Cathode biasing technique

[Voxbox]

Hello,
I've got an old, 1959/60-ish TV-front Vox AC10 in for a repair.
Its going fine, however I'm going to swap out the old, flat-sounding no-name Russian(?) EL84s for a pair of Groove Tube Premium EL84Rs. I've tried JJs and Harma Mullard clones as well, and prefer the sound of the GTs. Such a great, rich overdrive tone :) I dont want to give it back.......


So is there an easier way of biasing the valves other than swapping out the cathode resistor that is connected to both cathodes? Currently its a 130R 8W resistor (with 13v across it) and an 8 watt pot is an expensive thing!
Perhaps a resistor/pot combo (in series) with a lower power pot would do with a more affordable 5W pot? It's just not effective swapping out resistors as I like to hear the effect on the sound of changing the bias as I turn the pot.
Any ideas?


Cheers, Colin

[pdf64]

What issue are you trying to address?
Schematic?

[tubeswell]

No need to change the cathode resistor - just chuck the tubes in and play.

[PRR]

> Currently its a 130R 8W resistor

EIGHT WATTS?? The bias resistor dissipates about 1.69W. OK, doubled that is at least a 3.4W rating, another doubling rounds to 5 or 8 watts.

But you do not have to drop big power in a pot. Fixed resistor(s) with a pot tapping.

I know: wrong tube and only one. The principle is the thing. Pot wiper full-up, tube thinks it is biased by 100 Ohms. Full-down, seems like 200 Ohms. 120, 130, 140 Ohms will be in there somewhere. The user can't turn all the way to zero Ohms. The Bias_Check point is a fixed 100r metering resistor point.

[shooter]

when I "swap out resistors" it takes about 3 try's, no solder, maybe 5-10 minutes.  just have a range on hand, 4 gator clips, 2 for the R, 2 for the meter.  when you're happy, use solder on the winner.

[pdf64]

It seems somewhat crude to mess with the cathode resistor of a proven design. 
What’s the issue, and what’s its root cause?
Formulate and implement a logical plan once those have been identified.
As it is, the cart may be getting put before the horse.

[kagliostro]

In a catode biased amp the only reason to mess with the catode resistor (to me) is if you have tubes that don't match very close each other, in this case, may be, it will be useful to use separated resistor for each tube as to try to obtain similar currents on each tube


Franco

[shooter]

got an old, 1959/60-ish TV-front Vox AC10 in for a repair.

most folk with amps like this tend NOT to want it modified, just working well

[Voxbox]

Thanks all,


Maybe I'm missing the point here. Is it the case with cathode biased amps you dont need to set the bias when changing valves? The valves are a matched pair btw. However I was going by this info from Aikenamps on biasing cathode biased amps, that recommends biasing them.
https://www.aikenamps.com/the-last-word-on-biasing
Any amp I've biased and most of them are not cathode biased, benefitted from setting the bias, within current limits, to get the sound I wanted. i.e tune for maximum smoke then back it off a notch 


I'm not going to mod it for the sake of it, I just want a good-sounding and reliable amp that isn't going to fry the output valves because the current is too high.


Its already good-sounding. Its the first time I've played one of these and I'd have one in a minute, but not at £3-5k that I've seen them go for or Reverb.


So I dont want the customer coming back to me with fried tubes! At 50mA per valve it will be fine.


Take your point Shooter about having a range of resistors, and given that the ear/brain interface doesnt remember the sound of things for very long, I think its something like 19 seconds, I'd just like to do it quicker with a pot. I can get a 5W pot for a few quid so i'll probably just do that.


Thanks y'all, great advice!


Cheers, Colin

[pdf64]

So as per Aiken, you’ve checked the bias; what’s the issue?

[PRR]

Some of the EL84 amps were biased sinfully hot. Too hot? Or not?

Note that Aiken even says "Any amp I've biased and most of them are not cathode biased, benefitted from setting the
bias"

Yes, this is ambiguous. Did he bias the cathode bias amps? Or just the "most of them ...not cathode biased"?

I had an amp like this which clearly wanted to sound best about 150%-200% of rated Pdiss, plates glowing like ripe rubies. Didn't seem wise to leave it that way.

[Voxbox]

Time for some clarity!


Most of the amps I've worked on are fixed bias. Setting the bias has been done with a pot and I can easily hear the tonal variations as I adjust the bias. I usually end up setting it on the hot side as I prefer that sound, and, as it turns out, so do my customers.


I'm an electronics engineer to trade, but am only learning about valve amps in recent years as they stopped teaching about valves when I was at college the year I started my diploma in electrical and electronic engineering. Doh....


There seems to conflicting opinions as ever on the Internet about as to whether you tweak the cathode resistor or not in cathode biased amps. It seems reasonable to me, after reading the text from Aikenamps, that the bias current should be checked at least so that the valves won't get overcooked over its operation of a range of input signal levels. I've yet to do that with the AC10. At quiescent its fine at 50mA per valve.


PRR, I wasnt quoting Aiken, I was rather poorly trying to say that I hadn't  set the bias on any cathode -biased amps. All of them that I have set were fixed bias and definitely benefitted tonally from it being set. So these were usually set on the hot side, without red-plating, to get a tone I liked, which folks also seem to like fortunately.


My plan, pdf64, was to find an easy way to vary the bias in this amp without resorting to swapping the cathode resistors. simply because I cant remember sounds for very long at all, something like 19 seconds, so I'd rather use a pot, as I can instantly hear the difference when I tweak it. 8W variable pots are expensive. The reason to vary it is to ensure I'm getting the best tone out of the amp for my customer without frying the tubes. I've never been satisfied with "that'll do". I like seeing what the range of variation is, certainly out of curiosity, then set it to what I think is good. I'm certainly not going to change electrolytics for example, because one "should". The owner is a pro musician, and he likes it as it is, so I'm not going to mess with it. One of the output valves was microphonic, imparting a weird tail to notes, hence the change of output valves.


PRR, I guess they put that 8W resistor in there for a reason, most manufacturers dont spend money when they dont have to. Its original and as per the schematic. A 130R resistor with 13V across would be served by a 2W resistor, not an 8W one, so why install the more expensive 8W version? Is it required when the amp is turned up? I understand from Aiken that to an extent cathode-biased output stages' bias is self-regulating, however I dont know enough about this to have an answer and was hoping to gain enlightenment here. There's a shed-load of real-world amp and electronics talent in this forum!! 

[dwinstonwood]

For the last two cathode-biased amps I built, I just bought a range of three resistors. I started with the highest resistance, calculated the plate dissipation percentage, and either left it alone if I was over 90%, or went down in resistance until I was over 90%, but not too close to 100%. I also tend to run the amps I've built a little cooler, like 65% with fixed bias. I don't think I could really appreciate or notice a difference in tone between say, 5% or so hotter or colder. I'm more concerned with pampering the tubes, I guess.

[shooter]

best tone out of the amp for my customer without frying the tubes.

Like PRR mentioned, SE amps are biased at max, when you hammer the PA with signal power, the DC tube dissipation drops unlike Fixed bias where all betas are off.
I've only built 1 fixed bias, the rest were self biased.  using a pot and letting a guitarist fiddle may be asking for trouble.
I set the bias R typically between 90 and 120% of max plate dissipation with No signal.
what is more important in self biased is what happens at or beyond clipping (in the PA)  since SE amps are biased by nature class A, driving them into A/B can and IS a tube depended outcome.
that's why I quit using anything but KT88's, they seemed most "musical" when forced into cutoff, the rest was a crap shoot, hair pulling endeavor
I "design" the "sound" in the preamp.  You can feed a nice full square wave into the PA and it will replicate just fine.  when you force a nice clean sinewave into an SE PA that is big enough to overdrive it into Class AB, that's when the hair-pulling starts.
fwiw

[PRR]

Like PRR mentioned, SE amps are biased at max....

I don't think he is working with an SE amp? (Or am I on the wrong channel?)

Cathode resistor bias "CAN" raise current with signal but the transitions can be nasty. In speech/music audio we pick conditions (bias and load) so the rise of current with signal is not much over 20%, and often less.

[HotBluePlates]

... There seems to conflicting opinions as ever on the Internet about as to whether you tweak the cathode resistor or not in cathode biased amps. ...

There are 2 major components to biasing a pentode:  voltage between grid & cathode (here governed by the cathode resistor), and voltage between screen & cathode.

The grid-to-cathode voltage is more-impactful per-volt.  However, getting there by changing the cathode resistor is difficult as the tube seems to "want to run hot":  we raise the cathode resistor, but the dissipation doesn't seem to fall much.

The screen voltage is less-impactful per-volt.  On the other hand, high screen voltage gets ignored as a cause for excessive plate current & dissipation in cathode-biased amps.  When you need to bring down idle current, consider increasing resistance of the power supply dropping resistor feeding the screen node.  Lower screen voltage yields less plate current.

Why go that way?

High screen voltage & a larger cathode resistor means a larger grid-to-cathode voltage was used to tame plate current.  The output tube distorts when the peak drive signal equals/exceeds grid-to-cathode bias volts.  So the larger cathode resistor means the output stage is less-sensitive, and needs a larger drive signal to distort.

Low screen voltage & a smaller cathode resistor means a smaller grid-to-cathode voltage, for the same plate current.  Now the output stage is more sensitive, and distorts with a smaller drive signal.  In some amps/applications, that's Win-Win.

[pdf64]

…A 130R resistor with 13V across would be served by a 2W resistor, …

Just to note that in class AB, current typically increases with signal. The degree of change will depend on the bias point; closer to class A 360° conduction, it won’t be much. Closer to class B 180° conduction, it may be a lot.
So the voltage across the cathode resistor (and hence its dissipation) increases at high signal level.
And crossover distortion will take effect.
Hence technical and audio assessments really need undertaking over a range of signal levels, including overdriven, where the cathode voltage may double.

[shooter]

So is there an easier way of biasing the valves other than swapping out the cathode resistor that is connected to both cathodes? Currently its a 130R 8W resistor (with 13v across it) and an 8 watt pot is an expensive thing!

I assumed self bias from ^

[bmccowan]

I really like Hot Blue Plate's explanation in post #15 that begins "There are 2 major components to biasing a pentode..." When I first started on this hobby/addiction I spent a lot of time chasing my tail, changing cathode resistors. If only most tube amp books were written as clearly.

[dwinstonwood]

I really like Hot Blue Plate's explanation in post #15 that begins "There are 2 major components to biasing a pentode..." When I first started on this hobby/addiction I spent a lot of time chasing my tail, changing cathode resistors. If only most tube amp books were written as clearly.

I agree. Great information! I'm in that starting phase of this "hobby/addiction" now. Just when you think you're beginning to glimpse the whole picture of how these things work, you're shown that you're not seeing another important cause & effect piece of the circuit.

[PRR]

....including overdriven, where the cathode voltage may double.

If cathode voltage (and current) doubles, you will be in severe crossover distortion until the grid-caps charge, and again if the sound is stopped suddenly.

I grant this is very possible in some guitar styles.

[HotBluePlates]

Example of what I was describing.

1st example Shows a Red dot at 150v plate, Blue dot at 100v plate.  Dropping plate voltage by 50v only reduces plate current 5mA.

The 2nd graphic shows a Green dot at 100v plate and -4v on G1, with screen voltage is 300v.  Plate current is ~113mA.  Back on the 1st graphic, the Blue dot is at 100v plate and -4v G1, but screen voltage is now 50v lower at 250v.  Plate current has fallen to 85mA.  A 50v reduction of screen voltage lowered plate current by 28mA, so it was 5.6x more effective at reducing plate current than lowering the plate voltage.

On that 1st graphic, moving from the Blue dot to the Green dot holds plate & screen voltage constant, and reduce G1 voltage by 2v.  Plate current falls from 85mA to 60mA, or 25mA.  So a 2v change on G1 did about as much as a 50v change at G2.

So following my advice means a big change of screen voltage is necessary when that approach is used.  But it's often very effective.