Cathode bias in 5G9

[tubeswell]

A quick question while I am messing around setting this up.

I am installing a switchable cathode bias/fixed bias in teh 5G9 I built recently.

I am using a 5W common cathode resistor bypassed by a 22uF 63V electrolytic cap.

I have a range of values of resistors I can put in. Does 270R sound about right for a common cathode resistor for a pair of 6V6GTs with plate voltages around 370VDC (Before I switch it on an blow everything sky high)?

Cheers

[tubeswell]

Okay I fired it up with the 270R common cathode resistor and this is what I got at idle

cathode voltage = 22.7

Plate voltage = 355

Difference = 332.3

Cathode current = 22.7/270R = .084mA (Do I divide that by two for push- pull? if so that's = .042mA per tube)

.042mA * 332.3V = 13.95W

Is this okay for a 6V6GT push pull amp?

[Paul P]

.042mA * 332.3V = 13.95W
Is this okay for a 6V6GT push pull amp?

Isn't that a bit high for a push-pull amp ? I thought you were supposed to
aim for 70% of max (14w, so about 10w idle) or is that only for fixed bias ?
In my single-ended amp I've had to go to 680 ohms per tube whereas in a
Fender Champ it's 470 so something's changed since Fender made them.

I'd expect you to have to go at least to 300 ohms for a shared cathode
resistor, as I believe is fairly common.

Paul P

[tubeswell]

Isn't that a bit high for a push-pull amp ? I thought you were supposed to
aim for 70% of max (14w, so about 10w idle) or is that only for fixed bias ?
In my single-ended amp I've had to go to 680 ohms per tube whereas in a
Fender Champ it's 470 so something's changed since Fender made them.

I'd expect you to have to go at least to 300 ohms for a shared cathode
resistor, as I believe is fairly common.

Paul P

Yeah that's what i was wondering

[tubeswell]

...only, I read an old post by Bruce of Mission Amps that was saved onto the Blue guitar website, where Bruce recommends cathode biasing 6V6GTs in push-pull configuration with a common 270R cathode resistor, as long as the plate voltage is lowered.

So here the voltage is lowered to 355 (is 372 under fixed bias), and the cathode resistor is as Bruce suggested, giving 13.9 watt dissipation, but this seems to fly in the face of the 70% 'conventional wisdom' that is on the Aiken Amps site.

Is the cathode biasing in this way going to result in Class A operation, or A/B? and if so am I better going to 330R (the next size up), which will raise the plate voltage a bit again presumably?

Is it 'safe' to run push-pull at 13.9W per tube (or is that only for S/E amps)?

[Bruce / Mission Amps]

...only, I read an old post by Bruce of Mission Amps that was saved onto the Blue guitar website, where Bruce recommends cathode biasing 6V6GTs in push-pull configuration with a common 270R cathode resistor, as long as the plate voltage is lowered.

So here the voltage is lowered to 355 (is 372 under fixed bias), and the cathode resistor is as Bruce suggested, giving 13.9 watt dissipation, but this seems to fly in the face of the 70% 'conventional wisdom' that is on the Aiken Amps site.

Is the cathode biasing in this way going to result in Class A operation, or A/B? and if so am I better going to 330R (the next size up), which will raise the plate voltage a bit again presumably?

Is it 'safe' to run push-pull at 13.9W per tube (or is that only for S/E amps)?

Do you know what the screen voltage is?
Your power amp at 14 watts is a little high but with proper ventilation, it probably will be OK anyhow. They won't last as long so I'd still try to get the idle current down a little.
And you are also measuring screen current with that method.
Regardless, I would not go over 330 to 360 ohms in any case though.
If that doesn't do it... the tubes are soft and you'll need stiffer ones or lower B+.
Why?
Keep in mind that not all power tubes will draw the same idle current with the same bias and B+... you might have a pair that are soft and are naturally drawing more idle current anyhow.
As an example, when I order my 6V6s for the 5E3 kits, I always spec them at specific current rating range, based on what I know about the way they are tested and matched from my vendor.
That way I alway know that any pair will pop right in with very little consequential variation of idle current.

First thing I would do is try a pair of naturally cooler running power tubes and lower the screen voltage a little more. Then you can decide if 270 is too low.
330 ohms might be OK but 360 ohms is my absolute upper limit for cathode biased 6V6s under any situation.
If the idle current is still too high, then you have to replace the power tubes with stiffer ones and or reduce the B+ instead of increasing the cathode voltage.

Oh, ... just because the power tubes are idling in high current/power etc... or in the same idle current region you would expect a class A amp to be, doesn't mean they are operating in class A.
They very well still can be in class AB but with very high idle current.
Without know the design of the amp, the OT impedance etc etc... it is rather hard to say if yours is a true Class A amp or not.
I'd error in favor of it being a Class AB amp with high idle current.

[tubeswell]

Hi Bruce

FWIW

Haven't measured the screen idle voltage in cathode bias, but it was around 1V-2V lower than the plate in fixed bias mode at 22mA. The bias is now wired up switchable, either fixed bias or cathode bias. It has a small (Fender) choke for the screen supply with 470R screen grid resistors. I guess (with the plate at 355VDC) the screen could be 350VDC?

The OT primary is wound to give a 5k load resistance, with 3 secondaries (4R,8R,16R).

Everything else is out of the stock 5G9 schematic, i.e. parallel input stages, LTP, CF trem circuit, 2 x 6V6.

[MWJB]

The thing is you haven't actually measured the plate current. You have measured voltage accross the cathode resistor (& assumed the Rk value is what it said on the packet) and are now estimating screen current.

You're using a "rule of thumb" method to ascertain an informed estimate...so whilst we're going down that road you may as well assume you are including 1W of screen current per tube, which puts your estimated plate dissipation in the late 30's (38.85mA if you want an arbitrary number) - perfect.

I'd be very surprised if your 270R wasn't adequate. Another value may be more to your preference sonically, but things should be fiunctioning fine with 270.

I'd also be very surprised that if you actually measured plate current that it would be exactly the same as the figure you got from cathode voltage divided by resistor value. The good news is, if it sounds fine...it is.

[tubeswell]

Thanks MWJB

Is the only way to measure screen current or plate current by inserting an ampmeter in series with the screen or plate respectively, or can it be done by measuring the voltage drop across a screen grid resistor/OT primary respectively? - Never Mind I've just been reading Aiken Amps site on this subject. I see that using the cathode resistor method of measuring, that the screen current is also included in the total current at the cathode.

OK I will do some measurements.

[Bruce / Mission Amps]

Not a big deal but a 5K primary is a bit low for a pair of 6V6s used like this.
Depending on the OT, in cathode bias you'll probably find that the screen voltage is higher then the plate voltage... which will make the tubes draw even more idle current.

[MWJB]

"Is the only way to measure screen current or plate current by inserting an ampmeter in series with the screen or plate respectively, or can it be done by measuring the voltage drop across a screen grid resistor/OT primary respectively?" If you're going to be building amps then you'll save yourself hassle by getting some bias probes for measuring plate current. There is also the "transformer shunt" method of measuring plate current, google it and take proper heed of the safety precautions, use meter leads with clips/wire grabbers and make all connections/disconnections with the amp unplugged from the wall.

[tubeswell]

Not a big deal but a 5K primary is a bit low for a pair of 6V6s used like this.
Depending on the OT, in cathode bias you'll probably find that the screen voltage is higher then the plate voltage... which will make the tubes draw even more idle current.

Thanks Bruce

I suppose I can run a higher R (say 16R) speaker into a lower R (say 8R) socket to bring the load resistance up a bit? (But would that make the load resistance 10k? Too high?)

Is the higher screen voltage a drawback (apart from making it harder to determine where the bias point should be)? I suppose you aren't getting as much power out of the tubes? Could that be why it sounded better running at a higher bias? Is there any danger of overcooking the screen(s)?

I suppose the tubes would shut off if the screen voltage was too high. prompting me to wonder whether a marginbally higher screen voltage makes the tube less reliable

I noticed on one of the tubes when under fixed bias, that the screen measured higher by a couple of volts

Also I can see from the line of thought I have been following with MWJB that the 1R cathode resistor method of measuring bias in fixed bias mode has possibly been selling me short.

I was curious as to how 'flat out' a push pull 6V6 output section should be biased at. I see now that the 42mA calculation I arrived at for cathode current didn't take the screen current into account, but even if I did that, I am left wondering what a cathode biased push-pull 6V6 should be biased at? I am still slightly confused by the nuances of fixed bias versus cathode bias in this regard.

[MWJB]

Many fenders run the screens pretty well at plate voltages (this can vary with state of bias), it's common for SF Champs & Princetons to run the screens slightly higher than the plates...most people are unaware that they have a "problem" as the amps still sound great and are not unreliable.

"I suppose the tubes would shut off if the screen voltage was too high. prompting me to wonder whether a marginbally higher screen voltage makes the tube less reliable", In your case it wont make the tubes unreliable.

"Also I can see from the line of thought I have been following with MWJB that the 1R cathode resistor method of measuring bias in fixed bias mode has possibly been selling me short." How so? The 1R resistor method is a widely used & accepted method of measuring bias. Safety & ease of use (external test points) outweigh any negatives. Good enough for countless top shelf amp builders! Yes, there is the potential to read fractionally higher but this is where you use your ears & common sense...bias to 70% but find that 75% is what you like the sound of? Then run at 75%. As long as you are always measuring the current via the SAME method, you can build a picture of what's desirable/feasible. The number is a guide, you built the amp for the SOUND it makes - this is the priority (as long as you're not eating tubes/can't finish a set due to heat fade).

Many vintage cathode biased 6V6 amps run their tubes at over 100% plate dissipation, mostly due to higher wall voltages. How do you find out what they'll take? You throw more current at them until something gives. OR, you set to 70-90% and enjoy playing through the amp and leave the destruction testing to someone else. At the voltages you are running 40mA (of measured plate current) might be seen as a practical limit

Any 2x6v6 cathode biased Fender, running the voltages that your amp runs, would have been fitted with a 250ohm cathode resistor. I'd really be loking to go higher on the 270 if you were getting much over 380-390vdc to ground.

Cathode bias - high idle current, when hit with a signal the current can't rise appreciably, so voltage drops compressing the front of the note. As the note decays the power supply recovers, swelling the note, helping it bloom.

Fixed bias - Lower idle current, harder dynamics, less envelope of sag/swell, better transient response. The idle current rises considerably when playing (connect your voltmeter to a tube cathode 1R and watch the reading as you play...in fact don't - most folks scare themselves sh#tless when they try this!), you want to allow for that.

At extremes of bias the differences become less marked.

There is a tendancy with novices to obsess over specific numbers (most of which only exist in obsolete 50yr old schematics, the amps in the real world typically don't conform to these), as you become more experienced you will learn what are the critical boundaries and what you can push a little.

[Bruce / Mission Amps]

More rambling...

"I suppose I can run a higher R (say 16R) speaker into a lower R (say 8R) socket to bring the load resistance up a bit? (But would that make the load resistance 10k? Too high?"

Normally you would see between 6500 and 8000 ohms with a pair of 6V6s.
Yours is OK at 5K but 10K is even higher then I would go.

"Is the higher screen voltage a drawback (apart from making it harder to determine where the bias point should be)?"

As long as the screens are not way over the plate voltage, no.
The PA will make more power with the screen voltage higher but keep them closer to the plate voltage.

"I suppose you aren't getting as much power out of the tubes?"

You should be able to run the power tubes in class AB right up to their DC power rating and still get close to the same output power... ...IF, the OT is large enough and has low DC resistance to handle all the extra DC with out dropping B+ voltage too far, has
the proper impedance and the power supply can handle the high current load.
Of course it will still make more power if the B+ is higher and idling cooler Class AB

"Could that be why it sounded better running at a higher bias?"

Maybe but it probably just sounds better to you when the power tubes are idling higher... at the cost of shorter tube life and the amp will be able to be driven into power tube distortion easier but, with everything else being equal, (not including fixed vs cathode biasing) the ultimate power will not be much different.


"Is there any danger of overcooking the screen(s)?"

Not much danger if the screen resistors are at least 1watt and no more then 1000 ohms.

"I suppose the tubes would shut off if the screen voltage was too high. prompting me to wonder whether a marginbally higher screen voltage makes the tube less reliable..."

Well if not crazy high, not really but the tube will use more power supply current and there will be more voltage drop across the OT. If the screen voltage is too low the tube will shut off

"I noticed on one of the tubes when under fixed bias, that the screen measured higher by a couple of volts"

There really should not be much of a difference in screen voltage if the tubes are closely matched. Yes, each tube is different but there should be very little difference in screen voltage on each tube.

"Also I can see from the line of thought I have been following with MWJB that the 1R cathode resistor method of measuring bias in fixed bias mode has possibly been selling me short."

If you have separate screen resistors on the sockets, you can measure the voltage drop across those and use Ohm's Law to find the actual screen current used at idle... I'd be surprised to hear it is more then a couple ma each.

"I was curious as to how 'flat out' a push pull 6V6 output section should be biased at. I see now that the 42mA calculation I arrived at for cathode current didn't take the screen current into account, but even if I did that, I am left wondering what a cathode biased push-pull 6V6 should be biased at? I am still slightly confused by the nuances of fixed bias versus cathode bias in this regard."

Cathode biased PAs always sound better to me with the tubes running hotter, like this:

(idle current times actual measured plate voltage) being at least, 75% to 90% of the power tube's DC rating.

6V6s, being 12 to 14 watt each, would be adjusted accordingly to taste...
9 watts to 11 watts or, if you buy the 14 watt rating, 10 watts to 13 watts, so use that as a guide.
That means with lower plate voltages and an OT of around 6500 to 8000 ohms.

Fixed bias 6V6s always sound and feel right to me at around 60% to 70% ... or around 9 to maybe 10 watts each.
So you can see that there is a range of around 9 or 10 watts that both sound really good but the cathode biased amp will always sound less power full using the same OT and PT because the actual B+ on the power tube's plates will be lower and the variable current through the tube and biasing resistor will alter the bias level as the tube is driven harder and harder while the fixed bias is, well FIXED... exactly the reason to use it.

These variations are not subtle but they are not Earth shattering either.
Remember that the majority of the current in the power tube is delivered through the OT... any voltage drop (while the currents are swinging all over the place in the OT) will also appear at the power tube's plates.
But the screen voltage is not subjected to those instantaneous voltage drops and while the voltage is dropping at the plates, the screen node filter cap will help hold the screen voltage higher during those peak current nodes... or at least more regulated for bit.
Hence, the screen voltage could be higher anyhow at any given moment in AC power output.

In fixed bias PAs the power tubes will momentarily experience an over biased (colder) condition, (even while the main filter caps are acting like a high voltage battery)... the power supply rail still sags a bit but the voltage at the plate pins drops a lot under high power, AC peak outputs.
The cathode biased ones do that too but they seem to be out of phase with what is happening and the bias momentarily changes with the high current flow through the cathode biasing resistor.
That is why a cathode biased PA sounds stiffer if you use a VERY large value cathode bypass cap... it charges up like a battery and supplies a "fixed" bias voltage through the cathode for a longer time constant.

That was all pretty loose and shot from the hip but maybe you can sense the dynamics of it all...
Most players can hear all this nonsense.... ha ha

[tubeswell]

Thanks Bruce and MWJB. This is quite alot clearer for me now.