Twenty-watt, simplified, customized 6g3 build "SBS"

[Roe]

you should change the cathode resistor to 3k when you use split cathode instead of shared cathode.

[xtian]

you should change the cathode resistor to 3k when you use split cathode instead of shared cathode.

Which cathode?

What you see on V1 is not the first stage of the normal channel and the first stage of the bright channel, it's the first stage of the normal channel and the SECOND stage, as seen on the 6g3 schematic. These stages normally have separate, 1.5k cathode resistors.

Please clarify. Thanks!

[rdjones]

This is all guesswork, and the layout is all new, so all comments on my third draft are welcome and then some.

If it's not too late, this being a new customized layout I recommend a "universal" wiring on the output tube sockets.
The main reason to do this is to allow a wider variety of tube types including EL34/6CA7.
I've never been a fan of the Fender method of using unused socket pins as tie points and in particular pin 1.
The Fender methods were done for economic and ease of assembly reasons that are of very limited benefit to a one-off custom build.

The only resistor that needs to be mounted as close as possible to it's socket pin is the control grid stopper.
Mount the grid stoppers between pin 6 and pin 5 and wire the grid feeds to pin 6.

Ground pin 1. This is the key to allowing 7S, 7AC and 8ET tube base interchangeability.
This was standard convention for octal sockets before Fender changed it.
If using 1 Ohm current sensing resistors mount these from pin 8 to pin 1, or alternately directly to ground.
I suggest using ground terminal(s) under the tube mounting nuts and from there run a ground tie to the star point. (don't rely on the chassis as a current path)
Fly the screen resistors or mount on the board or a terminal strip.

rd

[xtian]

Very good advice, Mr. Jones, and I'll take it!

My grid stoppers are 1K5 2W and are connected to pins 5. My screen resistors are 470R 2W and are connected to pins 4. Any comments on these values?

I redrew my layout with a turret strip for the octal socket components. I'm grounding pin 1 as you suggest, and all ground points run to the single star ground on the chassis.

Let me know what you think!

[xtian]

Well, solder fumes wait for no man. Got most of this baby wired, just have to do the input, tone stack, and preamp sockets.

[rdjones]

Very good advice, Mr. Jones, and I'll take it!

My grid stoppers are 1K5 2W and are connected to pins 5. My screen resistors are 470R 2W and are connected to pins 4. Any comments on these values?

I redrew my layout with a turret strip for the octal socket components. I'm grounding pin 1 as you suggest, and all ground points run to the single star ground on the chassis.

Let me know what you think!

I think it's a great idea if you're willing to add the turret strip.
Your ground scheme looks good with the high current point (the cathodes to ground) connecting to the same place as the PT center tap and the filter cap negative.

The grid stoppers are a bit large wattage wise, these can normally be whatever wattage the rest of the signal resistors are - 1/2-1W.
The traditional Fender values are fine.

rd

[xtian]

Finished wiring, and powered up without tubes. DC on filter caps is looking OK. But the bias circuit is confusing me.

I have 51vac on the PT's bias tap. This is steady each time I've measured it. Across the 470R 2W resistor, I get 38vac. Next, across the diode, things get weird. At first, I measured only -9vDC. Way lower than the -34vdc I was hoping for. But during the past 20 minutes I've been trying to figure this out, the voltage has slowly risen to -24 vdc, and continues to rise. All other voltages are holding steady, for example, 474vdc at B+.

I lifted the bias pot's wiper to isolate the bias circuit, so now the entire voltage path is

PT's bias tap (51 vac) -> 470R 2W -> 1N4007 -> leg of bias pot (50uF cap with + to groun here), then other leg of bias pot -> 15K -> ground.

Since I've been typing, bias voltage has risen to -27vdc.

WTH?!?

[xtian]

I disconnected one leg of the 50uF cap, and this allowed the voltage to steady at -22.8vdc. As soon as I reconnect the cap, the voltage starts to rise. Does this sound like a bad cap?

Now that I've got a steady circuit, how do I raise the voltage?

Hoffman <http://www.el34world.com/charts/Biascircuits.htm> says reducing the resistance of the first resistor (470R) will increase the neg voltage. Not working for me! I added in parallel another 470R, and measured 233R, but no change in negative bias voltage. Still -22.8.

[xtian]

I'm reading on half-bridge rectifiers. To find output voltage:

(Vin x 1.414 / pi) - 0.7 (for the diode votage drop)

so,

(51 x 1.414 / 3.14159) - 0.7 = 22.3

and this is what I'm getting. -22.8 volts.

Is this an instance where the bias voltage will change once the tubes are providing a load?

[xtian]

Well, ain't things a LOT different with a functioning cap? I put in a 50uF 150v model, and now I'm getting -60 bias voltage. Guess I'll adjust the 470R "sweep" resistor, and I'll be in business.

[xtian]

OK, I get it. Without the cap, I'm getting half-wave rectified DC, and that duty cycle gives you voltage at the equation shown above. But WITH a reservoir, I get a full duty cycle, raising the voltage accordingly.

[xtian]

Here's where I am. My power section and PI are great. If I ground the input to the PI, I get very low hum and hiss. Sounds perfect. I injected audio from my iPod into the PI, and it's working nicely.

But my preamp sucks. Output is way too loud, and contains too much hum and hash. I've checked my grounds, bypass cap orientation, lead dress. Any suggestions?

[topbrent]

You can very easily dump some gain right at the 220k that leads into the second stage. Just run an additional 220k(or to taste) to ground right there on the board to reduce signal.

If that fails, you could also try wiring V1 up like a 6g2 Princeton.

[xtian]

I replaced the NFB pot with an on-off-on switch so I can have three levels of neg feedback: 1K (roaring), 2.2K, and 56K (nice, bedroom level, low hum and hiss).

Got some claro walnut and created the valence you see below. Is super pretty after a single coat of Tru Oil, so I'm leaving it as is.

Faceplate arrives in a week.

[Chad]

That cab sure looks real nice to me. Great work,hope she sounds wounderfull. Chad