Separate Cathode Bias Resisters for Dual 6V6 Amp -> Garter Biased 6V6 Songwriter

[bcmatt]

I picked up this amp this week for $10 on Facebook Marketplace.

20241015_190741.jpg

As I suspected from the horrible photo, it is a Garnet made Stencil amp (probably from the 70s). I don't even know what company Garnet made this one for because of the missing name badge. Maybe some company with the initials BJ?
She said it was working but it was clearly missing a 6V6. Haha!

20241015_181619.jpg

The model is the BJ45B and it appears to be this schematic:

Garnet 45B.gif

Here are the innards:

20241015_192439.jpg

I'd like to be able to use some mismatched 6V6s in it that I have kicking around and not have to buy another set of matched tubes at the moment. I'd also like to switch it to Cathode-biased like the schematic says the previous versions had. I have no concern to keep this as a true vintage replica. Even though Garnet is a reputable Canadian brand, these stencil amps have no real value beyond their usefulness. My other motivation for Cathode Biasing this thing is to install the simpler version of VVR that I have parts for.

Anyways, in order to play with some spare 6V6s, I'm thinking I would give each power tube it's own cathode resister and bypass cap.
From the different models of cathode biased 6V6 amps out there, I'm thinking I would start with a 470ohm/5W resister with maybe a 22uF/50V cap on each 6V6 cathode. For some reason, I thought you double the resistance and half the capacitance when going from 2 parallel cathodes to just an individual one, but every model of deluxe or champ that I looked at seemed to keep the same 22uF capacitor whether for one or 2 6V6s... maybe because the difference between 10 and 20uF is pretty subtle sonically?

Looking in my parts bin for about the right voltage, I do have some 22uF/50V or some 47uF/100V.
I'm already going to replace the big cap can with these ones on terminal strips inside:

20241018_112708.jpg

Anyways, any suggestions on these individual cathode bias resister/cap combos?

Also, the one thing I don't have at the moment is a 10ohm/5W resister for the VVR:

VVR in 18watt schem (1).png

But I do have several 20ohm/10W. I assume the VVR should be fine... I think the 20ohm resister would draw more power, but being 10W would make that ok?

[pdf64]

That all looks fine to me

[bcmatt]

I threw this idea out there on facebook and the discussion basically went in the direction of saying that splitting the cathodes to having their own resisters won't make them any more able to use mismatched tubes... unless I also make at least one of the cathode resisters variable to dial in the cathode bias on any tubes that are getting more severely mismatched.

[Helmholtz]

It is a misconception that a shared cathode resistor improves balance.
The opposite is true. It increases imbalance.
The higher idle current of the hotter tube increases the cathode voltage and makes the colder tube run even colder.
Shared cathode resistor requires well matched tubes.

[maxkracht]

Here's a build I did a while ago, still need to finish... It has some links to a circuit to use mismatched power tubes. A bit convoluted but it works. https://ampgarage.com/forum/viewtopic.php?f=6&t=37272

I think separate cathode resistors are good practice if you have the space, but it won't match the tubes.

[martin manning]

Also, the one thing I don't have at the moment is a 10ohm/5W resister for the VVR:

The 10 ohm 5W and the 6V Zener are a current limiting mechanism, so I would stick with 10 ohms. You could double the Zener voltage, but why waste power? Also note that schematic shows a JFET, in fact it is should be an N-Channel MOSFET.

[bcmatt]

Also, the one thing I don't have at the moment is a 10ohm/5W resister for the VVR:

The 10 ohm 5W and the 6V Zener are a current limiting mechanism, so I would stick with 10 ohms. You could double the Zener voltage, but why waste power? Also note that schematic shows a JFET, in fact it is should be an N-Channel MOSFET.

Thanks Martin. I keep using that old VVR schematic for the layout but the Solid State components are not what I use.

[bcmatt]

Here's a build I did a while ago, still need to finish... It has some links to a circuit to use mismatched power tubes. A bit convoluted but it works. https://ampgarage.com/forum/viewtopic.php?f=6&t=37272

I think separate cathode resistors are good practice if you have the space, but it won't match the tubes.

That looks super intriguing.
Do the 270K resisters connect to each other, or are they just crossing without any connection?

IMG_2488.jpg

[pdf64]

... I think separate cathode resistors are good practice if you have the space, but it won't match the tubes.

My experience of a individual cathode bias resistors is that valves that are a poor match tend to idle at reasonably similar cathode currents.

[maxkracht]

Do the 270K resisters connect to each other, or are they just crossing without any connection?

Just crossing. Tubecad drew it better than me.
https://www.tubecad.com/2005/May/blog0046.htm
https://www.tubecad.com/2009/04/blog0163.htm

My experience of a individual cathode bias resistors is that valves that are a poor match tend to idle at reasonably similar cathode currents.

Certainly better than shared, and probably close enough for most cases, but the complicated mess above takes it a step farther.

[bcmatt]

Do the 270K resisters connect to each other, or are they just crossing without any connection?

Just crossing. Tubecad drew it better than me.
https://www.tubecad.com/2005/May/blog0046.htm
https://www.tubecad.com/2009/04/blog0163.htm

My experience of a individual cathode bias resistors is that valves that are a poor match tend to idle at reasonably similar cathode currents.

Certainly better than shared, and probably close enough for most cases, but the complicated mess above takes it a step farther.

Thanks for sharing these, Max.

I am quite intrigued, and am wondering about application in this case. So, for your particular implementation with a pair of 6V6s, how did you land on using 470 Ohm resistors and the 220uF caps? Looking at it, the values are all higher than I would expect. Being in series, that is close to 1K to ground from each cathode compared old Fenders that are closer to half that. Also your value of 220uF seems 10x that of the typical Fender bypass cap used on the old Fender 6V6 amps... So far that's all I'm going on for sources of inspiration. Please guide me to your process on this. Thanks!

[maxkracht]

So, for your particular implementation with a pair of 6V6s, how did you land on using 470 Ohm resistors and the 220uF caps? Looking at it, the values are all higher than I would expect. Being in series, that is close to 1K to ground from each cathode compared old Fenders that are closer to half that. Also your value of 220uF seems 10x that of the typical Fender bypass cap used on the old Fender 6V6 amps...

Your amp wants 200r shared cathode resistor, which would be 400r per cathode. No different in the garter circuit, but you add another 400r in series. Each tube's grid leak resistor connects to the junction of those two 400r resistors on the opposite tube. The tube "sees" that junction as ground, so it's not an 800r cathode resistor, its 400r. Article explains how it works a lot better than I can.

The 220uf wasn't anything special to this circuit, I just like big bypass caps sometimes. I think I wound up with 47 or 100uf or so due to size constraints. Use whatever you would have normally used.

[bcmatt]

Ok, so this is actually representative of my amp currently. It has an extra filter cap like this:

Garnet 45B.jpg

I guess to make it Cathode Biased with individual cathode resisters for each tube, I would change it into this:

Garnet 45B Cathode Biased Split.jpg

And I'm supposing that I could do the garter method like this:

Garnet 45B Cathode Garter.jpg

Do you think it would be important to add grid stoppers or screen resisters of some sort even though they weren't part of the original design?
Would I be missing something else?

[maxkracht]

Looks reasonable to me. Might be running slightly cool with 470r opposed to 400r. Also, keep in mind you loose 20v or whatever on the extra resistor.

I almost always add grid stoppers and and screen resistors. Probably not necessary, but more benefit than risk in my opinion. I'm sure others would disagree. I doubt this amp was cooking tubes without screen resistors.

[bcmatt]

Looks reasonable to me. Might be running slightly cool with 470r opposed to 400r. Also, keep in mind you loose 20v or whatever on the extra resistor.

I almost always add grid stoppers and and screen resistors. Probably not necessary, but more benefit than risk in my opinion. I'm sure others would disagree. I doubt this amp was cooking tubes without screen resistors.

I may try to add some of those extra resistors depending on what I have in my stash. From the article, he mentions that the bypass caps need to handle twice the wattage with this method, but I assume the resisters probably only need half the power handling since there are twice as many.

I'm prepping the chassis for my changes. I'll wire the VVR after but I want to get all my chassis drilling and mounting over with first. I unwired the fixed bias parts and the cap can. I just added a VVR knob and a heatsink. Somehow I had one of these same knobs and I have no idea where from.

20241019_121257.jpg

20241019_121148.jpg