Does this project make sense? (I think it does)

[dwinstonwood]

Hi all, I'm brainstorming for a new amp project. I want a change from building things in Fender BF-style chassis. I've been looking at sluckey's Dual amps, and I like his board layout style, along with using a Hoffman Stout chassis (my Musing 40 was built that way).

So... since I started with this amp hobby I've wanted to build a 6G6-B. Would it be ridiculous to build one modelled on sluckey's Dual amp concept? It would be the Bass and Normal channels with a three-way switch for either, or both channels (each channel has an even number of triodes, so that would work, right?). I'd only need to drill one hole for the Presence pot, and use one tube socket adapter plate for the PI tube. At any rate, it would be a compact amp. I would probably go with 6V6's since I already have a PT for them.

Thanks for any and all comments and criticism. 

[sluckey]

Sounds like a nice spring project. However, you can't jumper the channels because the Bass channel has three inverting stages, and the Normal channel has two inverting stages.

[tubeswell]

In addition to what sluckey said, getting rid of oscillation is trickier in a 6G6B because of the way B+ is supplied to the pre-amp channels, and because the signals between Ch1 output and the Ch2 output are 'DC-coupled' to each other. I built a 6G6B about 4 years ago and found I had to add an extra 0.022uF coupling cap before each 470k mixing resistor to get the problem to go away. Your experience might be different.

[dwinstonwood]

Thanks tubeswell and sluckey!

sluckey, I missed that inversion issue. I guess I thought that because the Bass channel has four triodes, and the Normal has two that they would end up the same where they join. 

tubeswell, I didn't even notice that. Yes, now I see that coupling caps should be added before each 470K mixing resistor.
Also, it is a bit weird how the B+ string is connected. The Bass channel is fed off of "B" and the Normal channel off of "C." But, the Bass channel actually has lower plate voltages than the Normal channel, even though "C" is stepped down with a 27K dropping resistor! Seems backwards.
I would make four nodes: A for OPT > choke > B for screens > C for the PI > and add a D for both preamps.
Thank for the input! here's the schematic:
https://el34world.com/charts/Schematics/files/Fender/Fender_bassman_6g6b.pdf

[brewdude]

The bass channel has a 220k plate resistor and the regular channel has a 100k plate resistor.  I suspect that this could be the reason the voltages are not as you anticipated.

[brewdude]

Also, there are four triodes on the “B” and only two on the “C” node.

[AmberB]

It would seem to me that as long as both channels have an even number of triodes, the inversion should end up the same.  If one channel had an even number of triodes, and the other had an odd number, that would cause an out of phase problem where the come together for the PI.

[AmberB]

If you use a 0.022 decoupling cap before the 470k resistor on the output of each of the channels, you shouldn't need the .005 cap shown on the input to the PI tube.

[dwinstonwood]

The bass channel has a 220k plate resistor and the regular channel has a 100k plate resistor.  I suspect that this could be the reason the voltages are not as you anticipated.

Thanks brewdude!
The preamp plate load resistors are something I want to keep original. The Bass channel uses one 220K and three 100K's, and the Normal channel has one of each. I'll also want to keep the plate voltages "sort of close" to the schematic. But, I should be able to do that without using the 428V screen node (B) and forcing it down to 150V for the Bass channel. Plus, the grounding method I've settled on is easier to implement if I isolate the nodes for different parts of the amp, i.e., not using a power amp node for the preamp. I'll try to work something out in PSUD2.

[dwinstonwood]

If you use a 0.022 decoupling cap before the 470k resistor on the output of each of the channels, you shouldn't need the .005 cap shown on the input to the PI tube.

Thanks AmberB. Maybe I should use two .005's instead of the .022's, and then leave out the one you're referring to.

Does it matter if the coupling caps are before or after the mixing resistors?

And, yes, I was understanding the inverting the same way! Two or four should end up with the same non-inverted signal, yes?

[tubeswell]

It would seem to me that as long as both channels have an even number of triodes, the inversion should end up the same.  If one channel had an even number of triodes, and the other had an odd number, that would cause an out of phase problem where the come together for the PI.

Except that one of the stages in the bass channel is a cathode follower, so the signals will be out of phase if jumpered.

[dwinstonwood]

It would seem to me that as long as both channels have an even number of triodes, the inversion should end up the same.  If one channel had an even number of triodes, and the other had an odd number, that would cause an out of phase problem where the come together for the PI.

Except that one of the stages in the bass channel is a cathode follower, so the signals will be out of phase if jumpered.

Ahhhhhhhhh!

And that's all fine. I'm really interested in just having both channels to switch between. It looks like a lot of folks built this amp as a single channel. Some love the Normal, and others rave about the gain of the Bass channel. With a sluckey-style Dual config I can have both in a small chassis.

[tubeswell]

I'm really interested in just having both channels to switch between. It looks like a lot of folks built this amp as a single channel. Some love the Normal, and others rave about the gain of the Bass channel. With a sluckey-style Dual config I can have both in a small chassis.

The Bass channel is the best channel for guitar on this amp YMMV - check out this demo from about 2:50 on this clip

[sluckey]

I'm really interested in just having both channels to switch between. It looks like a lot of folks built this amp as a single channel. Some love the Normal, and others rave about the gain of the Bass channel. With a sluckey-style Dual config I can have both in a small chassis.

OK, but keep in mind that you can only select one or the other (Nor or Bass) but not both. You would need a common DPDT (ON-ON) switch as seen in my Dual Lite.

[bmccowan]

I have a Bassman 50 that I've rewired multiple times. Currently ch 2 is the normal 6G6-B channel - and I really like that channel. I tried the 6G6 bass channel and then decided I wanted something that diverged more and tried several different pentode circuits, but found that the tone stack loaded that pentode preamp down too much, and/or I'm just not focused on the right issue. So i took a break from that amp and ch 1 is up for grabs again. So I think you have cool project idea and I'll be watching to see how you make out.

[dwinstonwood]

Thanks for that vid tubeswell.

Thanks sluckey. I have a mini DPDT that I bought for my Hoffman 6V6 Plexi w/Hot Switch build, but didn't use (hot switch is on all the time). So, that's one less part to buy.

bmccowan, thank you for that info and your interest. If I can pull this off it will be cool to have both channels.

[AmberB]

If you're looking at the 6G6B schematic, there is no cathode follower in that preamp.  The 5F6A has a cathode follower to drive the tone controls, but the 6G6 does not.

[PRR]

...the 6G6B schematic, there is no cathode follower in that preamp. ....

[sluckey]

If you're looking at the 6G6B schematic, there is no cathode follower in that preamp.  The 5F6A has a cathode follower to drive the tone controls, but the 6G6 does not.

6G6, 6G6-A, 6G6-B, and 6G6-C all have a cathode follower in the bass preamp. So, which schematic are you looking at?

[dwinstonwood]

Not that I want to do this, but is there some kind of "little solid state thing," like a transistor, that will shift a signal's phase 180 degrees?

[sluckey]

Yes, thousands of them.

[dwinstonwood]

Yes, thousands of them.

Haha, well, I figured as much. Though, I know less about solid state that I do about tube amps. 

[bmccowan]

Haha, well, I figured as much. Though, I know less about solid state that I do about tube amps

They just took the stuff that was in tubes, shrunk it, and stuffed it in little aluminum cans. Now they are taking that stuff and stuffing it in little nano-cans.

[PRR]

some kind of "little solid state thing," like a transistor, that will shift a signal's phase 180 degrees?

[AmberB]

OK, I see that I was looking at the lines on the bass preamp wrong.  Once pointed out to me, I see the cathode follower. 

[dwinstonwood]

Thanks for that schematic PRR; I've saved it in my files for future reference.

I think every time I post a new project here I change directions and pursue something different. So, this thread is no exception.

I've redrawn my Musing 40 for 6V6's with some changes. The original wasted triode is now a Cathode Follower before the tone stack. Instead of fixed bias, I stole the cathode bias from the 5E3. I dumped the NFB. I thought I'd try the valve Wizard's "low-noise" grid stopper (it was recently mentioned in another thread). I don't know yet what values the B+ dropping resistors should be. I'll use PSUD2 to get in the ballpark.

I'm sure there are errors galore in this schematic.

[PRR]

V1: when driving a cathode follower, the 1.5k bias sets the plate too high. The cathode follower distorts way early. Some call this "toan". But will it be reliable in the first stage where signals are low? For reference, Fender liked 820 Ohms in this plan.

But you are driving a tone-stack with 100k first resistor. You don't need a cathode follower for that, Fender only did it for the 56k 5F6a tone stack.

Output stage: 0.1uFd grid caps is hi-fi, when you never clip. We clip guitar. Fat caps here take a long time to recover. Plagiarize many well-known plans and you will see 0.5u, 0.2u, even smaller when there is NFB around the power stage.

[dwinstonwood]

V1: when driving a cathode follower, the 1.5k bias sets the plate too high. The cathode follower distorts way early. Some call this "toan". But will it be reliable in the first stage where signals are low? For reference, Fender liked 820 Ohms in this plan.

But you are driving a tone-stack with 100k first resistor. You don't need a cathode follower for that, Fender only did it for the 56k 5F6a tone stack.

Output stage: 0.1uFd grid caps is hi-fi, when you never clip. We clip guitar. Fat caps here take a long time to recover. Plagiarize many well-known plans and you will see 0.5u, 0.2u, even smaller when there is NFB around the power stage.

Thanks PRR. I obviously don't know enough to know that I don't need the CF. 

OK, smaller output stage caps makes sense. An easy thing to change.

[dwinstonwood]

As far as the extra triode, I think I'll just not try to use it. The TW Express doesn't.
What I want to do is read up more on the effects of different cathode resistors/bias points and different plate loads. I'm convinced that one of the reasons the 6G3 sounds so good is because the input triodes have 220K plate resistors and share a 1.5K cathode resistor, which is sort of like having 3k (or, 2.7K) on each one, I think. Also, there's the whole thing of bypassed/un-bypassed gain stages. I've never experimented with that, either. I've read the theory, but haven't actually listened to the differences.
This is a pretty simple amp, and it would be "easy" to experiment with just those three things to see what I like.
Thanks

[PRR]

> sounds so good is because the input triodes have 220K plate resistors and share a 1.5K cathode resistor, which is sort of like having 3k (or, 2.7K) on each

There's no magic numbers. Ratios are more important. 220k and 3k is not a lot different than 100k and 1.5k. Though you also need to compare the 100k/200k to the load resistance.

Books may not tell you as much as an hour with a hot iron and a box of resistors.

[dwinstonwood]

...Books may not tell you as much as an hour with a hot iron and a box of resistors.

I think that will definitely prove to be true.

I did find a table of gain figures for different resistors and voltages on another site. Here's a portion of the data that the person gathered and posted:

Plate   B+ 180 volts

Resistor   Cathode   Gain
100k          1.8k        40
100k          2k           47
100k          2.2k        52
220k          3k           53

Thanks

[PRR]

> I did find a table of gain figures

That does not even look right. Larger cathode resistor, less current, makes more gain?? Then let's reduce current to zero, save lots of power!!(?)

Also, does this include any load? With any likely load, increasing the plate resistor can only go so far.

Also I wonder what "Plate B+ 180 volts" means. Is the plate forced to be 180V? (How?) Or is that the power supply, plate naturally sitting lower?

[dwinstonwood]

Again, PRR, I don't really know what I posted. 
Somehow I'm going to have to gain a "bigger picture" perspective of not just a common-cathode triode, but how it relates with the preceding guitar pickup and cable on one side, a tonestack on the other, and a certain DC voltage too. I realize I can't look at just one isolated component.
So, swapping real resistors and hearing the results will be the best way.
That's actually what I did with my AC4. After about five or six changes to EF86 resistors I finally hit on what sounds "right" to me. I may have broken several datasheet rules, but it sounds the way I want it to. 
Thanks

[shooter]

found this chart, came in handy when I was BB'n circuits

[dwinstonwood]

Thanks shooter!

[dwinstonwood]

I'm playing around with PSUD2, and I'm wondering if it's a bad idea to add another RC filter before the choke. The purpose would be to lower the screen voltage by about 15 to 20 volts, or so.
Thanks!

[PRR]

You don't actually need C2 for anything. There's as much filtering as almost any commercial amp already. You just did that because PSUD does not allow arbitrary placement of dropping resistors.

Skip R1 and C2. Tell PSUD that L1 is 4H 1100r. It won't know you really mean a 4H + 100r plus a 1k.

[dwinstonwood]

...You just did that because PSUD does not allow arbitrary placement of dropping resistors.

Yup, that's exactly why.

Skip R1 and C2. Tell PSUD that L1 is 4H 1100r. It won't know you really mean a 4H + 100r plus a 1k.

Perfect. Thank you PRR.

[dwinstonwood]

If I do decide to use a 1K dropping resistor before the choke, is this math correct?

20V dropped across 1K resistor with around 22mA of current (6V6 screen current + three 12AX7's current draw):

20 Volts x .022 Amps = 0.44 Watts

So, a typical 3W resistor would be more than enough for this location? Or, am I not seeing all of the current?

Thanks.

[tubeswell]

...
20V dropped across 1K resistor with around 22mA of current (6V6 screen current + three 12AX7's current draw):

20 Volts x .022 Amps = 0.44 Watts

So, a typical 3W resistor would be more than enough for this location? ...

Yes and Yes

[dwinstonwood]

Thanks tubeswell!

[dwinstonwood]

I've come up with this. Nothing really original. As the title block on the schematic says, it's an AA864 preamp with some TW Express component values, and a cathode biased power amp based on the 5E3. I do know that I really like this preamp from my 6L6 version.

[sluckey]

Nice schematic. Lot of potential.

[dwinstonwood]

Thanks sluckey! I still want to go the Hoffman Stout chassis route. So, the layout you designed is most of the way there, already.

What's great about this hobby is that everyone can go in their own direction - as long as the design is safe and stable. I doubt I'll ever get into the pursuit of replicating an exact clone of some famous/iconic amp. For now, my interest seems to be "simple, inexpensive, and (relatively) easy to troubleshoot."

[pdf64]

For zero effort, DC heater elevation could be implemented, by referring the 2x100ohm heater balancing resistor node to the 6V6 cathodes, rather than 0V common.

[dwinstonwood]

For zero effort, DC heater elevation could be implemented, by referring the 2x100ohm heater balancing resistor node to the 6V6 cathodes, rather than 0V common.

Thanks PDF64. That's something I've never tried, but have thought about. The Stout chassis is compact, and there's not a lot of open area to keep high-current wires far apart from each other. Your suggestion is a great one.

[dwinstonwood]

I've ordered the chassis and parts for this build. Now I'm on the fence as to whether I should go cathode, or fixed bias. I have the parts for either, and I'm finishing up layout drawings for both. I've been reading several threads over at Ampgarage about cathode biased Expresses (some dating way back). Some say it rocks, others point out the drawbacks like blocking distortion and flubby, loose low end. The purists think it's sacrilege. Cathode resistor and bypass cap choices are all over the place. I've gotten the impression that it can work well, but has to be dialed in with some trial and error.
Anyway, I'm leaning towards fixed bias with Express component values. It won't be a clone, but most of the signal path will closely resemble one or more of the (many) Express schematics found in the library here.
But, like I said, I'm still on the fence. 

[tubeswell]

Due to the high input impedance on the LTP, you can get away with a lot smaller coupling capacitance feeding the LTP input/inverting triode (without losing any useable 'guitar bandwidth'). So you might want to replace the 0.022uF with 1000pF

[bmccowan]

Now I'm on the fence as to whether I should go cathode, or fixed bias.

If you have room, you could do both - switchable. Sluckey has a sketch in his scrapbook. Outside of that, I would build it as Ken intended, although I know its not a full clone, and see what you think. Pretty easy to convert an amp from fixed to cathode bias. Although I really like cathode bias amps, I once converted a Princeton Reverb to cathode bias, and did not like it at all. But I did not give it much of a chance - 20 minutes or so. It was like the original designer knew what he was doing, better than me

[dwinstonwood]

Thanks tubeswell and bmccowan.

One reason I'm set on using turrets is that they make it easy for me to switch things like coupling caps. Just reflow a dab of fresh solder and parts come right out.

...I would build it as Ken intended, although I know its not a full clone, and see what you think. ...It was like the original designer knew what he was doing, better than me

Yeah, going against that logic is what's nagging me. Fixed bias it is.

I also went through a similar thing with my AC15 where I grafted in the Princeton Reverb cathodyne PI, never bonded with it, changed it back to the Vox LTP, and now it sounds "right."

Thanks.