Capacitors in Interlas Tig box

[jo1010]

This has been lurking at the back of my garage for a while with leaking electrolytics ....there's a bank of 4 of them rated at 20kV and 1nF each. There don't seem to many options for electrolyrics looking around - at least not at any kind of sane price. I have seen ceramic disc caps of that spec available cheaply but I'm not sure how these would be combined in a replacement circuit - any ideas about how to do any of this ?
......The tig box was gifted as part of a workshop clearance so doesn't owe me anything....the HF does work but I've not run it for long with the caps in poor condition...... It's obviously had a life and is a simple thing really ...post gas flow is controlled by mechanical timer....the original gas solenoid has died and the gas is now controlled by the water solenoid (they're identical), triggered by a torch button.....Looking around it there is also an unused terminal for "signal lead" on the faceplate - don't know currently what this is for......
So really I thought I'd try to get the caps in a decent condition and then see if it welds ok....(Will I be disappointed haha...idk only have a need for AC Tig very occasionally...)

 

[AKN]

Be careful here! Those old capacitor may have PCB polychlorinated biphenyls in them, @Parm , these are carcinogenic and were banned years ago.

 

[Erie Fred]

20kv sounds like a capacitor to keep the Hi-Freq voltage from back feeding into the welder.
A feedthru capacitor IIRC.
Similar to on a plasma cutter.

 

[eddie49]

Those four will be oil-filled high voltage capacitors, not electrolytics. They are of very low capacitance and work on high frequency AC, whereas electrolytic capacitors typically have high capacitance and are polarised so only work on DC. I think the four capacitors in parallel are C2 in the circuit diagram, and form a series resonant circuit with the variable resistor and the primary of the coupling transformer. They are this type of capacitor: https://www.ebay.co.uk/itm/127107591085

 

[Dragnet]

What's leaking? If oil from those four RS ones, be careful as AKN says. They could be rebranded Hunt's, and thus have an alphabetic date code on so you could work out when they were manufactured. I'd guess mid-60s. I wonder if a ceramic would be the modern equivalent at 20kV, but eddie49's your man here. The three grey moulded ones (look also like RS types) are probably OK - if they're what I think they are, a mixed dielectric, they can go high in value but with the insulation intact. If leaky, a common replacement for those is the axial yellow polypropylene film types.

 

[eddie49]

What's leaking? If oil from those four RS ones, be careful as AKN says. They could be rebranded Hunt's, and thus have an alphabetic date code on so you could work out when they were manufactured. I'd guess mid-60s. I wonder if a ceramic would be the modern equivalent at 20kV, but eddie49's your man here. The three grey moulded ones (look also like RS types) are probably OK - if they're what I think they are, a mixed dielectric, they can go high in value but with the insulation intact. If leaky, a common replacement for those is the axial yellow polypropylene film types.

Ceramic capacitors could be used as modern replacements, but I would worry about the current requirement of this circuit as well as the high voltage. Those bolted terminals on the original capacitors must be there for a reason! I don't think a small wire-ended ceramic disc cap would last long.
Reading about home-made TIG welders and Tesla coils I have come across so-called "doorknob" ceramic capacitors:

1PCS 102 3KV 202 10KV 1n0m 1000PF high-frequency machine spark filter button high-voltage ceramic dielectric capacitor - AliExpress 502

Smarter Shopping, Better Living! Aliexpress.com

www.aliexpress.com

Alternatively, these large polypropylene film/foil capacitors would probably be OK too:

2PCS high voltage polypropylene film capacitor 0.001UF 102/12000V 12kv 12X60MM - AliExpress

Smarter Shopping, Better Living! Aliexpress.com

www.aliexpress.com

 

[jo1010]

Aha, you've answered the question that was bugging me about the little ceramic caps @eddie49......I'll have a look through those links....since then I've also seen soviet type bobbins that are rated around 20kv.....on closer look today 2 of the original caps look ok, 1 is slightly damp, and 1 is leaking badly....so I'll see what appears (presumably a single 2nF 20kV cap could substitute for two the originals....)

 

[Dragnet]

Yes, a diddy disc type would be a poor substitute! Perhaps a chain of them in series-parallel with balancing resistors would be an option. I know old TV repairers have used them on CRT circuits at several kV, and as long as the leads are kept apart, or you're good at designing PCBs for such high voltages, they've been fine. Nosing around I see LCR offer a film EHT line, but I can't see any prices. I guess that means it's silly to ask...

I wonder whether the originals were simply the largest available, or whether having them in parallel is a design feature - they may be doing so to reduce inductance at high frequencies.

 

[jo1010]

It's certainly an interesting design....the Tx output is 3500v 50mA so I can only guess at the designed 20kV ratings of the caps. Some more views of the gubbins.....

 

[jo1010]

**MISLEADING** circuit diagram above.....I checked back thru the docs I had and the diagram above is taken from a Miller HF box! (HF15/20 to be precise), so there may only be passing similarity to the actual Interlas circuitry **** I'll have a look through the web and see if I can find the real thing ...(I seem to remember that Interlas re-badged Miller HF units, so there may be a passing similarity)

 

[Dragnet]

Well it all looks pleasingly solid - I think it's well worth getting going again!

 

[jo1010]

These are some ceramic high voltage caps on eblag atm - never seen anything like these before (soviet era construction).....but what kind of wire connections could be used....would they crimp onto some kind of bullet connector....idk
It's also difficult to work out from the legends exactly what the spec is - seller says 21kv and 1000pF, but are they

 

[jo1010]

....elsewhere there is similar but on these the round connector is screwed to a flat terminal coming out of the cap so tap the round connector for a screw fixing of the bridging bar....

 

[eddie49]

For Russian capacitor markings:
Φ is Farad
B is Volts
K means ceramic
MBM means metal-paper capacitor (металлобумажный)

"Seller says 21kv and 1000pF, but are they ?" -- I would say yes.

I wonder whether the originals were simply the largest available, or whether having them in parallel is a design feature - they may be doing so to reduce inductance at high frequencies.

Yes, I think the main reason for having banks of similar-value capacitors in parallel is to increase the current-carrying ability by decreasing the effective ESR - Equivalent Series Resistance.

 

[Dragnet]

I would tend to avoid any paper dielectric because they have demonstrable failure modes based on water absorption. 'Waxies' (waxed paper) in valve audio grid coupling locations are a classic 'change on sight' as any leakage here puts DC on the grid and can blow the output valves and transformers (expensive...). Some paper in oil metal cased capacitors are very resistant to this and can be good after 80 years, but others (as you've discovered in your box) leak their oil which can be very nasty and rather demolishes their electric capabilities too. Those Soviet ones are at least 35 years old and seem to have some potting compound as their seal. I'm not convinced that's as good an encapsulation against thermal cycling as a crimped and soldered metal can, so wouldn't not be certain they'd resisted water ingress or be able to withstand their rated voltage any longer. I have had some good Soviet film capacitors, with zero leakage at 630V, but they don't have the same tendency to absorb water so are very resistant to age-related failures. I would prefer ceramic or metallised plastic film for future reliability, but current carrying and high frequency resistance does have to be considered.

 

[jo1010]

@Dragnet yep thanks, good points. I think you're right. I'm looking at @eddie49 's links for the ceramics and film caps. Would these modern caps at 13kv reduce inductance enough......I don't understand enough of these circuits to say atm - I should look at what the Tesla coil builders are doing to compare

 

[Dragnet]

In parallel, yes. You could always experiment. Even with 2kV capacitors you can chain them as long as you put balancing resistors across each one so the weakest one doesn't blow up. The comprehensive answer here might help.