New Build Idea: 6G4 Based, Cathode Biased

[Rocketrob]

I've got a couple of small chassis left over that have 3 octal and 2 noval sockets so I thought I would see if I could build something different than a 5E3, but still fendery and bluesy. How about a single channel, no trem, early brownface, with cathode bias. I like the idea of two knob tone stack and LTP PI, and cathode bias for less volume and more compression. I drew a schem for it so please fell free to check for mistakes and offer suggestions for making this project work. I'm still a novice so please be patient.

Some questions I have: Is the first filter cap too big for a 5U4GB or 5V4? I don't have much room, so no place for a choke. Is it OK to use a two section cap for the PI node and the PA node? I thought it would be easier to adjust voltages this way. Will I have to change the values in the PI much going from fixed bias to cathode bias?

Please advise and I will post updated drawings as I complete them, as well as a layout drawing for others to use.

Thanks, RD

[Chuck H]

With no voltages posted I can't comment on the PI tail and bias resistor, cathode bias resistor or the feedback resistor. But I would speculate that you might use a 10k PI tail, 27k feedback and the 250R cathode in the schem. Not sure what the 1M series resistor behind the PI is about. Low pass filter I suppose. But because of the high impedance of the PI input there won't be much signal attenuation anyhow so you might avoid some hiss by eliminating that resistor and using a split plate load instead for attenuation. That would also allow you to eliminate the .047 series cap behind the 1M resistor feeding the PI and reduce phase shifts. Reducing phase shifts and sreies resistance are both good things. The gain after the tone stack is already degraded so I might use a 100k pot as the V1b plate load with the coupling cap off the wiper to dial in the amount of gain and then replace it with fixed resistors before buttoning it up.

Looks good.

Chuck

[Rocketrob]

With no voltages posted I can't comment on the PI tail and bias resistor, cathode bias resistor or the feedback resistor. But I would speculate that you might use a 10k PI tail, 27k feedback and the 250R cathode in the schem.

The EDCOR XPWR131 PT is rated at 330-0-330 @ 180 mA, so with a 5U4GB I think B+ would be about 400-410. A 5AR4 would be about 450-460; right around stock 6G4 voltage. I was going to try the lower voltages first, to see how I like it. The Tweed amps use the 10K/470 combination in the PI, that would work too. Just not sure yet.

No

t sure what the 1M series resistor behind the PI is about. Low pass filter I suppose.

I wasn't sure how to join V1B with the PI input; the Fender 6G4 schem shows a .05 cap off the plate , then the 1M R, then the the trem circuit, then the .001 cap into the PI. How do I join V1B and the PI without the Trem? See 6G4 schem attached.

Thanks for the suggestions. RD

[Chuck H]

Do it like this.

[Rocketrob]

Thanks Chuck! My theory is not so good, but I'm learning everyday.

In this voltage divider, if the two resistors are not equal, which resistor sets the gain, and it is setting the gain out of V1b to the PI input, right? I'm just trying to understand how it all ties together. RD.

[Rocketrob]

Ok, after further research I found this article about split load plate resistors and it was quite helpful in understanding this concept. It seems that the resistor connected to the plate is the boss here; that is, if you have a 90k connected to the power rail in series with a 10k connected to the plate, the output of the signal would be reduced by about 10% at the junction of the two resistors. Please correct me if I'm not getting this right...I have enough misconceptions already. Thanks, RD.

Link to article:

http://www.blueguitar.org/new/articl...s/bp010629.pdf

[Rocketrob]

I need some more help on this please.

I am trying figure out how to set the dropping resistors in the power rail. I don't have the PT yet, so no hard numbers, but this can be estimated right? I can't find any examples of amps with LTP and cathode bias 6L6GCs, so I have some questions before I try to do this. The PT is listed at 330-0-330, and with a 5U4GB I think about 410v at the first filter cap.

1. What should the voltage difference be between the plate suppply and screen supply on the power tubes? I have seen any where from 25 to 80v difference. Only a few volts for fixed bias.
2. Should the PI run at stock 6G4 voltage, or something else for this application?

I'm sure I'll have more questions, but I'm trying to work this out. BTW, i am just about done with the first draft layout drawing; will be posting soon.
Thanks for the help, RD.

[Chuck H]

With 6L6's at this voltage you don't really need to worry about reducing the screens other than to be sure it's below the plate voltage (and that doesn't even seem to be so critical when you look at some vintage designs). Just use a 1k 5W and call it done. No good technical reason or advice other than to say the tube will be fine without dropping a bunch of volts for the screen and you won't get a huge voltage drop with a 1k anyhow since the screens don't draw much current. So it's really just a generic, default value to seperate the two power nodes for a dedicated screen voltage filter.

If someone has some tech reason for choosing a specific value I'd also be interested. But really, I don't think it's critical in this app as long as it's not too low or too high.

[Rocketrob]

OK, it just seemed like all the schems (mostly old Fender) that I looked at showed cathode bias 6L6s had a bigger difference between plate and screen supplies. I thought it had to be that way. What about the PI voltage? RD

[Chuck H]

The larger difference with cathode bias could be due to the added sag. With voltage building up on the cathode there is less on the plate. So to keep the screen below the plates it would require a bigger differential to begin with... Hmmm, I never considered it before now and I may be full of crap but that makes some sense. I still think 1k would be fine. There are amps that use cathode bias and a choke between the plate and screen (these are the amps I was referring to above) where the screen voltage actually rides above the plate at many points in actual operation. Some of these amps are known for their characteristic tone. So... I say build it and see. Far less rules with instrument amps. Especially guitar amps since they are expected to create waveform distortions and color the sound of the instrument.

[Rocketrob]

Check my schem and layout, please

I've got all of the parts on hand now, so I should be making some progress on this one. In the schem I added split load plate resistors to V1b and a cap to the standby switch. On the layout I reversed the location of the plate resistors for V2 and the grid resistors for V3 and V4 to shorten the grid wires to the output tubes. Just some stability insurance. The values for the power rail dropping resistors and the plate resistors for V2 are not yet decided; I have to see what voltages I get out of the PT. Take a look and see if I've made any mistakes; my inexperience may be showing again. Any help and suggestions are much appreciated. RD

[Chuck H]

Yupper, looks good. I would include a pair of grid stopper resistors for the power tubes. 1.5k shouldn't mess with the tone but would be good insurance against oscillations. These would be right on the tube pin so no need to change your layout. The split plate load makes for a much more eloquent solution I think. I might plug in 1k for the +V rail resistor between the plates and screens and 5k for the other two. The 1k will probably stay and the other two can be adjusted if need be after you get voltage readings. Then you'll have a place to start and you can build this thing.

When buying parts I would pick up alternative values for a couple of components, in case you want to experiment. Obviously you'll get a couple of different values for the NFB loop and several variables for the PI split load. I would also get a 10k for the PI tail, a .0022 for the PI coupling cap, three .022 film caps as coupling options, a 100/100uf electrolytic for the cathode circuit, two 2.2uf and two 1uf electrolytics for preamp cathode circuit options and two 10k resistors for the preamp +V rail.

That way when you come back on and say 'sounds great but farts out on low notes' (just an example) I can say 'try smaller coupling caps to the power tube grids' or 'try a smaller cathode bypass cap on V1A' and you'll have the parts.

Chuck

[Rocketrob]

Thanks, Chuck.
I'm going to try it without the grid stoppers on the power tubes first, then add them if needed. I have the parts, so not a problem. I also have alot of different value caps and resistors on hand as you suggested; I learned that one already. Nothing like waiting for parts to arrive after you have put it all together. I'm sure there will be a fair amount of tweaking to be done on this one, so i'm prepared. RD.

[MWJB]

With a 330-0-330 PT do not use a 250ohm cathode resistor, start with 330ohm.

[Rocketrob]

With a 330-0-330 PT do not use a 250ohm cathode resistor, start with 330ohm.

Hmm...Is this because of higher voltages than the typical tweed super or other old design with cathode bias 6L6s? I know you need to keep the dissipation at a safe level. Please explain. Also what about the issue of the voltage difference between the plates and screens on the output tubes? I think they should be at least 25v apart. RD.