Spark plug opened

[Peter Schmelcher]

I opened up a spark plug looking for some inspiration and repair parts for a broken HV feedthrough.

The plug is rolled shut. Under the rolled edge, which I cut off with a lathe, is a packing powder that becomes solid and chalk like. I prodded it out with a small screwdriver in 3 minutes. Near the spark end is a small copper washer positioned between the metal housing and the ceramic body. The glaze is limited to the outside and where the copper washer seals.

Just food for thought
-Peter

[Steven Sesselmann]

Peter,

Thanks for sharing, I wonder how good the seal is?

Steven

[aka47]

Hmmm just trying to visualise that.

You didn’t happen to take some piccys did you ??

Take that back helps if I read the original posting before replying.

The piccys don't come through on my email.

Sorry guys

[Chris Bradley]

In my experience, typically vacuum tight - to at least an essentially undetectable leakage rate at 10^-5 mbar. I know 'cos I have eight plugs in my construction!

John F has also commented on the same and indicated that in his experience one in several leaks. I did try to pre-check the ones I installed, but after a few hours the outgassing rate (from the new plugs' exposed surfaces to vacuum) were still overwhelming any leakage rate past the seal. I couldn't be bothered to wait and took the risk of installing all 8 in one go (if there was a leak, which one would be leaking?). Fortunately there is no evidence I can measure of any leakage (that is, below any measurement threshold I can practically apply - compared with outgassing rates).

I have been unable to fault spark plugs as a perfectly adequate occasional 30kV-capable feedthrough.

[Richard Hull]

Way back on Songs.com in the 90's, when I used a formula one racing plug adapted to fusor use, I noted that you needed to put some vacuum rated epoxy at the point where the outside insulator entered and metal plug body. Rough up the glaze in a tiny ringed area at the base so the epoxy will grip the insulator portion well.

Such operations are, in general, limited to demo systems and only the base, low level fusor operation.

Richard Hull

[Doug Coulter]

I've used plugs into the 50kv region, if the air side is properly "booted", and little trouble, just tap a flange for it.

Joe Jarski found and traded me some special plugs for oil burners - no resistor, and a super long center electrode (a few inches sticking out) which are even nicer, as is, no need to do fancy machining. I have had some tiny leaks as Richard mentions, easily solved with Hysol-C at the ceramic/metal junction on the air side.

[Peter Schmelcher]

I have a 70kV coaxial cable that has a 0.5” diameter rubber dielectric. The spark plug is 0.618” in diameter at the wide bit. Looking closely at the glaze it covers the entire outside to just past the large diameter edge by 1/16”, a good location for an indium vacuum seal. I would expect a 50 to 70kV rating from the pair. One possible catch is permeation. If that turns out to be a deal breaker I’ll repair the commercial feed through using the spark plug core.

[Doug Coulter]

The likely catch to an indium seal is it will melt in operation. My fusor gets pretty hot in that area.

[Peter Schmelcher]

Good point Doug. It's always easier to make changes in the design phase. Thanks!
I had postulated that a failure of the HV stalk might create a projectile followed by a bad very moment. My scrounging journey will end in a rather underpowered 10” Pyrex bell jar fusor with a stainless foil liner.

Does indium have good surface tension when liquid?

The OD of the spark plug is about the OD of 1/2" copper pipe and all sorts of plumbing parts are useful. A 3/4" to 1/2" copper pipe reducer looks good at this moment. The plug clearance is about 0.01”. I think I could roll the copper edge and make a proper mechanical seal, and use the indium as a backup. The 70kV cable will slip inside a 1/2" copper pipe with just a small amount of boring or reaming. I could make a heat pipe from the space between the concentric pipes so the indium seal and plug would never get hot.

-Peter

[Doug Coulter]

If you're stuck on the spark plug idea, I think you should look for this gem (H/T Joe Jarski who found them). See pic. They're cheap, they seal well, you don't have to do much, and there's plenty of rod sticking out to support a grid or allow for welding on even more. Standard threads and a washer should do it though say a bell jar baseplate. Chris has this working, as do I for my lower voltage/inexpensive stuff. Worst case, you use some long-time set epoxy (JB weld or my preferred Hysol-C) to stop any leaks. Either will take more heat than indium will (In melts at 156c but starts out weak and gets weaker fast).

Indium is cool stuff, but it's very weak even at room temperature. I bought some from rotometals for activation studies, and it wasn't hard at all to nick off a pea sized piece (with my thumbnail!) and pound it out 2" square with a 4" handle tiny hammer. Hot, it's about like water or solder - only it takes less heat than solder to melt. Neat material, wets everything, but not the right tool for this job - use epoxy, if anything.

I too use a lot of commonly available copper plumbing parts for vacuum work. I silver solder them so as not to have high vapor pressure stuff in my tank. I also etch off the surface layer, which tends to be loaded with gunk that outgases seemingly forever, even though it looks clean in there. Ammonia does this fine, or reverse plating.

Here's an example spark plug use written up by John Futter: http://www.coultersmithing.com/forums/v ... ?f=9&t=158

I think the oil burner plug is even better, though - less modification work required.

Don't forget - make whatever you make easy to replace, or you WILL be sorry you didn't.

And I'd be dishonest not to tell you what I use myself:
http://www.coultersmithing.com/forums/v ... p?f=9&t=72
If I didn't think that was best of all, I'd be using something else.

[Peter Schmelcher]

Doug the advice and links are all good. I think the siren’s call of the plug for me is EMI. I’ve read your posts about fusor screening, and I’m starting with glass. When I look at the spark end I see a male contact for a HV coaxial connector. Life is about the journey, I’m going to give it a go and see what materializes.

I’ll make a female boot out of the coaxial cable end so it will slip over the plug, and use a copper compression seal like the big boys. More complex than I wanted but C’est la vie.

-Peter

[Richard Hull]

If you must use a boot over the plug body, slather it and the plug up good with silicone HV putty. In general, these things are better run naked in air after a good alcohol rub down and cleaning before each run.

The spark plug boot, as is normally used in a car over the plug, is more about keeping the insulator free of conductive carbonized oil and grit fouling than insulation.

Richard Hull

[Peter Schmelcher]

Thanks Richard. I read some horror stories about silicon oil/grease contamination on the inside of fusor chambers and thought I would try and steer clear. The spark plug taper section is about 8mm long so air probably won’t work (typical value 30KV/cm), however the cable stands off 70KV with just 4mm of radial rubber. I’m optimistic that I can get a dry rubber-ceramic seal on the nose of the plug, mostly because it is tapered. I poked around the internet and machinists recommend diamond and carbide cutting tools, plus freezing during machining of the rubber. Personally I know that a regular grinding wheel is very slow, and it makes a huge mess. My 70KV cable is 0.03” over sized for my 50KV power supply HV female receptacle. I successfully reduced the cable diameter using a cylindrical carbide burr in a Dremel (with no freezing), and the resulting surface finish is good. Using a carbide ball burr I hope to machine a taper into the rubber that matches the spark plug. However it has been quite some time since I last wrote g-code.

Plan B is lots of silicon grease and now with your suggestion HV putty.

Some practical research on HV vacuum breakdown, talks about conditioning. http://alexandria.tue.nl/openaccess/Metis123209.pdf

The pic is an improvised HV Bertan 740A connector made from half of a 1/2" compression fitting. The braid is clamped using the olive.

[Chris Bradley]

Peter Schmelcher wrote:
> The spark plug taper section is about 8mm long so air probably won’t work (typical value 30KV/cm), however the cable stands off 70KV with just 4mm of radial rubber.

The way HV goes can be a mystery at first sight. Judging anything on '30kV/cm' is inadvisable.

Anything towards 10kV wrt ground on pretty much any 'non-spherical' electrode will begin to build up some degree of corona around it. At 20kV it is a pronounced effect on all connections not designed to avoid corona, such as smooth spherical surfaces. It is this ionisation of the space around the electrode that will lead to the breakdown. What breaks down is when the field strength very locally to the electrode exceeds this value, and judging how high that field strength may or may not get is guesswork at best. I'd say assuming anything over 5kV can jump a cm is not unreasonable, 20kV jumping an inch or more.

Anyhow, the upshot of all that is that you'll need some solid insulator between the conductors and parts at ground, as you have already discovered. Just a few mm of a good insulator maybe all it takes. This won't be the problem ... the next problem is always the next weakest point, and that will be where the cable/rubber connects to the spark plug. Just a teeniest of gaps will lead to an electric field 'line' running between the electrode and ground, and if that field exceeds the breakdown value of the space then you'll get a discharge. It may even only be a 'little corona leakage' to start with, but that will degrade the insulator around it leading to a bigger failure.

So the key point to focus on is that interconnection. In this regard, I am limited only to comment on the things I have tried myself - the taper at the end of the plug is ideal for finding adhesive shrink-tube of the right size and working it on as far as you can. The sparky-end ceramic usually has a rough texture too, which helps the heat shrink key into the ceramic. I used repeated layers of heatshrink with a sandwich of kapton tape.

You can see the construction I used here;
viewtopic.php?f=11&t=4893#p32037

Anyway, one way or another you will need to put a lot of thought-power into accomplishing that DIY interconnection which I suspect will be the main weakness. As you can see in mine, I simply used '600V' hookup wire - that's good enough to suppress the local corona, so as long as you keep good clearance and keep it well away from making physical contact with anything grounded then it seems to do the job OK. That made it easy to create a solution (I used a piece of PE dielectric from a coax cable and threaded the wire through it, which gave the heat shrink tube the right size to shrink on to). How you accomplish it with you wider cable will likely be problematic. It will be interesting to see the solution you arrive at.

[Peter Schmelcher]

Chris I searched this site, read all the spark plug related stuff, and I liked your solution. The Metis link mentions the 30KV/cm in passing and I guess I should have pointed out that this is the maximum electric field strength that air can support. For any interested I believe that hyperphysics has a field calculator for a few simple electrode geometries.

The cable center conductor is 0.2” in diameter and will contact the plug on the radius at the start of the tapered ceramic section. The KIS plan is grind a taper on the electrode and drill a slightly smaller hole in the cable center conductor. As the cable is pushed over the electrode the cable will bulge slightly and make a contact. The spark plug tapered section will get a similar approach. I should mention that the ceramic core is a high precision part. The cylindrical regions do not have any measurable taper.

Reflecting, I think that ball burr machining of the taper will end poorly. I can collet hold a 5” length of cable. This is long enough to get into the metal box for shielding the current limiting resistor. Turning the cable will result in the most accurate and the best surface finish. I just need to grind a tool bit for boring.

[Richard Hull]

I thought the boot you were talking about was on the outside! You will have far worse issues with outside arcing than with inside if you use a spark plug. All my adivse was for the use of an insulating boot over the outside of the plug much as in an automobile.

As for inside insulation.....I have found that only an alumina tube of the correct diameter and length to be retained by a screw-on grid will assist to any significant degree. A spark plug should just not be considered at all over about 15-20kv.

Richard Hull

[Peter Schmelcher]

Your initial assumption was correct, spark end outside, cable-boot over spark end plus 1” of alumina. My guess is the spark plug alumina core can support 50KV. The original spark plug copper sealing washer has an internal radius of 0.125”. The redesigned seal radius will be about 0.3”. I figure the failure mode will be alumina surface flash over. So if I get a good cable-boot fit/seal it will work fine on the outside, on the vacuum side I have no experience, a potential later project. My biggest concern currently is the rubber cable dielectric. It’s difficult to get a smooth surface finish on rubber.

It’s not over till the fat lady sings.

-Peter

[Chris Bradley]

Given that you have indicated it'll all be going in a pipe, it is probably advisable that you see if you can construct it in a way to fill it with oil. Depending on the 'angle' of installation, you may need an additional pipe that runs vertically as an 'expansion' chamber so to speak, where all the bubbles can end up.

My double-layered adhesive shrink-wrap set-up appears to be stable to 30kV for short periods, but I can't imagine it'd hold up much more.

... now that I think about it, possibly an even easier way to do this is to attach a flexible PVC pipe, instead of the copper. You could then, even, use any old sort of wire through that pipe to the spark plug, seal the pipe to the spark-plug with a jubilee clip, top it up with oil and flush the bubbles out, hold the 'open' end vertical so the oil doesn't come out, and that end of the job is done!

[ed]

I gather from this discussion that the idea is to use the boot-connection end of the spark plug as a post inside the vacuum chamber, and make the external HV connection on the spark electrode end. The main problem is to get a good, reliable connection that can withstand enough voltage.

I don't know if this has been suggested before, but how about using two spark plugs lathed properly and mounted with their spark electrodes nose to nose within a threaded coupling? The chamber-side one would have its threaded section shortened as in the discussion, and would be threaded through the vacuum bulkhead, and provide the vacuum seal as normal, with an o-ring in a chamfer. If the bulkhead is thin enough, the spark plug thread length can be made such that some thread extends enough to screw on a coupling with matching threads.

The outside connector-end spark plug would have its threaded section shortened also, to increase clearance from the electrode, and it would be screwed into the coupling, bringing its electrode near or in contact with the other's electrode. If the coupling is just the right length, the electrode tips will just make a light compression contact - it could also be made with a slight clearance, and have a small spring or piece of compressible metal to bridge the gap. The volume of the coupling can be filled with dielectric such as a silicone insulating sleeve or silicone goop. Since the chamber-side plug provides the vacuum seal, the coupling volume will not see vacuum - but the fill material may even assist if any leakage does occur in the crimped metal to ceramic seal of the plug body.

The equivalent structure could be made by using a very thick section for part of the bulkhead unit itself, having a threaded through-hole deep enough to mount the two spark plugs. This style of construction would lend itself to easier maintenance, such as if the chamber-side plug bodies are consumed rapidly. The plug can just be unscrewed and replaced, without having to take apart the rest of the assembly.

To make this, or a coupling, the threads from each end only need to be deep enough to accommodate the remaining thread length of the plugs, but for versatility and experimenting, it's probably best to make them as deep as any maximum expected plug thread length, or even have a continuous through-thread. Since the thread is straight, not tapered, it should be relatively easy to tap clean through a fairly thick section. On the other hand, a smooth inner wall would be better, high voltage-wise.

This type of assembly would provide good mechanical integrity, and maximize standoff capability. The critical electrode joint would be fully enclosed and protected. The external spark plug can of course be nicely connected and protected by a booted HV cable made from standard automotive parts.

Ed

[ed]

One thing I forgot to mention - as long as the coupling volume does not have to seal anything like against vacuum, or hold liquid dielectric, the threads can be quite loose. This makes them much easier to cut by using oversized tapping holes. There should be plenty of mechanical strength to hold these small items together even with very loose fit - it's not like those plugs are going into a cylinder head. A little trial and error should determine the right size and compromise between strength and machining effort. The vacuum-side threads should be nice and precise, but don't worry about the others.

Ed

[Peter Schmelcher]

Chris Ed reasonable solutions.

I picked up $20 of material the other day, 2’ of 0.75” 316L SS tube, 1’ of 0.81” copper rod, and a few plumbing bits, stared at them long enough, and have decided how I am going to make the feed through. A single point diamond will shape the rubber boot. It will take me a few weekends to make the feed through. I’ll take some pictures along the way and post the results good or bad.

I searched the site for vacuum friendly silver solder but never found a specific recommendation. I have 0.001" pure silver foil that I could use straight or alloy with lead free solder or...

-Peter

[Roberto Ferrari]

Hi Peter

The standard vacuum friendly silver solder is the eutectic alloy 92% Silver-28% Copper, commercially available in many forms, mainly wires.

Cleaning the surfaces and then using a non-oxydizing flame will prepare the surface to be wet by the alloy, without using fluxes.

Regards

Roberto

[Peter Schmelcher]

Thanks Roberto. I picked up 1oz of sterling silver wire 92.5%Ag balance Cu.
-Peter

[Roberto Ferrari]

You are welcome, Peter

The advantage about the eutectic allloy is the lower melting point in order to not to stress too much any ceramics close to the brazing area.

good luck!

Roberto

[Peter Schmelcher]

Just an update.

Not yet HV tested. The structure went through a couple of changes. This design is unlikely to ever fail catastrophically in my glass bell jar.

The spark plug to copper to stainless seals are great and I can't detect any leak or additional out gassing. The two copper rings are made from 1/2" copper water pipe.

The matching spark plug HV connector needs some better tooling before I can proceed. The issue is the small diameter of the spark end and making a diamond tipped boring bar small enough for interior connector profiling.

BTW the 2.75" copper gasket was rejuvenated with solder and then used in the test.

-Peter