Amp (JCM800ish) popped on startup, then one power tube redplated

[SnickSound]

With some luck it's probably just a failed power tube, but the pop right on startup puzzles me. There shouldn't be any sound until the tubes warm up!

After the initial pop, I looked at the tubes and one of the EL34s had a very pronounced blue glow (more than usual, and way more than the other tube next to it) and then the plate started glowing bright red. I shut it off.

Now, do note that this amp does not use a typical Standby switch (it's just a mute switch that grounds the PI input instead), so it applies full plate voltage on startup. And since it uses JCM800 iron (Hammond 290GX), it probably goes over 500VDC for a short while before tubes start conducting. These EL34s are about 2 years old, but they haven't been used THAT much.

So maybe voltage went too high on startup and damaged the tube and that's why I heard a loud pop.

But before I go ahead and try another pair of tubes and risk ruining it, I want to ensure I rule out anything else (which I need to do without running power tubes). Things I'll be looking at:

- Bias supply (confirm it's providing the right amount of negative voltage to each tube's grid)
- Tube socket damage (looks for any sign of arc damage)
- Visual inspection of the board (NB: this one uses a PCB from Headfirst, it's not a turret board build) and main caps
- Confirm I get audio out of the FX loop

Assuming it's just a failed tube, I'll probably go ahead and install a proper standby switch. Usually not required, but this one idles at 485VDC with the tubes hot, so probably too much voltage for the first 30 seconds or so.

EDIT: Note that the amp also features elevated heaters. So heaters would have been at +50VDC or so right at startup. That said, this amp has been running fine and gigged regularly for the last 3 years

[Latole]

As I learned from sluckey (a member you probably know), a "high" plate voltage is not dangerous for the tubes as long as the screen grill voltage is not too high (350 volts, for example).
And, in my opinion, 500 volts on the plates is not too high.

The first thing to do is to measure the bias voltage on the base of the EL34 tubes without them in place ( -30 vdc ? )
This voltage must be present at all times.

A stanby switch is not mandatory.

I suppose it's a home-made amp?
There could be a lot of mistakes
Do you have the schematic as built?

[SnickSound]

As I learned from sluckey (a member you probably know), a "high" plate voltage is not dangerous for the tubes as long as the screen grill voltage is not too high (350 volts, for example).
And, in my opinion, 500 volts on the plates is not too high.

The first thing to do is to measure the bias voltage on the base of the EL34 tubes without them in place ( -30 vdc ? )
This voltage must be present at all times.

A stanby switch is not mandatory.

I suppose it's a home-made amp?
There could be a lot of mistakes
Do you have the schematic as built?

As mentioned it's built using a PCB from "Headfirst" (now Evolve DIY). But I built this one exactly as per the JCM800 2204 schematic.

Measured voltages are a bit higher than what's on the bit you shared though, most likely due to higher wall voltage.

This was my first time using a PCB, I liked that I could more easily use radial caps (cheaper and better spec'd). But now I hate that I don't really know exactly what's going on under the hood. The loud popping sound at start up is what really puzzles me.

[Latole]

Read bias negative voltage on EL34 pin 5, first. No EL34 tubes in.

It's important to proceed step by step and in the right order

[Latole]

You wrote ;

"Now, do note that this amp does not use a typical Standby switch (it's just a mute switch that grounds the PI input instead) "

"So heaters would have been at +50VDC or so right at startup. "

_____________

That is not what I call a real JCM800 / schematic is .

[SnickSound]

You wrote ;

"Now, do note that this amp does not use a typical Standby switch (it's just a mute switch that grounds the PI input instead) "

"So heaters would have been at +50VDC or so right at startup. "

_____________

That is not what I call a real JCM800 / schematic is .

Lol, true. There are 2 main differences: no Standby and DC elevated heaters :)

[SnickSound]

Well this is getting... confusing.

Opened up the amp, no signs of electric arcing anywhere, no bulging cap.

But I was struggling to confirm the value of the resistors in the DC elevation circuit (it's supposed to be  820k over 100k for a 11%-ish DC elevation, or around 52VDC) but the 820k resistor kept measuring at 120k or so.

So I tried to measure resistance to ground from B+... fluctuates all over the place which I'm assuming is due to the caps charging with the voltage from the multimeter, but it seems to settle around really low values, like around 70k. I removed the B+ fuse and all the tubes, the only path to ground is through that DC elevation circuit (920k combined resistance) and the two 220k resistors across the main filter caps (without the fuse the main ones are disconnected so just the screen supply is accounted for). Altogether I should be reading almost 300k, 5 times higher.

The lowest measured resistance to ground is on the cathode follower node, at 65k.

So something is leaking somewhere (again, this is with the tubes out), or I'm an idiot and I'm missing something.

[AlNewman]

It can be hard to measure resistance in circuit, especially without an as built schematic.
One thing I don't see is voltages without tubes.
Do you have a light bulb limiter?

[SnickSound]

It can be hard to measure resistance in circuit, especially without an as built schematic.
One thing I don't see is voltages without tubes.
Do you have a light bulb limiter?

I never drew up a schematic for this one cause it wasn't built from scratch.

Here are some of the doc provided, the schem doesn't show the power supply but it's pretty boiler plate: doubled diodes into a pair of reservoir caps with balancing resistors, feeds the OT center tap through the fuse then into the choke. Followed by another pair of caps with balancing resistors for the screen node. Then 10k resistor into the preamp.

The layout image only shows the components for the preamp (this PCB is not a paint by number, you have to know a bit already), but we can clearly see what's happening in the power supply. I have 220uF caps for the plate node, and 100uF caps for the screen node. Balancing resistors are all 220k 2W

The DC elevation circuit is composed of R61 (I used 820k), C37 (10uF 250V) and R62 (100k), fed from the screen node. 10K resistors between each preamp power node.

Really nothing original there, except I didn't bother with a standby switch.

Didn't measure voltages yet, wanted to measure what I could without power first. I'll bring it up with the lightbulb limiter when I get the chance.

[AlNewman]

Did you try switching the tubes around and see if the problem follows the tube or stays with the socket?

[Latole]

Resistor cannot be measured without disconnecting one end of the circuit.

It's disappointing that you don't follow the advice already given by those who know more than you by messing around.
 I've stopped following the subject.
I wish you good luck.

[sluckey]

But I was struggling to confirm the value of the resistors in the DC elevation circuit (it's supposed to be  820k over 100k for a 11%-ish DC elevation, or around 52VDC) but the 820k resistor kept measuring at 120k or so.

Disconnect one end of the 820K to measure it accurately. That resistor is unlikely to be causing the pop.

So I tried to measure resistance to ground from B+... fluctuates all over the place which I'm assuming is due to the caps charging with the voltage from the multimeter, but it seems to settle around really low values, like around 70k. I removed the B+ fuse and all the tubes, the only path to ground is through that DC elevation circuit (920k combined resistance) and the two 220k resistors across the main filter caps (without the fuse the main ones are disconnected so just the screen supply is accounted for). Altogether I should be reading almost 300k, 5 times higher.

This sounds like a leaky cap and could possibly be the source of the pop.

[SnickSound]

Sorry if it looks like I'm ignoring suggestions, just haven't had much time to work on it and I'm being extra cautious cause power tubes have gotten expensive! Pretty much all suggestions were already on my to do list.

Disconnect one end of the 820K to measure it accurately. That resistor is unlikely to be causing the pop.

Yeah I don't figure that would be the culprit, just had a brain fart reading the color code and thought for a second it was an 82k (which would have brought my elevation close to 300VDC! but amp's been running fine for over 3 years so can't be that)

That is however what made me test the resistance of the rest of the B+ line (to which that 820k resistor is connected).

This sounds like a leaky cap and could possibly be the source of the pop.

Yeah that's what I'm afraid of. PCB and radial caps make for affordable easy builds... but removing the caps to test them is now a board out operation.

[SnickSound]

So in case anyone's curious: the bias supply failed, without bias voltage both power tubes turned into diodes and started drawing way too much current. One of the 1W bias measuring resistors burned out quickly which lifted that cathode and stopped the current flow in that tube (no visual damage but it now tests open), which is why only one tube was red plating.

I had been waiting for an excuse to rebuild that amp into something else for a while, so I just replaced the PCB with a new turret board, didn't bother trying to fix the original bias supply.

[tdvt]

I DO remember reading this thread earlier & I think these follow-ups are always helpful.

Was there any indication as to why the bias failed?

[SnickSound]

I DO remember reading this thread earlier & I think these follow-ups are always helpful.

Was there any indication as to why the bias failed?

Can't really say for sure. Whatever I could test with the board installed tested fine (and no visible signs of damage anywhere), but no voltage was getting through. Could be a bad cap or trimmer, or a burned trace. Again, that loud pop on turn on is very puzzling and the amp was running fine the last time I had used it. Further investigation would require taking the original board apart, unfortunately I threw it away.

Really I only solved one of the two odd things
- Why was only one tube glowing red? Answer: because the other tube's cathode resistor (1R for bias measuring only) had burned out first
- What was the loud pop? Jury's still out...

[Merlin]

With some luck it's probably just a failed power tube, but the pop right on startup puzzles me. There shouldn't be any sound until the tubes warm up!

A spark in the tube or between the pins/attached components could defnitiely cause a pop before it's warmed up. A spark between anode and heater pins is possible since they are right next to each other

[sluckey]

Merlin, it's good to see you hanging out on the forum again. We refer to your books and website constantly. Even better to sit in the same circle.