6V6GT Single-Ended Project

[dwinstonwood]

This project has morphed a few times over the last couple of months. It started out intended for two 7C5's. It still uses some other less common tubes. I could even wire it for a single 7C5 since I have some NOS examples, and some unused Amphenol Loctal sockets.

These previous threads "chart the path" this project has taken:
https://el34world.com/Forum/index.php?topic=31870.0
https://el34world.com/Forum/index.php?topic=32131.0

So, a big thanks to DummyLoad, HotBluePlates, bmccowan, shooter, and others who shared their insights and knowledge. I learned some good stuff by stumbling my way through this. There's also a lot of things borrowed from Merlin's books in this design.

At any rate, I either have, or have ordered, all the bits I need to build it. Here's the schematic. Or, should I just post my layout...

[shooter]

looks like it'll rock! 
once it's "good enough to jam", find a 4 X12 cab, The guitarist I used, plugged into one, played a Hendricks solo lead and everyone stopped what they were doing.   That was through a similar "Champ" 4w build.  The amp went into the wild that day
 


Spend extra brain-time on the TS position and layout.  I would pilfer through Steve's pages til I found something close, then steal the man blind, yet, never quite able to achieve his A-game quality.

[bmccowan]

Here's the schematic. Or, should I just post my layout...

 
It'll be a gainy little beasty, that's for sure. Speaking of stealing from Steve (I've already got his car and a couple of TVs) the raw control is one of my favorites.
Since I think 3 preamp tubes is a bit much for a SE 6V6, I'll be watching to see how you make out.

[shooter]

Since I think 3 preamp tubes is a bit much

I'm beating he's a closest death-metal guy 
they might come in handy when he finally realizes anything less than a KT88 or EL34 just isn't worth the time 


take out V1 n the 6V6 might stand a chance.

[dwinstonwood]

Aw, come on! There's a gain pot before the pentode, and the third tube is a ~unity-gain CF. I'm betting it can be reined in.

[shooter]

I'm betting it can be reined in.

ya, volume on guitar 1, gain 1, Drive 3-4.  when you wear out that half of the pots just rewire backwards to take advantage of the un-used section 

[Willabe]

I'm betting it can be reined in.

ya, volume on guitar 1, gain 1, Drive 3-4.  when you wear out that half of the pots just rewire backwards to take advantage of the un-used section 

       

[dwinstonwood]

Ok, I've been warned (which amounts to just being more determined).

I'll have fun building it. But, it still has less gain than my Hoffman 6V6 Hot Plexi. Besides, it'll be easy enough to solder in voltage dividers between stages if I think it's necessary. And, I've already made the board...

[AlNewman]

Looks pretty hairy.  Maybe preemptively lower value volume pots could be used, like 500k or even 250 k...? 

[bmccowan]

No doubt you will get there - you always do. One of the nice things about a unique circuit is that you feel more free to experiment with resister, cap, and pot values until you get what you want. FWIW I've built a couple amps with a pentode in V2. One has a gain pot between V1 and V2. I had a blast experimenting with plate/cathode/grid resistor values on V1 and V2. I ended up with values very similar to what Tubenit uses in various HoSo56 circuits. After cleaning out Steve's house, I stole the stone stack from Dr. Z. That was harder as unlike Steve he locks his doors.
BTW - nice clean schematic. Do you draw in _.sch format? I think not, but if you do, please share that file.

[dwinstonwood]

No doubt you will get there - you always do. One of the nice things about a unique circuit is that you feel more free to experiment with resister, cap, and pot values until you get what you want.

Thanks bmccowan. That's just what this project is all about. Learning by doing. And, like half of my amps, I doubt it will ever be played through by anyone other than me.

BTW - nice clean schematic. Do you draw in _.sch format? I think not, but if you do, please share that file.

I use DIYLC. I have JSchem and ExpressPCB installed. I'll make an sch file and share it!

[kagliostro]

Ciao Dwinstonwood

I like the idea 

Just a question for you

There is a particular reason for the choice of a 5879 instead of an octal pentode (6j7 - 6sj7 - 6sh7 ...... ) ?

Franco

[dwinstonwood]

There is a particular reason for the choice of a 5879 instead of an octal pentode (6j7 - 6sj7 - 6sh7 ...... ) ?

Franco

Thanks Franco. No particular reason for the 5879. But, maybe I consumed a bunch of internet advice saying the 6SJ7GT is too microphonic? They are affordable, though.

One plus for using the 5879 is that rewiring the socket for an EF86 is not complicated. The heaters and cathode connections are the same.

[kagliostro]

5879 is surely a good tube, if you go for noval there is also the 6AK5 (5654), a cheap tube, used in tube mikes

BTW, to defeat microphonic effect on some penthodes go for lower gain settings

Franco

[dwinstonwood]

bmccowan, converting my schematic into an .sch file is slow progress. I've never really used JSchem before. I'm maybe halfway there. Here's a screenshot. I'll share the .sch when I get it finished and proofread a few times.

[bmccowan]

Hey thanks - that is really nice of you. I almost feel guilty for asking. But not quite as I think that once you figure its kinks, you will enjoy using it. The snap feature is helpful as are the component libraries And Doug maintains a library of sch schematics. You can open something similar and edit rather than starting from scratch. You also can cut and paste sections of schematics. The instructions for doing that are in the help docs. The only nag with that is when pasting a section that was drawn using snap with a section that was drawn without snap. I just get it close enough for rock & roll and then go back and clean it up.
Maybe someone here will say, "hey, you are doing it the hard way!"
About the 5879 and microphonics. Its smart to be careful with that socket. I just use tiny rubber washers or O rings from the hardware store on either side of the screw and nut and avoid using those self tapping screws. Some folks "prove" that pentodes are microphonic (actually all tubes are) by tapping them with a pencil. The cure for that is easy - stop tapping the tube with a pencil. A while back PRR posted a pic of a rubber headed tube tapper. https://www.hammermuseum.org/hammer-highlight-tube-tapper/

[dwinstonwood]

Thanks bmccowan, good advice on the socket mounting. With this being a head, it should hopefully have adequate isolation from the speaker.

I pretty much finished the .sch schematic in ExpressSCH. I need to review it a few times before launching it onto the web. But, here's an image of it.

I finally got started on the build this weekend. I ordered a blank 17x6.5x2.5 chassis from Hoffman (not only the best price, but very fast shipping). It's nice, thick aluminum like the Stout chassis.

I've finished building the board. The under-board wires come up through small holes near the turrets and wrap around them. I think this is more reliable than soldering into the bottom of the turrets. The chassis is marked up for drilling; I'll get that done next.

[DummyLoad]

5693 is my go-to for octal pentode - Industrial low microphonic sub for 6SJ7. They sound really nice.

5789 is much more tame, am assuming your B+ is going to be around 250-300V, so should have AV of about 100 - You could take that down to around 50-60 with 470K g2 and 1.5K cathode R, might be bit more stable. If it were mine, I'd start with 100K anode, 220K g2 and 1K cathode.

A V divider for the gain pot will take some of the touchiness out of that control.

Your plan is going to want to howl. Metal amp?     

--Pete

[shooter]

 
ya, he's been warned on the other hand I just seen in my feed mullets are making a comeback, big production metal bands might be around the  corner 

[dwinstonwood]

5693 is my go-to for octal pentode - Industrial low microphonic sub for 6SJ7. They sound really nice.

Thanks Pete. Is the 5693 that red metal tube?
I've been considering adding a voltage divider before the pentode. Something along the lines of the pic below.

As for the 5879, I spent a lot of time studying the datasheets and thought I had it set up good. From what I gathered you could either go lower gain and higher voltage output, or higher gain and lower voltage output. I'll revisit the datasheets on that point.

But now... I'm thinking of buying one of these:
https://edcorusa.com/products/xse15-5k-15w-5k-ohms-single-ended-tube-output-transformer?_pos=15&_fid=c861be190&_ss=c&variant=41118788223163
They're only $47 and are rated for 100mA (the Hammond 125ESE is rated at 80mA). I've read where people say the Edcor works great with a 6L6GC. I have Edcors in my EL34 SE hifi stereo amp and they are solidly built USA tx's.

If so, I'll get one of these chokes, too. They're 6H/200mA at around $20:
https://catalog.triadmagnetics.com/item/inductors-and-chokes/smoothing-filter-chokes/c-14x

That will put me back to where I started this whole project with a 6L6GC or 5881.

And shooter, try as I may, my hair won't grow long enough in the back.

[shooter]

 


mines just past the shoulders now and i keep scissoring around the ears, yesterday I laughed as i'm shaving thinking it's sorta mullet looking, then I see the feed today.....


Pete has put me on the right path with my builds every time so don't overthink just steal whatever he's offering, then when the fingers get sore from playing, you can sit back and think

[dwinstonwood]

These are the tables I was referring to above. To be honest, the practical distinction between "Voltage Output" and "Voltage Gain" is not totally clear to me.

[Merlin]

These are the tables I was referring to above. To be honest, the practical distinction between "Voltage Output" and "Voltage Gain" is not totally clear to me.

That datasheet does have an odd way of putting it.

[kagliostro]

If the 5693 is not disposable (or too expensive) I can suggest to give a try to the russian 6Ж8, it was planned for military use and it has a sturdy structure

Here’s a detailed comparison between the **6SJ7** and the **6Ж8**:

### 1. Origin and Purpose
- **6SJ7**: An American-made tube widely used in audio amplifiers, radios, and low-frequency circuits. It was designed to provide stable gain with low noise for commercial and home applications.
- **6Ж8**: A Russian tube with similar specifications, often produced for military and industrial purposes. It's designed with greater resilience to mechanical and environmental stress.

### 2. Structure and Build
- **6SJ7**: Typically comes in a glass tube or metal case (depending on the version), offering decent shielding. The glass versions tend to be more prone to microphonics than the metal-bodied ones.
- **6Ж8**: Generally more robust and built to withstand shocks and vibrations better, making it less susceptible to microphonics. Its construction quality tends to be more durable due to military-grade standards.

### 3. Performance
- **Gain**: Both tubes offer similar gain levels, though the 6Ж8 can sometimes provide slightly more stable performance under high-gain conditions, depending on specific production batches.
- **Noise**: The 6Ж8 typically generates less noise than the 6SJ7, especially in noisy environments or with potential interference. This makes it better suited for applications requiring signal purity.
- **Microphonics**: The 6Ж8 generally has lower microphonics due to its sturdier construction, while the 6SJ7 may exhibit microphonics at high gain levels or in vibrating environments.

### 4. Availability and Compatibility
- **6SJ7**: Still relatively easy to find due to its historical popularity and large-scale production, especially in vintage American versions.
- **6Ж8**: As a Russian-origin tube, it may be less common in some Western markets, though it's readily available in Europe and the US through specialized suppliers. It may also have slight pin configuration differences, so compatibility should be checked before use in circuits.

### 5. Ideal Applications
- **6SJ7**: Recommended for vintage audio amplifiers and projects where microphonics are not a major concern. It’s also appreciated in DIY audio projects.
- **6Ж8**: Better suited for applications with concerns over vibration or electromagnetic interference, such as precision instrumentation and circuits intended for military or industrial use.

### Conclusion
Both tubes are great choices but with slightly different optimal use cases. The **6SJ7** is well-suited for general audio applications and hobbyist projects, while the **6Ж8** provides enhanced durability and lower microphonics, making it preferable for demanding environments or where high reliability is needed

.

Of course the 5693 is the better of the three

Franco

[shooter]

output voltage is simply that what you measure at the output and it's "peak" voltage so 1/2 the sin wave.
voltage gain is Vout / Vin typically expressed peak to peak or rms, so 1vac in, 10vac out = voltage gain of 10

[DummyLoad]

Yes, 5693 are RCA Special Red tubes.

Choke is a bit overkill, but at that price, go for it. Your amp pulls 60mA or thereabouts worst case, depending on how hot you run the 6V6. 7KΩ OT would be a better choice if you're building with Champ power supply plan, however, 5KΩ will work fine. The multi-tap primary Hammond is also a good part.

I <3 Edcor iron, good stuff, only thing I don't like about their "off the shelf" parts is the single tap secondary, but reducing secondary winding complexity keeps their costs down and reduces build time. Meh! I'd rather have the extra taps.

--Pete

[dwinstonwood]

Thanks Franco! That's great info. I'll search around online for the mysterious 6Ж8.

And thanks Pete! I'm also interested in this red metal 5693. I will do some searching for it, too, and see if I can get one at a good price from a reputable seller.

I agree that having multiple primary impedances, and secondary taps is the way to go, since this will be a head. I can order both a Hammond 125ESE OPT and a 159R choke from Hawk and save on shipping. I've decided I want to be able to use a 6L6GC. You can never have enough spare transformers on hand.

I added a couple of voltage dividers that each chop the voltage in half after the gain and volume pots. Sorry for all of the schematic images (they're small PNG's).

[DummyLoad]

I designed a PCB for Champ amps and I added a couple of mods other than the up-rated output stage. I got tired of punching G10 and staking turrets or eyelets, for every build, so now I just stuff a PCB. They are high temp. FR4, 2.4mm thick, 3oz copper clad. Good stuff.

Anyway, if you're going to jump down the 6L6GC/5881 rabbit-hole, the attached schema with telemetry should help with your final design goals. The One Electron UBT-2 is 4.8K:4/8/16 and is no longer made.

If you're still in the buying phase of the OT, then consider the Heyboer 15W SE OT sold by Amp Parts Direct part no. APD-8031, a clone of the ClassicTone 40-18031 - Details here. I have built several EL34/6L6GC Champ amps with the ClassicTone part. Very positive reviews on that part. I have a pair of the ClassicTone part for spare/future builds, so I haven't auditioned the APD part as of yet. If it's a true copy, it should please. If I were to build another one of these for a customer, I'd probably use the Edcor, or Heyboer part.

--Pete

[dwinstonwood]

I scored a tested NOS 5693 from a 99.8% positive feedback tube seller.

Can I assume that Leo's operating conditions for the 6SJ7 in his 5C1 are a solid starting point for the 5693? I assume they're probably borrowed from a datasheet?
https://el34world.com/charts/Schematics/files/Fender/Fender_champ_5c1.pdf

Thanks.
David

[shooter]

I would assume
I would double check the datasheet pinout to doublecheck they're the same and the key is solid

[kagliostro]

Ciao David

I'd love to hear how your project is going

Franco

[dwinstonwood]

Hey Franco. I plan to drill the chassis today; the weather's good. I live in a townhouse, so I have to drill metal outside which means setting up a bench and the drill press. Setup and cleanup take much longer than the actual work. Since I bought a 5693 it's going to be "All-Octal" so, it's good I didn't get around to drilling holes until now. Once that's done I can start building the thing.

In the meantime I've been messing with PSUD2. If these voltages are close, then I'm right where I want to be: about 335 on the plate, and 250 on the screen.

[kagliostro]

Ciao David

Many thanks for the update

townhouse, do you mean something like this ?





dont you arranged a workshop in your garage ?

the 5693 is surely a good choice, now, as you say, is a full octal tube project, I like very much this idea

keep us updated

Franco

[shooter]

dont you arranged a workshop in your garage

 
my thought exactly, which is more important, obviously a work bench!

[dwinstonwood]

I don't have a garage. And, yes, it's like a rowhouse. I'm in the city.
But, I live 9 minutes from work. I bought my car new in 2018 and the odometer just hit 13K miles.

All of the larger holes are drilled, along with the rectangular IEC inlet hole.

After I'm sure how I want the sockets rotated for the shortest wiring, and the least crisscrossing, I'll drill the socket mounting holes. I also need to arrange the choke and the OPT to decide where I want to drill those mounting holes.

The board fits good (I measured for the mounting holes three times and drilled once).

[dwinstonwood]

Weekend progress.

[DummyLoad]

I scored a tested NOS 5693 from a 99.8% positive feedback tube seller.

Can I assume that Leo's operating conditions for the 6SJ7 in his 5C1 are a solid starting point for the 5693? I assume they're probably borrowed from a datasheet?
https://el34world.com/charts/Schematics/files/Fender/Fender_champ_5c1.pdf

Thanks.
David

 

5C1 plan uses grid leak scheme and is biased for highest gain, so I presume that the theory was that with just two tubes the 5C1 plan was not going to distort as easily, however, running the pentode with so much gain, input sensitivity can be so low with so much gain that input saturation with modern pickups is a reality. With your plan, and so many cascaded stages, it may potentially be unplayable.

Suggest starting with the pentode biased for lower gain, experiment swapping different values of R Anode, G2 R and cathode R increasing gain to find the distortion & tone to your liking. In other projects I've used the Matchless DC30 EF86 values that result in a pleasing tone, mostly tweaking for less gain to tame things a bit, usually by just fiddling with G2 voltage.

--Pete

[dwinstonwood]

Thanks Pete. Here is that table that floats around the various forums (I don't know who put it together).

I can start with the DC30 values...

[dwinstonwood]

Or, looking at other pentode values, I might start with this:

Rp - 100K
Rg2 - 330K
Rk - 680Ω
Ck - 4.7uF
Cg2 - .047uF

I'm sure it will take some time experimenting to find what I like.

[kagliostro]

Ciao David

Here is how Matchless arranged the circuit on the Clubman (6SH7 tube)



Franco

[HotBluePlates]

... I'll search around online for the mysterious 6Ж8. ...

We might think of "ECC83" as the "European name for a 12AX7."  In general, the ECC83 has all the same characteristics as a 12AX7.

"6Ж8" is just a "Russian-made metal 6SJ7".  A challenge may be that the Russian data sheet specifies a maximum screen voltage of only 140v, where the American 6SJ7 can go up to 300v.

The 5963 is de-rated to only 125v on the screen (though a 300v supply is allowable on the other side of the screen resistor), so perhaps you don't need to worry too much for either the 5963 or 6Ж8.  Besides, a lower screen voltage will probably yield higher gain, especially if you also up the plate load resistor (there's a balancing act between screen volts, bias volts, Gm, and plate load to get the desired stage-gain and maximum output voltage swing).

[dwinstonwood]

Thanks Franco and HotBluePlates.

Tonight I'm going to re-read Merlin's chapter on small signal pentodes. It's been a long time since I read it, and I don't remember much of it. I think he does say that you can still retain the pentode's tonal qualities at lower gain settings.

Just glancing at one of his EF86 examples, he has a 47K Rp, 330K Rg2, and a 680 Rk for a gain of 85. That's with 300V before the Rp. My B+ should be somewhere in the neighborhood of ~225-ish.

David

[kagliostro]

I remember that he say something like in the past pentodes were used in V1 because of theyr high gain but there may be microphonic problems due to the high gain settings, to maintain the positive effect on tone of the pentodes his council was to use it as V2 at a lower gain setting, this way they are less prone to be microphonic and give the best they can

Or this is what I remember

Ciao


Franco

[dwinstonwood]

...to maintain the positive effect on tone of the pentodes his council was to use it as V2 at a lower gain setting...

Yes, this is exactly what I'm aiming to due.
I have the pentode in the V2 position, and I'm reading up on how to set it up for lower gain.

David

[kagliostro]

You told it previously but seemed unsure about what was saying Mr Merlin so I told what I remembered as to confirm also you remembered correctly 


Franco


p.s.: Do you have the glass or the metal version of the 6SF5GT tubes ? (may be I have some glass version but not sure about)

[dwinstonwood]

Thanks Franco! I have the 6SF5GT glass tube. The 5693 is the only metal tube.

[kagliostro]

If I've those 6sf5gt also mine have to be glass, not metal, but my tube stock is placed in a place in my garage that is difficult to reach because there is letterally a mountain of stuff in front of it

Franco

[dwinstonwood]

there is letterally a mountain of stuff in front of it

Franco

Do you own climbing gear?

I'm going to attempt Merlin's "Designing a Pentode Gain Stage the Easy Way" process.

So, if I want a gain of say, 80: The 5693 datasheet gives a gm range from minimum to maximum with 1650 uMHO as the average. That's 1.65mA/V, so 80 divided by 1.65mA/V = 48,484 ohms, or a 47K plate resistor.

With a 300V plate supply, the load line hits 6.4mA. I'll just pretend the -1V grid curve is my 0V curve (Merlin says to draw one in where we want it), and levels off around 7mA (the load line passes a bit low through the knee).

Looking at the 5693 Average Characteristics graph, 0V and 7mA lands a bit less than halfway between the 100V and 75V screen curves, so I'll fudge it at 85V screen voltage. According to the graph the tube reaches cutoff at Vgk = -4.8V. Biasing warm would be about -2V with a plate current of about 3.2mA.

5693 Anode current is 3mA and the screen is .85mA, so the ratio is 3.5. So, 3.2 / 3.5 = .91mA screen current.

300V - 85V = 215V. 215 / .91 = 236,200; a 220K resistor is closest.

Cathode current is 3.2mA + .91mA = 4.11mA. 2V / 4.11mA = 486.6 Ohms, so a 470 Ohm cathode resistor is close.

Finally, I have:
Plate resistor - 47K
Screen resistor - 220K
Cathode resistor - 470

Hopefully, my math is correct and these values would give a gain of about ~80.

I did make an error in my calculations, and did my best to fix it.

[kagliostro]

Great job David 

I always get punched in the face by those calculations

I would like to be able as you are

Franco

[dwinstonwood]

Thanks Franco! I'm sure I'm about 20%+ off.

I tried again with my supply voltage of 225V B+ and got these values:

Plate - 47K (rounded from 48K)
Screen - 270K (rounded from 285K)
cathode - 680 Ohm (rounded from 700 Ohm)

Obviously, I would round to the nearest resistors, maybe 680K and 2.7K or 3.3K?

Edited to fix math mistakes.

David