Why would a tube socket arc?

[pfrischmann]

I've been working on my homebrew '68 era non master Marshall for a while. I've had two instances where two different tube sockets have arced between pins 2 and 3 on the el-34's. The first time I noticed a little scoring on the v5 socket when I got the amp but didn't figure out what it was until it was too late (an expensive lesson as it took out the OT)

The second time it was (v4) a brand new NOS amphenol socket. It arced so badly, the socket actually felt lighter than the others from all the carbon displacement....Luckily the ot seems fine.

I've just replaced the socket and have talked to a few people but I'm looking for some thoughts.

About the amp....

This was a mkII suberbass. It now has a ptp metro board, new power tube sockets. New caps a MM OT and a MM 10H choke.

***The original PT is in there. It has some outside rust and is pretty covered in varnish on the outside of the laminates ****

The last time this happened.
The amp was biased at 35ma (490 volts on the plates)
I had my OCD pushing the front end of the amp. the amp was dimed and going into a Hotplate....then fzzzzzzt! both times this has happened the amp has been on for a while. I can play regularly for a while, 2+ hours with no problem.

I've talked to a few techs...here's what I've heard

I bet it's parascitic oscillation
I'd bet you have too much current going through the heaters
shrink wrap the heaters
replace the PT, that will solve it
It's more likely an impedance issue. If the tubes are arching it may be a mis match when the amp is under an extreme load.......check to make sure your impedance selector is wired correctly. (it's very easy to do this in correctly)

any thoughts???

[jbltwin1]

You don't specify the one thing that I feel is critical-what kind of tubes are you using. Almost 500v is a lot to suppress with 35ma of idle current. I usderstand that they "supposedly" will take it but dimed into a hotplate is pushing it, I'd try cooling the bias down to around 30mA and installing ceramic tube sockets and try it.

[jbltwin1]

Also, you need to make sure that the HOTTEST tube is only pulling 30ma-not just an average for the two.

[pfrischmann]

These were JJ's....You make a good point. I biased them a while ago. I'm not sure if I counted for the top and bottom range in the bias.

I talked to ED at THD. He mentioned cooling the bias down as well.
More like 25ma. All things being equal (I know there not). That would put the tubes at about 12 watts stattic dissipation.

Seems cold to me....but what do I know

[VaughnC]

It almost sounds like to me that some time in its life the amp was played without or too high of a speaker load causing flyback. For starters, I'd replace the tube sockets with high quality ceramic one's.

[hasserl]

The most common reason for arcing at the tube socket is from high flyback voltage due to incorrect impedance loading.

The tech who told you that is correct, the others are full of crap. With all due respect to Ed, he is not a tech, a great guitar player, but not a tech.

The OT failure you mentioned was not caused by the arcing tube socket, it was killed by the same high flyback voltage. YOu've got a serious impedance mismatch there and until you correct it you are going to have these problems.

[hasserl]

One other thing, once the arcing happens at the socket a trace of carbon is deposited, which is conductive and will continue to give you problems even after you get the impedace mismatch corrected. Ceramic sockets may be better for this, but can still give you problems, leading to another socket replacement as the final cure. But you've got to get that load correct first.

[VacuumVoodoo]

One more thing: recommendation to bias the tubes colder in this case is wrong. With colder bias the tubes draw less current and present a lighter load to the power supply so the B+ voltage will go up. With higher B+ the flyback voltage will also be higher and therefore more prone to arcing. Arcing threshold is a function of voltage between the points and physical distance between them. As said by others, get your load impedances in order, change sockets and bias to normal 70% plate dissipation or a bit higher.

[pfrischmann]

Fly Back voltage makes sense. Is it possible it could be something other than the OT/impedance selector?

Could the Hot plate affect this at extreme volumes?

[hasserl]

Fly Back voltage makes sense. Is it possible it could be something other than the OT/impedance selector?

Could the Hot plate affect this at extreme volumes?

Absolutely, that is exactly where the problem lies. There is a mismatch between the tubes/OT/attenuator/speakers. When you mentioned Ed at THD I suspected you were running a Hot Plate. Is the Hot Plate impedance matched to the amp/speakers? i.e. a 4 ohm amp into a 4 ohm Hot Plate into a 4 ohm speaker cab? Or is one of the items not matched?

[pfrischmann]

Yup,
16 ohms on the amp, 16 ohm HP into a 16 ohm cab...assuming I wired the impedance selector correctly..

It is possible I didn't bias to the hottest tube so one of the tubes could have been 35+ up to 38 MA...I just don't remember

[pfrischmann]

BTW it's a 100 watt amp

[John Phillips]

bias to normal 70% plate dissipation or a bit higher.
I would definitely NOT do that with this amp. 70% is way too hot, and although it might not be the cause of the arcing in this case it's still bad for the tubes. These amps are extremely hard on the tubes (especially the screens) anyway, and biasing them as hot as that just makes it worse, especially into an attenuator - have a look at what happens to them under load when they're biased like that. Biasing cooler does make the plate voltage go up, but so slightly it isn't important - a few volts at most, and it's not the static voltage which is the problem for arcing, it's the induced dynamic voltage... which might actually be higher if the tubes are biased hot since the overall current draw through the OT is slightly more.

I agree with hasserl, but I would also suspect the impedance selector if its the original. Check your speaker cable (the one from the amp to the attenuator) too.

Replacing the sockets with ceramics might stop the sockets arcing, but unless you find the cause of the voltage spikes, it may just do it somewhere else instead... like in the OT.

[hunter]

Certain 8 pin JJs have been manufactured with undersized pins. I don't recall specifics but Jim McShane over on the Weber tube page can provide some insight. May even still be a thread on the subject. Sounded like a bunch were - you might have to measure yours to know for sure. If so, then even a new socket is subject to arcing. If the pins on your tubes are undersized, the solution might be to either not use the JJs or tension the new sockets to grip the smaller tube pins (which could lead to problems later on when you use a standard spec'd pin diameter tube).

hunter

[pfrischmann]

:messedup Certain 8 pin JJs have been manufactured with undersized pins. I don't recall specifics but Jim McShane over on the Weber tube page can provide some insight. May even still be a thread on the subject. Sounded like a bunch were - you might have to measure yours to know for sure. If so, then even a new socket is subject to arcing. If the pins on your tubes are undersized, the solution might be to either not use the JJs or tension the new sockets to grip the smaller tube pins (which could lead to problems later on when you use a standard spec'd pin diameter tube).

hunter

What????

I'm not saying this is the problem but this seems very irresponsible...

Thanks for the insight.

[pfrischmann]

I would definitely NOT do that with this amp. 70% is way too hot, and although it might not be the cause of the arcing in this case it's still bad for the tubes. These amps are extremely hard on the tubes (especially the screens) anyway, and biasing them as hot as that just makes it worse, especially into an attenuator - have a look at what happens to them under load when they're biased like that. Biasing cooler does make the plate voltage go up, but so slightly it isn't important - a few volts at most, and it's not the static voltage which is the problem for arcing, it's the induced dynamic voltage... which might actually be higher if the tubes are biased hot since the overall current draw through the OT is slightly more.

I agree with hasserl, but I would also suspect the impedance selector if its the original. Check your speaker cable (the one from the amp to the attenuator) too.

Replacing the sockets with ceramics might stop the sockets arcing, but unless you find the cause of the voltage spikes, it may just do it somewhere else instead... like in the OT.

O.K.....I think I found my problem, if so. I feel really stupid.

I switched out the impedance selector for one of the new ones. I may have wired it wrong.

Does anyone know if the Mercury Magnetics color codes follow the same color scheme as Marshall. If so, I goofed.

I believe it should be
Green 16
yellow 8
Black 4

If this is the case, I switched the black and green wires. I did a continuity check on the selector..I had the green wire going to the 4 ohm and black to the 16....

Can anyone confirm the MM secondary color codes?

[jca345]

pfrischmann,

Dust is conductive. It is also easily attracted by opposite electrical charge. Dust tends to accumulate on amplifier chassis and in/around sockets. It can and will arch--especially if the humidity is fairly high. Check to see if the amp was/is dusty. Also, a lot of contact cleaners are conductive. Most guys pile too much of these "enhancers" on without thinking about the fact that they leave behind conductive residues. Use this stuff very sparingly. Also, discharge your power filters and re-tension your tube sockets after de-dustifying them.

Most high quality sockets are non-hygroscopic [able to absorb moisture] and tend to have more than sufficient dielectric strength to suppress arching at 500VDC operating levels.

[jca345]

OK, I guess I should have read the entire post before replying to your first post.

[John Phillips]

O.K.....I think I found my problem, if so. I feel really stupid.

I switched out the impedance selector for one of the new ones. I may have wired it wrong.

Does anyone know if the Mercury Magnetics color codes follow the same color scheme as Marshall. If so, I goofed.

I believe it should be
Green 16
yellow 8
Black 4

If this is the case, I switched the black and green wires. I did a continuity check on the selector..I had the green wire going to the 4 ohm and black to the 16....

Can anyone confirm the MM secondary color codes?
Marshall color codes vary a bit, the one I was working on yesterday is yellow 4, green 8, grey 16, ground black.

It's fairly easy to check it - just use a multimeter on the output taps (to the ground connection). The DC resistance of the windings is very small, probably less than the zero error of the meter, but you should get three distinct readings (eg .5 ohm, .6 ohm, .8 ohm). The lowest is the 4-ohm tap and the highest is the 16.

[pfrischmann]

pfrischmann,

Dust is conductive. It is also easily attracted by opposite electrical charge. Dust tends to accumulate on amplifier chassis and in/around sockets. It can and will arch--especially if the humidity is fairly high. Check to see if the amp was/is dusty. Also, a lot of contact cleaners are conductive. Most guys pile too much of these "enhancers" on without thinking about the fact that they leave behind conductive residues. Use this stuff very sparingly. Also, discharge your power filters and re-tension your tube sockets after de-dustifying them.

Most high quality sockets are non-hygroscopic [able to absorb moisture] and tend to have more than sufficient dielectric strength to suppress arching at 500VDC operating levels.

Cool stuff,
thanks Jason. I believe I typically overlook stuff like this.

[hunter]

:messedup

What????

I'm not saying this is the problem but this seems very irresponsible...

Thanks for the insight.

FYI

http://www.webervst.com/tubes/messages/3061.html

hunter

[pfrischmann]

It's fairly easy to check it - just use a multimeter on the output taps (to the ground connection). The DC resistance of the windings is very small, probably less than the zero error of the meter, but you should get three distinct readings (eg .5 ohm, .6 ohm, .8 ohm). The lowest is the 4-ohm tap and the highest is the 16.

I'm afraid my DMM isn't good enough to be sure of the difference. Fluke for my birthday I guess.

[John Phillips]

Another easy way is to measure the output voltage at each tap, with a steady signal going through the amp. You might even get enough if you just turn it up full with nothing plugged in, or if that's not enough to get a good reading, an unconnected cable - the hum is itself a signal. Just meter the AC voltage at each of the three taps. The lowest is the 4 and the highest the 16. (It doesn't matter which the speaker is connected to for a low-level test like that.)

[TheAmpNerd]

About the amp....

This was a mkII suberbass. It now has a ptp metro board, new power tube sockets. New caps a MM OT and a MM 10H choke.
The MM tranny you have should use the same color code as
the original Marshall transformer that came in that amp.
EDIT: or the original amp.

What are the MM tranformer and the choke codes?

Then contact Mercury and verify it with their schematic.

PM me if you need a replacment (read discount) as I'm also a MercuryMagnetic Dealer.

[jbltwin1]

People have been pi$$ing and moaning about JJ's with undersized pins for months. I'm surprised you guys haven't heard about it. It's pretty much a given that SOME of the older stock JJ's have small pins and even the AES buy website states "JJ's, now with larger pins". I think that's more likely than anything, henve my FIRST comments about WHAT tubes were in the amp.

[pfrischmann]

O.K.,
So I verified that I did indeed have the 4 and 16 ohm taps wired backwards.
I'm good now and verified.

BUT... I'm using a bastardized set of tubes as I don't have a coplete factory matched set. I have three winged c's and a svetlana and JJ I swap out. I got a little nervous about the JJ's pins

the tubes a biased like this with 490 volts on the plates

v4 v5 v6 v7
31ma 30.7 28 32 (JJ)
27 (svet)

....So , I turned out the light so I could really look at what's going on. playing through a 16 ohm THD hotplate (it's 5.a.m.) into a 16ohm cab....


v4 and v5 seem to be pulsating to the music, v6 not at all v7 is a bit. When I dime the amp and really start playing, the plates on v4,v5 and v7 start to glow a nice red/orange. The more I play the more they glow. When I stop, it fades.

What is this all about??

[jbltwin1]

Just as a follow up, I sat down just now and measured some pins. They are listed below:

70's Fender branded 6L6 .091
70's sylvania 6ca7 .091
70's amperex el34 .092
2 yr old JJ 6L6 .091
60's national union 6l6ga .093
new GT 6l6 .091
2 yr old JJ el34 .090
60's 6dq6d .092
brand new JJ 6L6 .091


Conclusion? None. they're all pretty damned close and NOT enough to argue about. Maybe a .001 or two hear and there but as a regular service, nothing that I see that a normal retension wouldn't cure. Maybe a little more "mojo" talk, I don't know. Mike.