who knows?

[grunt]

BFDR has power tube that sparked like a Roman Candle. Replaced tubes and played some more (amp was dimed) and new tube in same position glowed cherry red. Shut down amp, and pulled chassis. Checked B+, ok at 420V-checked plate voltage, ok at 410 approx. on both power tubes. Checked for bad coupling cap on power tube by pulling tube and measuring negative voltage, -43 to 44V on both power tube control grid teminals with tubes removed. Re-installed power tubes, turned amp on and let sit to get up to operating temperature. After about an hour, tried listening to the pops on the plate and control grid terminals. The pop of the plate terminal on the power tube closest to the rectifier tube was quieter than the plate terminal on the tube that had sparked/glowed. The quieter tube plate voltage was 410V, and the other tube's plate voltage pegged my analog voltmeter set at 1000V. Looks like this may have caused the sparking/glowing. Can someone explain with the limited information I've given what might be going on? Is transformer three history? When I checked with an auto-ranging digital voltmeter, it reads the same as the quieter tube, about 448V. Is it the analog voltmeter or the amp?

[stokes]

Could be losing bias voltage to the tube that is malfunctioning,maybe a bad solder joint on the 220K bias feed resistor, malfunction could be caused by vibrations when playing that is why it seems normal when the tubes are out.Could be the heat causing the resistor to drift too.

[stokes]

Check the screen grid resistor also.

[Steve Conner]

It sounds like a bad bias feed resistor on the tube that red-plates. The tube grid does draw a tiny current, so if the resistor had gone way higher than the 220k it's meant to be, the bias voltage could change with the tube in place and running.

[Chris G]

Check the screen grid resistor also.

Yeh, I've seen a bunch of those fried in Fender amps from the 70s and 80s.

If you can, get 5W metal film replacements. Shouldn't be more a couple bucks apiece, and the higher ratings will make them less likely to fry or drift with heat.

Digikey has 5W metal films from Yageo and Ohmite for next to nothing, like $1.15 each.

[stokes]

Yeah,I would do what Chris says even if the old ones seem good,a little extra protection.If the screen grid resistor is fried I think it is more likely a result of the tube going into a runnaway condition rather than the cause.I think the likely cause is a total loss of bias voltage.If the resistor in question drifted it would cause a change in bias voltage but not likely enough to cause the catostrophic failure you describe.More likely you are losing your bias voltage altogether causing the power tube to runaway like that.The plate voltage going to 1000v is pretty drastic.

[Don Symes]

It could also be as simple as a loose socket at the grid pin. A little retensioning might be a quick fix.

[Enzo]

I like Don's idea. Just because the bias voltage gets to pin 5 doesn't mean pin 5 is gripping the tube socket well. And don't discount the possibility a new tube was bad. SOme old tubes out of something that works will be good testing tubes.

I use 5w screen grid resistors too. I just use the plain old rectangular cement ones. Inexpensive and reliable. I mount them outside the socket circle instead of across it, with enough lead length so they sit on the chassis floor

[grunt]

This is how I checked the 220K resistors. I shut off power after the problem shows itself. I check across the resistors with a dvom. The resistance starts out low (approx 80k) and slowly rises on both resistors to within spec. This is a problem, right?

[stokes]

I would test it for bias voltage with the tubes in and poking the parts in the circuit with a chopstick,after you let it heat up.The check you describe for the coupling caps will tell you nothing about the cap,to check these,pull the power tubes and with the amp on,check the tube side of these caps for dc,if you see more than .2vdc the cap is too leaky.The "drift" you describe when checking the resistance of the 220k resistors could be caused because they are connected to the cap in the bias supply.Check the socket tube pins like Don said,it is rare that the power tubes get that loose,but it is possible.When performing the chopstick test for the bias voltage,use the alligator clips on your DMM,not the probe.If you see the voltage disappear when you poke a component you have pinpointed where you are loosing contact or a bad component.Resistors sometimes develop hairline fractures not visible,but from heat or vibrations can open.

[grunt]

I heated the amp up and did the chopstick test. Seems like it knows I'm trying to find a problem, because it now is behaving. While playing the amp, I noticed with a more discerning ear that the Electro-Harmonics tubes were breaking up at 3-4 on the volume knob. I decided that this was not what I wanted. Tried some 6V6's out of my '65 Princeton (Sylvanias) and these were not well-matched. Biased to the tube with the higher millivolt reading (about 25 millivolt) and played. Much cleaner sound at volume of 4, good grind at 7-8 and up. Much more of what I want.

Can anyone tell me what numbers I look for when shopping for a tube? Do I just take the sellers word that the tube breaks up later? Or what test data will verify what they claim? I know groovetubes have a rating system, but even that has no explanation other than they've "tested" them and this is the rating due to that testing. What is that testing? What makes one tube distort sooner and another distort later?

[stokes]

The BFDR has an adjustable bias,so dont sweat the tube ratings.You can adjust the bias for the tube to break up as you want.I would also suggest finding some NOS tubes,the new productions like the EH arent as good.If you have gone thru everything pointed out in these responses,it is likely that the 6v6's you were using just cant take the volts where you had it biased.If you look at the tube data for a 6v6,you will see the design max is about 350v,which is not a hard and fast rule but some tubes wont take the 410 volts you describe,and be careful when setting the bias,if it is set to high (drawing tooo much current) tube life will deteriorate.Make sure you are not red plating.