Fuchs Mantis Electric fire!

[pompeiisneaks]

I did a bias adjust on this fuchs that had a nice light show on one of the tubes.  check it out :)

https://www.youtube.com/watch?v=KZuso-i18kY

[HotBluePlates]

In the middle of the plate on that JJ EL34, there is a "window" (rectangular hole).

When you took the amp off standby it started glowing blue in between the plate & cathode, viewable through that window.  Arcing quickly followed.

That particular blue glow (as denoted by location of the glow) indicates a gassy tube, and the arc was likely facilitated by the gas.  So it was totally a tube issue, not an amp issue (and nice that nothing burned along the way).

[pompeiisneaks]

True, but I think the tubes got pretty abused with the plate dissipation at like 120% and climbing :P

~Phil

[HotBluePlates]

True, but I think the tubes got pretty abused with the plate dissipation at like 120% and climbing

The type of "gassy tube" I described causes runaway plate current.  The gas ions counteract the grid bias, and plate current rises.

There's a bit of "chicken & the egg" thing going on here:  Gas ion cause excessive plate current, but excessive plate current overheats tube electrodes, and can liberate more gas.

At the end of the day, it's a tube failure rather than an amp failure, as you found out when the amp goes back to normal just by replacing the bad tube.

[pompeiisneaks]

True, but the bias was still running away with the new tube, but I dialed it down into the sub 70% range and it was fine thereafter, maybe the other tubes weren't in great shape.  I told the owner it may not be a bad idea to replace all tubes

[HotBluePlates]

True, but the bias was still running away with the new tube ...

Define "running away."

I bought an amp recently, and checked bias only as an experiment to see what was happening with another part of the amp.  I used a pair of tubes in it that have sat unused a very long time.  I didn't allow any warm-up or settling-in period, and the idle current of both tubes inched up a few-tenths of a mA every few seconds.  I called the bias "measured"  after a minute or so when the tubes had crept up maybe 3-4mA.

But that was definitely insufficient burn-in & settling time.  And tubes drift anyway.

I once worked on a different amp that had a tube "running away" (to my way of thinking).  With the meter in place to measure idle current, I flipped the amp off Standby.  Idle current shot up within a second or two to 120mA (for the single tube) and I heard loud hum at the end of that through the speaker.  I immediately powered off, tried a different tube.  It didn't have the same plate current spike, just the gentle mild drift of a couple-mA.  I threw the first tube in the trash.

If a tube has good contact in the socket, and the coupling cap to the grid is good, screen voltage is steady and the bias voltage reads steady...  Then it's pretty easy to know the rapidly-climbing plate current is due to some internal tube defect.

Don't forget it's possible to have bad luck and replace a malfunctioning tube with another tube, also malfunctioning.  I know everyone says that is "modern crap QC" (maybe it is), but I've had a bag full of old production tubes that I tried in a new build and had multiple samples I had to trash because of gas.  I found others (some with the same date codes) that worked just fine.

Note there is a huge difference between "blue glow" that is the glass envelope fluorescing (1st pic below) beyond the plate structure due to stray electrons hitting the glass, and "blue glow" (or purple/pink glow; 2nd pic below) that is within the plate structure and indicative of a gassy tube.  The 2nd pic shows a tube about to melt down because gas ions are defeating the effect of grid bias.

The tube in your video showed similar behavior to the 2nd picture.

[pompeiisneaks]

So when I'd replaced the bad tube, the plate current was somewhere around 70mA, which I knew was too hot, so I monitored for a second and it went up to about 86 or so before I quickly dialed it down to the 40's that I wanted, and then it stayed stable.  (I'm going from memory now, and this was a month ago so I could be recalling wrong)

Basically it seemed like if I hadn't dialed down the bias resistance it would have kept climbing and gone to redplating.

~Phil

[HotBluePlates]

That sounds believable.

But that it stayed stable is a sign it wasn't in runaway (unless the bias voltage was sketchy, and wasn't sufficient to reign in the tubes).  Once you get a tube that immediately upon warming up shoots up to 100-120mA and climbing, and you see the plate glow red in front of you, you'll know what I was talking about with runaway.

[pompeiisneaks]

Yeah, I get it, I got there once with my AC100/2 I did a while back (Over a year ago wow!) it also died on me and the original died too, so I'm suspicious of the OT, or the PT doing something wrong, since it's blown up with completely different circuits in it with the same transformers...

~Phil