What happened to this tube?!

shoshin

Site Supporter
Site Supporter
This 2A3 ran for a month or so in my previous amp without issue.

I put it in my new amp (which has only slightly high filament voltage - below the 5% tolerance) and could not adjust the hum to settle - measurements were jumping all over the place. Then I noticed the tube was not as warm as the one in the other channel.

Thinking the problem was with the amp, I opened it up and measured:
  • 496VDC plate-gnd
  • 170VDC grid-gnd
  • 4.43VDC cathode-gnd
This is quite different to the baseline measurements (obviously!) I had for this amp. For comparison, putting the tube from the other channel in the amp measures:
  • 455VDC plate-gnd (perfect!)
  • 154VDC grid-gnd (close to perfect!)
  • 204.6VDC cathode-gnd (close to perfect!)
A quick inspection of the problem tube shows that the getter flashing near the base has blackened with the appearance of what might be heat rings:

problem_tube.jpg


So, it seems like the voltage from the cathode has been added to the plate - is this even possible? How would this happen?

Any thoughts most appreciated.

Cheers.

P.S Yes, those are EML mesh plates - not an inexpensive tube to go south.
 
Can you share the schematic of your DC 2A3 amp? If there is not a mechanism in the circuit to prevent full B+ from appearing at the grid of the 2A3 during startup, then damage can occur.
 
There is a CL-80 thermistor between the switch and primary of the B+ transformer; filament transformers are wired directly to the switch (no thermistor). The thermistor is bypassed by a 0.1uF film.

I have not monitored voltages upon switch on. Sound - a slight pop and mild thrum lasting a second or so - is heard from the speaker after ~6 seconds.

I would have thought the filaments would have time to power up before substantial B+ hits the plate, but maybe I am wrong.

Direct coupling seems to involve a level of tweakiness that is not worth the reward... seriously tempted to start a transformer-coupled build.

Cheers.
 
Yeah, a thermistor isn't going to get the job done. If you have a resistively loaded driver stage with the plate load tied to B+ and a really quick warmup rectifier (SS diodes or directly heated rectifier), you can end up with B+ on the grid of the 2A3 until the driver tube starts conducting.

I agree that the fiddly nature of directly heated 2A3s makes them a bit of a PITA.

FWIW, the Sovtek 2A3 will tolerate these conditions pretty well.
 
Starting to think that if going direct-coupling, something like the DRD/monkey style amps with inductive loading of the driver might be the least troublesome... won't work with a pentode driver.

While I think about how to progress, the JJ2A3 can probably handle those conditions too - EML not so much. Sad that an EML meshie had bite the dust for me to learn that lesson.

Thankfully, I had the power transformer wound with secondaries to accommodate RC coupling, so may give that a crack... or just start a different type of build altogether...
 
Wow, lost an EML mesh 2A3? Really sorry to hear this, that is a 2A3 I've always wanted to try but never could justify the 'stiff' price of admission.

I put in a separate switch to allow filament start-up before HV and usually use slow-start indirectly heated rectifiers but certainly forget about that start up sequence sometimes and have used directly heated rectifiers - think I'll stay away from direct-coupling!
 
Yep. It hurts. :(

I would have thought the thermistor would have slowed the B+ ramp-up in the same way an IDH rectifier does... but apparently there is more to it than that. Not sure how Serious Stereo managed it?

The EML solids are on the same level as the meshies, IME. Though, depending on preferences and system, one could have a preference for one over the other. The JJs are nice as well, especially in inherently transparent and dynamic sounding amps.

Cheers.
 
The JJ 2A3-40 will absolutely not handle those conditions.

If you are using a pentode and the plate current isn't too high, you can put a CCS between B+ and plate and the pentode loading resistor between plate and ground. When you turn the amp on, you'll get the CCS current across that plate loading resistor, which may be good enough to prevent damage. If it's not, there are ways to slow down the start-up behavior of a CCS so that the driver plate will sit at ground potential for a short time while everything warms up.

This is a can of worms, but it is ultimately solvable.
 
Thanks Paul. I'm going back to my old amp until I think through a way forward with this... perhaps even changing the design significantly. For clarity, I'll post the schematic in the next day or so. Clearly, I need to better understand what's going on and why this is not an issue with more directly-coupled amps.

Hey, that's part of the fun, no. Please someone, anyone, tell me this is fun! :rolleyes:
 
Last edited:
The easy way out is to switch on the filaments first, especially easy if using separate filament transformers.

Only requires the addition of a switch.

Somewhere in Sound Practices I published a 2 switch arrangement, a variation of upstairs/downstairs switching for home lighting, that would flip on the heaters first regardless of which of two switches was thrown. The other switch would flip on the B+.

I have done that but i also built an number of DC amps using indirectly heated rectifiers and never had any problems. I thought that a GZ34 turned on slow enough to avoid problems, certainly slower than 2A3s turn on. Maybe not much faster than the input tube heats up though and that affects voltage at the 2A3 grid also.

With the price of NOS and EMLs these days, probably wise not to take chances.
 
I have two 2A3 parafeed amps I recently converted to monkey/free lunch circuit.
Been playing for about 2 months with no problems.
Have read this circuit is safer/easier on the 2A3.

They have SS rectifiers.
 
Tieung,

That is a workable solution but you will have to switch the switches in the correct order.

You can also switch a relay in the center tap lead or the positive buss (use a HV relay) using a time delay device. This is a one power switch solution but takes up some space.
 
Missed the discussion here. I’d thought of using a second switch but thought the thermistor would sort it. I now assume the cl-80 does not put enough resistance on the primary to reduce the initial B+ sufficiently.

Due to the construction an additional switch will require new (or hacked) a/c and switch mounting plates... might be worth looking into before I trash the idea of direct coupling.

Thanks.
 
I have two 2A3 parafeed amps I recently converted to monkey/free lunch circuit.
Been playing for about 2 months with no problems.
Have read this circuit is safer/easier on the 2A3.

They have SS rectifiers.

They are safer, but I don’t think they are an option for drivers with plate resistors.
 
The easy way out is to switch on the filaments first, especially easy if using separate filament transformers.

Only requires the addition of a switch.

Somewhere in Sound Practices I published a 2 switch arrangement, a variation of upstairs/downstairs switching for home lighting, that would flip on the heaters first regardless of which of two switches was thrown. The other switch would flip on the B+.

I have done that but i also built an number of DC amps using indirectly heated rectifiers and never had any problems. I thought that a GZ34 turned on slow enough to avoid problems, certainly slower than 2A3s turn on. Maybe not much faster than the input tube heats up though and that affects voltage at the 2A3 grid also.

With the price of NOS and EMLs these days, probably wise not to take chances.

Found it for ya.

Screen Shot 2018-06-02 at 1.46.45 PM.png
 
Sovtek 2A3s...

I seem to be having no issues with the Shuggie 2A3C tubes. Still, I'd like to be able to run a wide variety of tubes. I might be reading into this, but I get the impression this issue might go beyond the filaments simply sufficiently warmed before applying B+?


Thanks! I'll have that buried somewhere in my SP archives... but maybe not so easy to locate. That's neat and avoids the use of timed delays and thermistors. Come home, flick a switch, feed and water the cat, flick second switch. Even I can manage that!

Thanks all.
 
Back
Top