Red-plating and More

[Fletcher Munson]

I'm working with a new build based on a blackface Princeton. Been playing it for about a week and it sounds great. Yesterday I decided to try some NOS tubes... RCA black plate 6V6's and one RCA 12AX7 in the V1 position. The amp sounded fabulous.

Today I started playing again and after 30 minutes or so, one of the 6V6's was red-plating. So I figured... bad tube. Put another RCA in there and carefully biased at about 65% max plate dissipation. Started playing again. Got through one hour-long session with no issues. Came back later and played it again. This time, after 20-30 minutes one of the 6V6's red-plated.

So at this point I'm baffled. I decided to re-install my (modern) Tung-Sol 6V6's and check voltages, etc. I put the amp on the bench, plugged in my bias probe with the Tung-Sols, and fired it up. After what seemed like too much time, voltage started rising on the bias probe display and shot up to about 450V. But with the "low B+" tranny I'm using (Mercury FBFPP-M), plate voltage should be closer to 375V. At the same time, there was zero current flowing in the tubes according to the bias probe. So I quickly powered off and immediately smelled something burning. Sure enough the 1k/2W power rail resistor was fried.

So what's happening here? What should I be looking for?

[Chuck H]

voltage started rising on the bias probe display and shot up to about 450V.

I'm confused. Are you saying +450V of bias voltage? On the control grids?

You have told us what the plate voltage "should be around" but you haven't told us what it normally is. And you haven't told us the bias voltage. And if that 450V anomaly is the plate voltage then you haven't told us what happens to the bias voltage when the problem is manifesting. And since the 1k screen grid resistor is toasted, why not tell us the screen grid voltage too.?.

The original bf Princeton (the only circuit reference we have) doesn't have control grid stop resistors or individual screen grid resistors. I'd suggest adding both if you plan to overdrive the amp much. I'm wondering if you may have an ultrasonic oscillation.

[Fletcher Munson]

Sorry I was unclear. The amp has been working correctly for about a week (or so it seemed). Here are the normal voltages:

6V6 plates: 376V
Bias supply: -33V
6V6 screens: 374V

The 450V anomaly happened very quickly. There was no time to measure the bias voltage. I saw the plate voltage shoot up, switched the amp off, then smelled burning resistor. However, I just now pulled the tubes and measured the bias voltage at -36V.

The amp, like the original, doesn't have control grid stop resistors or individual screen grid resistors.

The amp is modified. The vibrato circuit has been removed and replaced with a fixed bias supply circuit. A standby switch was added.

Here's a schematic of the power section with measured voltages (from a couple days ago, before any symptoms appeared):

Power Section.pdf

[The Dude]

I agree with Chuck on adding the screen resistors. I suspect your bias voltage is fine. The tubes stopped conducting because you lost your screen supply when the resistor burnt. Some possibilities are, bad tube, lack of screen resistors, intermittent/leaky filter cap on screen node.

Edit: And,........yes. You did measure the screen voltages. It's right on your schematic where it says 374V.

[Fletcher Munson]

I agree with Chuck on adding the screen resistors. I suspect your bias voltage is fine. The tubes stopped conducting because you lost your screen supply when the resistor burnt. Some possibilities are, bad tube, lack of screen resistors, intermittent/leaky filter cap on screen node.

Edit: And,........yes. You did measure the screen voltages. It's right on your schematic where it says 374V.

Ha! Sure enough.

Any suggestions on the values for those resistors? Maybe 1.5k on the control grids and 470R on the screens like most other blackface amps?

[Chuck H]

Ha! Sure enough.

Any suggestions on the values for those resistors? Maybe 1.5k on the control grids and 470R on the screens like most other blackface amps?

Good starting choices. Right at the pins too. Also, and again if you will be overdriving it much, increase the rating on that screen node resistor to three watts.

[DrGonz78]

Wouldn't it be a good idea to bump that 1k/2w resistor up to at least a 3 watt? In addition to adding those individual screen grid resistors, it just seems that the first power rail resistor will last longer over time at 3-5 watts.

[Chuck H]

Wouldn't it be a good idea to bump that 1k/2w resistor up to at least a 3 watt? In addition to adding those individual screen grid resistors, it just seems that the first power rail resistor will last longer over time at 3-5 watts.

That's actually the "that resistor" I was talking about. Good call. Due to high frequency additions to the signal when the amp is clipping hard and the impedance through the screen circuit supply rivaling the plate circuit impedance at those frequencies I've had to use a five watt resistor there on many amps to keep it from burning up. Sometimes a choke can mitigate this because it has inductance and the impedance at those higher frequencies is greater than a plain resistor. And sometimes when it's just a resistor there I find it necessary to increase the value to stop the screens from over dissipating under hard clipping. Those golden era designers wouldn't have given too much thought to the idea that someday players would be cranking these amps for hours on end and even plugging boosters into the front end to crank them more

[Richardvet]

Hi, not long joined this forum. I would also remove or disable the standby switch as it serves no useful purpose on “receiver” style amplifiers.

[Helmholtz]

When the screen node dropper opens, you lose your screen supply.
Consequently power tube cathode currents will drop to almost zero and B+ will rise to its no-load value of maybe 450V (depending on PT regulation).
So your readings seem to make sense.

But a bad screen dropper can't cause redplating. Redplating means excessive plate current, which is typcally caused by a loss (or drop) of grid bias.
So you should look for bad connections or components in the bias supply circuit.
Also check for bad/intermittent power tube socket contacts.
Seems your root problem requires some time and playing to build up enough temperature. But could also be triggered by vibration.

Loss of grid bias would also strongly increase screen currents which in turn may burn the screen dropper resistor.

Bias supply troubleshooting should be done without power tubes inserted. If necessary use a hair dryer to (carefully) heat up components and joints.