Bad tube takes bypass cap

[jimiclow]

Hi Paul!
I turned on the SEX this afternoon and noted that the left channel was "scratchy". At first I thought it was just a dirty pin, so I took out the tube and cleaned the pins with Deoxit. I put it back and it was still scratchy. I never thought of swapping the tubes. Instead, I checked for some loose connections under the chassis. I resoldered some joints and turned the unit on for a few minutes. It's still scratchy. This time I noticed that the affected channel had a warm choke. I checked the bias and it was reading 2.75 volts. I turned the unit off and check the resistances. The cathode resistor resistor was reading 60 ohms while the good channel was reading 680 ohms (which is corect). I cut one of cap's lead and it was reading 680. I think that I have a bad cap so I replaced it with the same value. I changed both tubes with another set before powering up. All readings now came out fine. No more noise.
I turned the unit off and put back the "scratchy" tube on one channel. The noise came back. I quickly measured the bias and it read 40+ volts. I turned the unit off and measured the resistances again. The capacitor shorted again. Good thing I have a few capacitors left. I now have the good set of tubes and all readings were okay and sounding good.
So my question is, can a bad tube cause the capacitor to short? If I had a higher voltage rating cap, would it still short?
Thank you.

 

[Paul Joppa]

It appears that the tube must have "run away," that is, it ran at excessive current, which raised the bias to a voltage higher than the bypass cap could handle - thus shorting the cap.

The question is, why did the tube run away. You said it has a scratchy sound, but it was not clear whether any actual music came from the tube. That would be a clue - if there is no music, then it is likely there is a grid-to-cathode short inside the tube. That's fairly rare, and is easily solved with a new tube. However, if the grid bias came disconnected (usually that would be a broken 220 ohm grid stopper, but could be a bad connection ion the grid to negative bias resistor) you might get runaway with music still sort of coming through. If the tube had lost its vacuum, then grid emission could do this for a short while, but you would see the silvery getter deposit turned white.

To answer the question, yes a 63-v capacitor would probably survive better than the stock 35v. I say "probably" because electrolytics run well below their rated voltage will often develop a memory effect, and lose the ability to take their full rated voltage. Conventional opinion is that 50% to 80% is ideal, so 20v bias on a 63v cap (30%) could see this effect.

As I said, an internal short is rare - this is the first time I've heard of it in the 8 or 9 years the SEX has been in production.

 

[jimiclow]

Thank you very much Paul!
I think you diagnosis is correct.
As for the sound, it was intermittent. Sometimes it only came in when the volume is turned up to about 12:00.
Second, on visual inspection, the 249k resistor (grid to negative?) wasn't properly soldered because I could not get resistance reading, probably also the reason why I didn't get bias voltage at one point.
The bypass cap was actually an upgraded Panasonic 135C, only to be destroyed by high voltage.
Everything is fine now. The ebay seller agreed to send a replacement. I have another coin-based tube that is also scratchy to a lesser degree. I might put it one day, when I have a higher voltage rating cap.
If not for the bad tube, I wouldn't have discovered the bad solder on the grid-to-negative resistor.
A learning experience...

 

[jimiclow]

I got the replacement tube and it took my third capacitor. So I started investigating again. If I wiggle the tube socket, it makes a weird noise/thump. I checked the solder on the pins and they all looked okay. To be sure, I resoldered all the joints around the socket. Then I put new Elna 1000/50volts on both channels. No more noise if I wiggle the socket. All readings okay. It's been playing all day. I hope I've fixed the problem.