Really high peak plate dissipation. Is it safe?

[ECV02]

Hi all,
After assembling a single ended, cathode biased 6BM8 guitar amp, I've stumbled on something fairly peculiar to me. I usually see people cathode bias at close to 100% (usually around 90%) max plate dissipation. The logic I've heard behind this is that every peak is followed by an equally low trough. I measured a -19 volt bias on the tube with a 730 ohm resistor, which equated to roughly 6.6 watts of plate dissipation at my 253v B+ (the 6BM8 is a 7 watt pentode). I figured that was that, and started running the amp HARD. Full volume, overdrive pedals and everything. After awhile I noticed the 6BM8 was red-plating. I rechecked my bias and found nothing wrong, so I kept playing. After it started red-plating a second time, I clipped my meter to the cathode resistor as I played. I measured a whopping 38 volts across the resistor, which calculated out to a staggering 13.2 watts of total dissipation! After subtracting the 1.5 watts of screen dissipation, I was still left with 11.7 watts of plate dissipation.

Is this safe??? I see people bias and run amps like this all the time, without any issues. If I were to hypothetically run the amp like this for several hours, wouldn't the plate eventually melt? That glowing red anode has me nervous already...


Thanks,

E

[PRR]

You should also deduct the power in the load.

But a SE audio amp should NOT change cathode current, much, maybe 20%.

Your rise suggests a very low load impedance (and severe overdrive).

[ECV02]

Severe overdrive is kinda the point... best part of a guitar amp  . When I look at models of most guitar amps, I see power tubes run like this all the time.

As for the load impedance, I'm using a 4.5k-16ohm transformer I salvaged from an old amp. 4k5 is one of the suggested load impedances on the 6BM8 datasheet...

Thanks,
E

[mresistor]

what speaker(s) are you using? and what is the impedance?

[ECV02]

I'm running a 16 ohm Celestion greenback.

[PRR]

> I'm using a 4.5k-16ohm transformer I salvaged from an old amp. 4k5 is one of the suggested load impedances on the 6BM8 datasheet...

.... for specific voltage and current. Looks like 200V 35mA.

For 275V 28mA they want 8K load.

You have 234V plate-cathode and 0.026A cathode current. For a SE pentode, the happy-load tends to be nearly V/I. In this case, 8,990 Ohms. Not 4.5K.

The half-value load will pull double the current at Full Wail. Which is what you observe: 26mA to 52mA.

Add a B+ resistor and re-bias to get close to the 200V 35mA suggestion.

[ECV02]

D'oh. Didn't even think about that! I figured if I just biased accordingly I'd be fine. I've got an old 8k-16 lying around somewhere. Would that be close enough? It seems I've abused this poor tube enough already.

Thanks,
E

[Willabe]

After awhile I noticed the 6BM8 was red-plating. ........ After it started red-plating a second time,.....

Is this safe???

No it's not safe for the tube. It's set up where it's not able to dissipate the heat, so the plate is overheating and turning red.   

....wouldn't the plate eventually melt? That glowing red anode has me nervous already...

It probably will melt the screen grid and short out before the plate melts.

But yes it could melt the plate.

[PRR]

> I figured if I just biased accordingly I'd be fine.

Bias sets the no-signal conditions.

B+ and load determine the FULL-signal conditions.

Most audio averages 10% of Full-signal with a few short peaks.

As you say, in guitar "Severe overdrive is kinda the point...", and you need to consider that extreme also.