Flashover damage

[Mick Bailey]

A porcelain socket is fired at anything up to 2500 deg. F to vitrify the body. The purpose of glazing the socket surface (which is carried out at a lower temperature) is to prevent surface contamination which could decrease the resistance to breakover from surface contamination. Because the initial firing fuses the materials within the ceramic body, the material is impervious to moisture absorption, hence for many sockets only the upper surface is glazed.

Unless contaminated the porcelain has a high resistance to electrical breakdown. However, arcing in close proximity can reach temperatures in the tens of thousands of degrees. This localised temperature deposits metallic and carbon particles onto the surface of the ceramic, maintaining the discharge. More importantly, the glazed surface melts and metallic oxides are highly soluble in molten glass (in fact many coloured glazes are composed of metallic oxides). The abrupt nature of a breakdown also means that micro-fractures can be created through the material, on or near the surface, and these can in turn be contaminated and form conductive pathways.

Now, if that material is cleaned and put back into service there is still the possibility of surface contamination through;

1. Microscopic carbon and metallic particles deposited on the surface
2. The presence of dissolved metallic oxides within the fringes of any remaining glaze or within any momentarily liquified porcelain.
3. The presence of contaminated micro-fractures.

There's an interesting property of a spark gap; if two energised electrodes are spaced such that they are just beyond the breakdown voltage of that gap, then another (electrically isolated) conductive point interposed between the electrodes will cause a discharge within that gap. Once initiated an arc can still be maintained even if the voltage is reduced well below the breakdown voltage of the gap.

So, why is this relevant? The particles which remain on an arced porcelain socket act as numerous interposer points and reduce the breakdown capacity between the socket pins, making failure more likely in the future. As the breakdown of any gap is more likely (for a given voltage) with an abrupt voltage change rather than a gradual one, the socket is more likely to fail again under stressed conditions than steady state conditions.

So, a cleaned socket may work for a while until the gap is electrically stressed. Now, the question is, would the stress be within normal operating conditions, or would that stress be caused by the fault that caused the breakdown in the first place?

[km6xz]

This is something I learned the hard way in the late 50s when I was 10 years old using surplus transmitter tubes and sockets. They were cheap but I quickly learned to identify which theater of the war the parts were stored or used in when hunting for parts for transmitters using 1500-3000 volts on the anode. ANY contamination, like the deposits acquired in the Pacific tropics acted like the little spark gaps Mick mentions particularly when dealing with RF because slight out of resonance of tuned plate circuits can send the load Z up to 10s of thousands of ohms or down to single digits of ohms. Ceramic is stable and has a high breakdown voltage if very clean but not if contaminated. You learn a lot about layout, stray capacitance and materials building RF systems.
Guitar amps used in clubs or on the road have a lot of dirt, kitchen grease, smoke and moisture exposure but do operate at pretty low voltages and have less risk change of major mismatch. But it does happen. I see more pc board contamination causing carbon trails than ceramic sockets mounted on chassis however. For pc board burns, I remove the portion of damaged board with a Dremel tool and fill it with epoxy since we can't get replacement pc boards here.

[Sowhat]

Looking at those 3 sockets, maybe it's just the perspective, but it seems there's less actual ceramic material between pins 2/3 of the charred socket. On the one on the right, 5/6 seems the same. Does it look even on your end, or is it just my perspective? I wonder if the tube lit up like my friend's 6L6GC when he didn't realize the guide pin was snapped off that one time...

Justin

No, that's just an illusion caused by the burn.

[soundguruman]

A porcelain socket is fired at anything up to 2500 deg. F to vitrify the body. The purpose of glazing the socket surface (which is carried out at a lower temperature) is to prevent surface contamination which could decrease the resistance to breakover from surface contamination. Because the initial firing fuses the materials within the ceramic body, the material is impervious to moisture absorption, hence for many sockets only the upper surface is glazed.

Unless contaminated the porcelain has a high resistance to electrical breakdown. However, arcing in close proximity can reach temperatures in the tens of thousands of degrees. This localised temperature deposits metallic and carbon particles onto the surface of the ceramic, maintaining the discharge. More importantly, the glazed surface melts and metallic oxides are highly soluble in molten glass (in fact many coloured glazes are composed of metallic oxides). The abrupt nature of a breakdown also means that micro-fractures can be created through the material, on or near the surface, and these can in turn be contaminated and form conductive pathways.

Now, if that material is cleaned and put back into service there is still the possibility of surface contamination through;

1. Microscopic carbon and metallic particles deposited on the surface
2. The presence of dissolved metallic oxides within the fringes of any remaining glaze or within any momentarily liquified porcelain.
3. The presence of contaminated micro-fractures.

There's an interesting property of a spark gap; if two energised electrodes are spaced such that they are just beyond the breakdown voltage of that gap, then another (electrically isolated) conductive point interposed between the electrodes will cause a discharge within that gap. Once initiated an arc can still be maintained even if the voltage is reduced well below the breakdown voltage of the gap.

So, why is this relevant? The particles which remain on an arced porcelain socket act as numerous interposer points and reduce the breakdown capacity between the socket pins, making failure more likely in the future. As the breakdown of any gap is more likely (for a given voltage) with an abrupt voltage change rather than a gradual one, the socket is more likely to fail again under stressed conditions than steady state conditions.

So, a cleaned socket may work for a while until the gap is electrically stressed. Now, the question is, would the stress be within normal operating conditions, or would that stress be caused by the fault that caused the breakdown in the first place?

Thank you, Doctor Bloody Bernoffsky. You said it well.

AND if there is a flashover arc across FIBERGLASS:
On a circuit board, You have a similar problem.
The fiberglass must be removed, mostly or entirely, or the flashover continues to occur.

(I hope Mesa Boogie READS this)

[olddawg]

I wonder what would have happened, if after grinding out the affected area and cleaning with non residue solvent, if a dab of good old red TV arc dope would have made it functional? Or even painting the porous area with some epoxy? I agree the best course is replacement. But ..... if it were mine or I was on the road, I would try it.

[soundguruman]

I wonder what would have happened, if after grinding out the affected area and cleaning with non residue solvent, if a dab of good old red TV arc dope would have made it fictional? Or even painting the porous area with som epoxy? I agree the best course is replacement. But if it were mine or I was on the road, I would try it.

Constantly looking for a way to avoid replacing tube sockets...
there is none.

You should paint on the TV arc dope BEFORE the flashover occurs.
Although, it does not seem 100% reliable...

[Alex R]

Eewww!

Heater to plate.

Quite. I believe that the DR rectifies the heater supply to give DC to v1, then uses the supply to run switching relays. Plenty of potential damage there if the heater circuit saw plate voltage. Sorry... I once saw a Single Rectifier head with that problem and the damage went a long way into the switching circuit.

[Jazz P Bass]

I had a Bogner head in a while back.
An Alchemist.
Same senario.
Plate to heater. Heater used as a low voltage switching supply.

What a bitch it is tracing out these pcb's.
They are friggin black.

[soundguruman]

I had a Bogner head in a while back.
An Alchemist.
Same senario.
Plate to heater. Heater used as a low voltage switching supply.

What a bitch it is tracing out these pcb's.
They are friggin black.

Just keep a careful watch on the speaker impedance.

Use a very good speaker cable.

And you won't have this problem anymore.

[olddawg]

I have never had trouble with a 2:1 impedance match on any amp and my speaker cables are 30 year old lamp cord. Lol.

[soundguruman]

I have never had trouble with a 2:1 impedance match on any amp and my speaker cables are 30 year old lamp cord. Lol.

Old Dawg: you invite trouble. Furthermore, you set a bad example.

[tedmich]

and impedance mismatch pushes the PA harder to get more of those nice even ordered harmonics out of the crystalline lattice!



right before it blows it never sounds better...

[g1]

Old Dawg: you invite trouble. Furthermore, you set a bad example.

No one in the history of this forum has set the kind of bad example that you do. You constantly spread BS around and make valid points maybe 5% of your rantings.
Go bias some amps in the dark with your "redplate" method.

[olddawg]

Old Dawg: you invite trouble. Furthermore, you set a bad example.

Yeah, and I drive the living crap out of my amps. Also was involved with the initial marketing of Monster Cable. It's all hype. Most of their specs mean nothing for guitar amp application at audio frequencies. I A/Bed their speaker cables with lamp cord, Romex, even claymore mine cable. As long as they had good Switchcraft plugs attached correctly there was no difference in sound or function.

[Leo_Gnardo]

Old Dawg: you invite trouble. Furthermore, you set a bad example.

Yup, Old Dawg and me too. When I have an arced octal socket I get all dentist-like, break out the Dremel and grind deep into the affected zone, leaving what the medico's call "adequate margins." No corona-dope glyptol greazy kid stuff. "Maybe" a tiny dab of clear GE RTV cement - hell I've used that in power supplies up to 25 kV and had no arcs.

Most of the time I'd say 95-98% I win, no further problems. Especially if flyback quench rectifier is applied to the plate connection.

Dawg will enjoy this: April 1987 Bowie rehearsals, production manager Joe Wirsing grabbed Monster Cable's PIA sales rep Mark Dronge (yes the one from Guild) and physically kicked him out of the building. While the band and crew looked on, laughing our guts out! Claymore cable indeed! You bet - wire up that perimiter Dawg!

You, me, and Warren Z, Old Dawg! Mess'rs Bad Example are we. Hey this photo looks a bit like me today, just more hair on top.