, there are newer moisture resistant CF's

Cheers

Reply to
Martin Riddle

[snip]

You mention a vessel. I take it that this is a gas insulated accelerator. If that is the case, then what you see can be caused by the new resistors not being gas tight. When evacuating the accelerator vessel, gas trapped in voids of the resistors "blows" them apart.

Regards, Klaus

Reply to
Klaus Bahner

THese R's are inside a vessel pressured with SF6 gas and the dew point is kept down to at least -35 min. The stock that is not in use is kept in a air tight bag and those bags are also kept in a air condition room.

But keep the suggestions coming, we're looking for ideas that have not been thought of, yet.

I see a large order of R's in our future, made to order..

Jamie

Reply to
Maynard A. Philbrook Jr.

It's most likely an over voltage problem.

At 1.1W, you're developing almost 5 kV across each resistor... and worse with resistor tolerance imbalances. This is too high for standard 2W carbon comp resistors under quasi-DC stress. When overvolted, they'll eventually internally flash-over/arc across the resistive element, splitting the body. Charring then causes their resistance to drop, further stressing the remaining good resistors in the chain, leading to cascade failures.

Substitute HV metal oxide film resistors rated for at least 6 kV continuous to fix the problem and provide a bit of voltage margin.

Bert

--
Bert Hickman 
Stoneridge Engineering 
http://www.capturedlightning.com*********************************************************************** 
World's source for "Captured Lightning" Lichtenberg Figure sculptures, 
magnetically "shrunken" coins, and scarce/out of print technical books 
***********************************************************************
Reply to
Bert Hickman

It rains a lot in China.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

Is this system in a dielectric gas, like sulfur hexaflouride? Is it possible the resistors are having problems due to gas being absorbed inside and then rupturing the package when the pressure is released? (Just a thought for an outside-the-box problem, not related to electrical environment at all.)

And, if the system runs in vacuum, then you have the pressure changes in the opposite direction when it is pumped down, as well as problems with unwanted charging and gas discharges in strange places.

Physical movement? Seems unlikely the resistor strings would have enough current such that the magnetic fields would move them much.

Is it possible they are arcing to some OTHER component in the system, such as the rectifier strings? That might put a lot of voltage across a shorter portion of the total resistor string.

Jon

Reply to
Jon Elson

Ah! I suspected they might be. You might want to put a sample of the new resistors in a chamber and pressurize with SF6, let it soak for a while and then depressurize rapidly. If they burst, you have your culprit. Any voids or even porosity in the resistors will make them susceptible to bursting when depressurized, and it may take extremely slow pressure reduction to allow the thick gas to diffuse out slowly. My guess is typical carbon comp material is pretty porous.

It is possible you need to bake them for several days before assembling into the strings.

Yes, but you may have to qualify them to some spec, and you don't even know the failure mechanism (yet).

Jon

Reply to
Jon Elson

An early poster gave me the answer I was looking for and I then did some test over here tonight.

The problem is over voltage, even though the resistor is not getting over currented or wattage etc..

I put a 8kv variable power supply across a 22MEG 2 W R and at 3kv all is normal, at 4k all is normal. However, starting at around

5500v the RGA starts to detect ozone when getting very close to the body but it is still ok. However, at 6kv I don't see anything significant via the IR but wait about 20 secs and pow! The phenolic case will split open with no charring, no smoke, no signs of heating.

Interesting what pressure does to a sealed unit :)

So I guess the only fix is to correct the machine from running away out of control voltage wise, which is currently what is causing this. With normal operations, the voltages will only reach about 4k max per resistor on this unit because its a smaller unit but the power supply is capable of generate much more than that.

I may offer to install a secondary over voltage sensor to switch things off.

Jamie

Reply to
Maynard A. Philbrook Jr.

How old is the resistor stock and have you measured the resistance? I have a fairly large collection of antique carbon comp resistors (mostly Ohmite) left over from a radio service shop. My guess is 50 years old. Most of the resistors have drifted erratically off tolerance, all towards the high end. Baking the resistors moves them back down in value, but they soon drift up again. My guess(tm) is the carbon "rod" inside is crumbling. More:

I don't have a clue what an increase in value might do to your arrangement, except maybe promote arcing.

The only real high voltage experience I had was fixing a first generation image intensifier tube power supply. Only 1.5KV as I vaguely recall. The resistors were carbon film inside a glass tube. Someone had gotten their fingerprints all over the glass, causing leakage currents. Problem solved with an alcohol wash. Since you're changing the resistors constantly, has there been any effort to keep fingerprints off the resistors? If one arcs, the remaining resistors in the string see a higher voltage, which might cook them.

My crystal ball sees a redesign for Hi-V carbon film resistors or you will probably be repeating the replacement exercise.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558
Reply to
Jeff Liebermann

It is rare tat a carbon comp splits, but it always is due to a current surge that heats the core; a slower current impulse would allow more uniform heating with a typical "blow-out" or cracking. I think the Allen Bradley brand gave the best resistance to that. Experiment with thin films?

Reply to
Robert Baer

Wiki is in error; typical tolerance was 2% low with about 1% or better spread (AB brand,natch).

Reply to
Robert Baer

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.