B15N cathode resistor/bias (+ thanks!)

[boroman]

At first - I want to THANK YOU all of you that helped me over the last months with old Ampegs. I've been collecting then I learned so much from you, reading and reading and reading...! Few days ago I did the most complicated repair job on my '62 B15N. The amp came to me with bad power transofmer (already non original and dead/shorted), leaked caps, sloppy wiring, non functional standby switch. etc... I have installed new PT which is exactly up to specs with original one, set new caps (along with bias 50uf bias cap and 270ohm resistor), new 1k 10W resistor too. I cut the thermistor (it didn't look right and measured strange...), cut the death cap and made a ground point along with 3-prong cable.

THAT moment when you first hit "Power" and after that, the "Stanbdy" - the heartbeat and you know... The thing wasn't powered since 1970s. The amp actually turned on and its working, super quiet, God, it all went OK. What I don't understand - and of course try to read and learn but it all is too technical - the difference between cathode biased amps and fixed bias amps. In fixed bias ampegs I try to change the bias resistor to make it around 60-70% dissipation. But in cathode biased amp they are all "self-biasing", right?.. You put a set of EL84 in your AC30 and don't even look on the board... that what I remembered.

So I checked the bias in Ampeg and it's near 70mA per tube. Isn't it a bit high? Cathode resistor is 270 ohm, VDC across the resistor 35V and 50uf bias cap and 450 PV. I noted that when I chaged the 6L6GC sets, the bias changed as well +/- 10mA per tube (as with fixed bias amps), so it's not that plug-and-play as I thought (mind the stupid question...). I know that changing the catjhode resistor will result in bias change. So really similar to fixed bias amps?

Is there any "safe" correlation between plate voltage and bias amperes like in fixed bias amps? I have checked my other '61 B15: 250 ohm CR, voltage across it =30V, 100uf cap and 425 plate voltage. Tubes is that particular amp draws like 60mA/pc. Should I worry with 70mA/tube in cathode biased amp in the '62?

[Helmholtz]

With your data (270R, 35V) I calculate a cathode current of 65mA per tube. From this I get a plate dissipation of around 26W. While this is high (87%), someone here (Dave H ?) lately simulated that with cathode bias it is possible to bias close to 100% without overdissipating.

Cathode bias provides some stabilization of idle current against tube variations by local DC NFB. This works best with higher gm tubes like EL84s. Nevertheless idle current with different tubes should vary less than with fixed bias.

If idle current is much higher than expected, it makes sense to check grid coupling caps for leakage.

[Chuck H]

There are a couple things that might clear this up for you. Fixed bias grounds the cathode and applies a negative VDC to the grid, thereby establishing the grid as negative with respect to the cathode. Cathode bias has the grid referenced to 0VDC and the cathode is elevated to give it a positive VDC, thereby establishing the grid as negative with respect to the cathode.

Cathode bias is often called "self bias" because as current increases so does the voltage that the cathode is elevated to. That means that the bias cools with increased current and heats with decreased current. Self biasing,.. Sort of. It's not a magic circuit that accepts all tubes as equal. Certainly not today when wall voltages are high and tubes are all over the map in operating characteristics.

Cathode biased amps in AB class typically idle hotter than fixed bias amps for the simple reason that current will rise as the tube conducts. This raises the voltage on top of the cathode and cools the bias. So an idle is chosen that allows for this and puts the tube in the desired state of bias while conducting.

When establishing actual dissipation for cathode bias you have to subtract the cathode voltage from the plate voltage. Because the tube is "elevated" the plate does not reference 0VDC as an operating point, but the cathode voltage. So it's the plate to cathode voltage and NOT the plate to ground voltage that is the working voltage for the tube.

Also, many methods for taking idle current measurements don't subtract screen current. It's often small enough to be neglected, but not always. And the confusion is cumulative. There are guys out there that firmly believe in the 70% dissipation thing like it's a speed limit. So when they see a cathode biased amp running at 85% or 90% it confuses them for three reasons. They measure plate volts from plate to ground (wrong). They failed to account for increased current with signal conduction (wrong). They failed to subtract screen current (wrong). They might think the amp is too hot when in fact that amp is probably set up perfectly.

Understanding some of this might give you more confidence in your figures.

All that said... Due to modern tube nature and modern wall voltages I think both your amps are running hot. But I do not know how you derived your current measurements. How did you isolate each tube to determine it's current? Are you using a bias socket type tool? Did you plug them in one at a time? Are you measuring across the OT primary?

[boroman]

With your data (270R, 35V) I calculate a cathode current of 65mA per tube. From this I get a plate dissipation of around 26W. While this is high (87%), someone here (Dave H ?) lately simulated that with cathode bias it is possible to bias close to 100% without overdissipating.

This is probably the info I was looking for, thanks!

Cathode bias provides some stabilization of idle current against tube variations by local DC NFB. This works best with higher gm tubes like EL84s. Nevertheless idle current with different tubes should vary less than with fixed bias. If idle current is much higher than expected, it makes sense to check grid coupling caps for leakage.

They all look fine and measure into 10% specs... Do you think putting there like 300R or 350R cathode resistor will bring the current down?

Also, many methods for taking idle current measurements don't subtract screen current. It's often small enough to be neglected, but not always. And the confusion is cumulative. There are guys out there that firmly believe in the 70% dissipation thing like it's a speed limit. So when they see a cathode biased amp running at 85% or 90% it confuses them for three reasons. They measure plate volts from plate to ground (wrong). They failed to account for increased current with signal conduction (wrong). They failed to subtract screen current (wrong). They might think the amp is too hot when in fact that amp is probably set up perfectly.

OK. I'm starting to understand. A "measurement" from the '62:

All that said... Due to modern tube nature and modern wall voltages I think both your amps are running hot. But I do not know how you derived your current measurements. How did you isolate each tube to determine it's current? Are you using a bias socket type tool? Did you plug them in one at a time? Are you measuring across the OT primary?

Yeah... I do it via sort of self-made octal type bias tool with a DMM.

[Helmholtz]

They all look fine and measure into 10% specs

You can't tell cap leakage by looks and cap value. For a test lift one end of each cap and remeasure cathode voltage.

Do you think putting there like 300R or 350R cathode resistor will bring the current down?

Yes. Too low cathode current might introduce some crossover distortion at low output.


Your chart above show wrong (way too high) plate dissipation percentage. Did you insert the correct tube type?

[Chuck H]

The attachment you posted seems to be missing two important details. What is the tube in question? And what is the plate to cathode voltage? I assume these are included in the input screen for the calculations, but the information there means little without those details.

[Chuck H]

And don't ignore Helmholtz observation of possible coupling capacitor leakage. This should be tested for before changing the cathode resistor to affect bias conditions.

[boroman]

I hand-matched and put new coupling caps.

I think it's stable now, take a look:

plate voltage measured from 6L6 pin3 to ground 440V (was 480V), between pin3 and pin8 = 400V
voltage acroos 270 resistor: 33V
voltage across coupling cap(s): 280V
tubes at idle: around 64mA

[Helmholtz]

plate voltage measured from 6L6 pin3 to ground 440V (was 480V)

Your voltage data doesn't look completely consistent. In post #1 you mentioned a plate voltage of 450V (not 480V).
Good coupling caps won't reduce plate voltage.
What is the B+ measured at the OT primary center tap?
When measuring amp voltages it is essential take the readings in short sequence. Otherwise fluctuations in mains voltage may cause inconsistent results.

voltage across 270 resistor: 33V

As this is cathode (pin 8) voltage, the voltage between pin 3 and pin 8 should be 440V -33V = 407V.

Seems like the new coupling caps reduced cathode voltage by 2V, right? If so, the old ones were leaking.

[boroman]

Sorry, I have meant 450V (like in the first post).
Just after powering it on it's now 441, to be clear. If I measure right I have 4V coming from center tap OT (near capacitor) its 448v?
By the way. I see different revisions on different schematics and bias cap changed in some years to 100uf/100V. What's the real difference between 50 and 100uf in this particular place?

[Helmholtz]

Sorry, I have meant 450V (like in the first post).
Just after powering it on it's now 441, to be clear. If I measure right I have 4V coming from center tap OT (near capacitor) its 448v?
By the way. I see different revisions on different schematics and bias cap changed in some years to 100uf/100V. What's the real difference between 50 and 100uf in this particular place?

Sorry, this is all very confusing.

Just after powering it on it's now 441

We can only use stabilized voltages.

If I measure right I have 4V coming from center tap OT (near capacitor) its 448v?

Why do you ask us? You have the amp and a meter.
If the voltage at the OT center tap is 448V and both stabilized plate voltages are 441V, this means a voltage drop of 7V across each side of the OT primary corresponding to a DCR of around 100R. If so, everything is fine.

By the way. I see different revisions on different schematics and bias cap changed in some years to 100uf/100V. What's the real difference between 50 and 100uf in this particular place?

What about posting schematics?
Generally a higher filter cap value can provide better hum filtering.

[boroman]

Sorry about that. I think it's all OK. Will leave the 50uf there as the amp is dead quiet now.
I see some guys adding 2x 1N4007's in series from pin 4 to 8 of the tube recitfier and another from pin 6 to 8. Is it considered a "clever" mod to protect tube from flashing, or protect the transformer?

[g1]

That is a solid state rectifier meant to replace the tube.
For protection, the diodes need to be in series with the tube.

[Helmholtz]

That is a solid state rectifier meant to replace the tube.
For protection, the diodes need to be in series with the tube.

Means that the amp is running at increased B+ (maybe +15V). The tube is shunted by the diodes and has no influence.