Leaderboard

arc in tube

dbw

Master Member
Messages
4,531
I was playing with my new attenuator (not the topic of this post) which means I was driving my the EL34s in my Carvin V3 flat out.  I was watching the tubes as I play (I like the blue glow  :laughing7:) and suddenly something arced visibly in one of them... YIKES.  So I immediately hit the power switch.  Nothing seemed to be wrong with the amp, but I didn't know what was going on so I pulled the plug.  The plates were red-hot but this is normal, right?

So what happened there?  What do I do now?  Is the tube OK?  Is the amp OK?
 
Depends on the color of the arc....

If it was a blue flash, usually thats just a gassy tube doing its thing, and on occasion they will flutter or flicker, especially when driven hard.

If it was a red, orange, yellow...white flash in there... not good.  You CAN get arcing between the plate and the suppressor grid... not good since the suppressor is typically tied to ground in a pentode.  You can also arc from screen grid to suppressor grid, but this is not likely as the screen grid is intentionally current-limited as part of most amp designs (not all, but I suspect it is on yours, being a modern amplifier).  And, you can arc from control grid to cathode sometimes, but... again not too likely.  Sometimes if you REALLY crank it, the plates will go cherry, warp from heat, and short to screen grid or push the screen grid into the control grid... makes for blown screen grid resistors when the latter happens.
 
The flash was bluish.  The plates look OK.  They were orange, not red.  Just being paranoid I guess :)
 
I got a second opinion from an amp geek who I work with.  He said the glowing plates are a bad thing... they mean I'm biased too hot.  Does that make sense to you guys?

Edit: Just talked to a local amp tech... he said it may be nothing or it might need to be rebiased or there could be a bad connection in there.  He charges $60/hr.  Worth bringing it in?
 
No... build yer own bias probe for a few bucks, and learn to bias it yourself!  cost... under $15, if you already have your own multimeter.

Blue flashes... no problemo... that's the gas swirly thing..... I used to watch that in my old Princeton Reverb all the time.

Plates glowing orange... could be an issue.  It might not be bias.  There are dynamics at play... the limits of THOSE individual tubes themselves, the bias, the plate voltage... all of which have to work together.  I will say, though... <gg> that when the plates glow orange, you usually get some outrageously nice breakup.

As the customer said to the amp tech "it sounded GREAT right before it blew up, can you make it do that again?"

Without knowing some more intimate details, just saying "its the bias" is just a guess (probably a good one, but... never hurts to consider the whole shamoo).
 
I would caution against biasing it yourself.  That procedure is potentially deadly, especially if you are unfamiliar with what you're supposed to be doing.  This is not a simple truss rod or intonation adjustment.
 
=CB= said:
No... build yer own bias probe for a few bucks, and learn to bias it yourself!  cost... under $15, if you already have your own multimeter.

Do you have a good link that explains how to do this?

=CB= said:
Blue flashes... no problemo... that's the gas swirly thing..... I used to watch that in my old Princeton Reverb all the time.

That's good to hear :)

=CB= said:
Plates glowing orange... could be an issue.  It might not be bias.   There are dynamics at play... the limits of THOSE individual tubes themselves, the bias, the plate voltage... all of which have to work together.  I will say, though... <gg> that when the plates glow orange, you usually get some outrageously nice breakup.

It did sound pretty damn good  :guitaristgif:  But I bet I'd fry the tubes if I kept on pushing it like that.
 
Super Turbo Deluxe Custom said:
I would caution against biasing it yourself.  That procedure is potentially deadly, especially if you are unfamiliar with what you're supposed to be doing.  This is not a simple truss rod or intonation adjustment.

Only if you go poking around inside.  Make up some probes... do it the easy way.  Jeeze.
 
dbw said:
=CB= said:
No... build yer own bias probe for a few bucks, and learn to bias it yourself!  cost... under $15, if you already have your own multimeter.
Do you have a good link that explains how to do this?

Its pretty simple really.  Get an octal base from www.tubesandmore.com (Antique Electronics).  Get a socket from them too. The Russian ceramic sockets seems to fit the bases fairly well.  Take some 22gauge bare wire about three inches long and solder it to all the pins on the socket, except pin 8 - IMPORTANT - loop the wire thru the hole in the socket pin (tab really...).  Make two holes in the octal base, close to pin 8.  Pass a test lead wire thru the hole closest to the bottom of the base, and solder it into pin 8 on the base.  Pass another test lead wire thru the hole closest to the socket, and solder that wire to the pin 8 tab on the socket.    Now take those 22 gauge wires and pass them through the holes in the pins of the base, keeping them in order!  (HINT: staggering the lengths slightly makes this go REALLY easy).  Pull all the wires snug as you do this... 22gauge bare wires, and test lead wires should have the minimal "play" inside the unit.  Solder all of the remaining wires to the pins - keeping them tight as you do.  (HINT: work fast on each wire, so you dont undo the soldering on the socket side of the wire).  Let it all cool.  Now put a banana plug that fits your multimeter on each test lead wire.  Many folks like to epoxy the unit at this point (pot it).  You can use just about any consumer grade 2 part epoxy, but I suggest the model aeroplane 30 minute epoxy, as it has great temperature conformity and flows well.  What you want to do, if you pot it, is to mask off all the pins on the socket, and press a hole in the masking tape in the middle of the socket - fill it thru that hole using a little paper funnel you can make yourself.  Once its set - undo the masking and make sure there's no epoxy in the hole in the middle of the socket (if there is, just use a small screwdriver to scrape it out).

Its all done.

To use it, plug it into the amps socket and put the tube into the probe.  Put one lead into the ground of your multimeter (common) - hopefully thats the lead that went to the base, not the socket.  Put the other lead into the amerage (milliamps) input on your multimeter.  Set the meter to read someplace in the 200ma range, max scale.

Turn on the amp and observe the meter, which is now reading all the current going through the tube.  After some warmup, adjust the bias to read about 60 percent of max dissipation (you can figure this out by getting the plate voltage from the schematic).  Barring that, set 6L6's to about 35-40 milliamps at idle - close enough for most practical uses.

Be sure to let it all cool before you take it apart.  Note - extended playing with potting compound (epoxy) in the base is not suggested.  Just get the job done, let it warm up about 5 minutes or so, set it and shut down.

And there ya have it.

If you IM me... I might have access to some under the radar items for ya <gg>, good will, and braggin' rights inclusive of course.
 
I don't have a schematic but the manual says this (I have EL34s):

29. POWER TUBE BIAS SWITCH
If you desire to change from EL34 to 5881(6L6GC) power tubes,
you may do so by selecting the external BIAS switch to the 5881/6L6GC
position on the rear panel. Be sure that this switch is set to the proper
position or excessive heat will damage your tubes. The internal bias
trim control P27can be set by a qualified technician. To set the bias,
measure the current across the terminals of the STAND BY switch
(set this switch to the off position when the amp is on). Set the idle
current to 100 mA for all tube types.
 
Back
Top