• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

DC coupled single ended cathode follower 6SJ7/6V6 amp project

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Maybe I could also ask you guys what it would take to make the output stage work as it’s intended. I am ready to give up the dc coupling since it seems to be the most limiting factor. I’ll re read the articles about the cathode follower output stage on tubecad since it is the most reliable and extensive information source about it.

By the way, the tektronix power transformer I intend to use has many different HV secondaries, so that could be useful to make the driver stage swing higher than the output stage. All this discussion has been really informative for me since i’m tackling something that is a lot deeper into tube circuit than what i’ve done before. I thank you all guys !!

You could make a simple SE. It won't be hi-fi, but it could make pleasant noises.

I could also order my parts and populate the Tubelab SSE board that I have received two months ago. It may be more logical to put in use the big edcor OPTs that are waiting in their boxes :D
 
Maybe I could also ask you guys what it would take to make the output stage work as it’s intended. I am ready to give up the dc coupling since it seems to be the most limiting factor.

You're on the right track of investigation with forgoing the DC coupling.

Also read here:

Cathode Follower | Tubelab

Because you've already got the octal sockets, two stages of 6SN7 (maybe choke loaded) could also be a solution to get the voltage swing you're looking for.
 
I have quite a nice stash of tubes and sockets so i’m not limited to the octal ones. The reason why I got into tube amps is that someone gave me all his material for building guitar amps, so I started out with a lot of parts. It kinda helped !!

Tubelab_com : your cathode follower amp is very interesting ! If you have any new informations, I would be more than happy to learn more about it !
 
Member
Joined 2015
Paid Member
The Raymond H. Bates circuit was published at page 62 of Radio News, 1949, issue 11. The same year, mr. Bates published the non-direct coupled version of this amplifier on the june issue of Audio Engineering, page 30. This was presented as "low power cathode coupled amplifier". The more professionally-oriented magazine distanced itself from the autor by presentig it as novelity in the "reader's contribution" section. The author also toned down the claims, and dropped the absurd 4.5W power specification.
 

Attachments

  • bates_audio_1949.jpg
    bates_audio_1949.jpg
    333.2 KB · Views: 272
Triodes are more linear. It's not nonsense, just an observation. I use gNFB on my triode amps, too. Open loop is a pain in the butt to get flat IMHO. I'd triode-strap the 6SJ7, and use a triode strapped 6DQ6 or something, but that's just me, right? Also I only build PP amps... YMMV
 
What's wrong with a 6SJ7 driving a 6V6 RC coupled?

That's where I started in the early 60's at around age 10. This was before computers and the only place to find schematics was a magazine article, a book from the library, or by ripping something apart and tracing it's circuit, something I did a lot. The 6SN7 driving a 6V6, powered by a 5Y3 schematic was obtained by carefully tracing the circuit of a friends old Fender Champ. That design worked for guitar and my record player. A ham radio guy taught me how to convert grid leak bias into cathode bias and it got better. When I couldn't find 6V6's in the trash, I used a 6BQ6.

Tubelab_com : your cathode follower amp is very interesting ! If you have any new informations

I did all the cathode follower stuff 12 years ago mostly for one purpose. Microchip had sponsored a design contest in Circuit Cellar magazine featuring their then new dsPIC chips. They wanted some unique designs featuring their new chips, so I set out to build a rather unique vacuum tube HiFi amp with some efficiency and power output improvements afforded by modulating the power supply voltage under DSP control. Low power output....low B+ supply voltage.....big transient coming, crank up the supply voltage just for that instant, then turn it back down. This is somewhat similar to a class H solid state amp.

I had just finished up a two year research project in my day job designing such a system for cellular towers and phones, so I just re-applied it to a tube amp. Changing the supply voltage on a conventional audio amp while it's playing music will cause all sorts of audible artifacts or worse, so I first needed to create an amp where the power supply knob could be turned without any change in the sound, provided it furnished enough voltage to avoid clipping. The best way to do this is a cathode follower due to its extremely high power supply rejection ratio (PSRR).

The amplifier that I designed for this project could be used without the DSP controller, but you certainly could not run it on 450 volts and get 20 watts out of it for long. The design could be adapted for several different tube complements, but the 6336 and 6LW6 tubes performed the best of any tubes I tested. I am enclosing the entire magazine article from the October 2009 issue of Circuit Cellar magazine. It did win a prize in that contest which resulted in me writing this article for the magazine. That provided me with a good chunk of cash, and got me a phone call from Hartley Peavey asking if the technology could be adapted for guitar amp use.....I had to tell him that it would work, but just didn't sound good when cranked to 11....

I built a few cathode follower amps during this design project. The first was a breadboard that worked quite well. It's schematic can be found in post #22 of this thread, pictures of the PC board version in post #55, schematic (without the bias pots) in post #62:

20W cathode follower amplifier
 

Attachments

  • 2910014_Anderson.pdf
    289.1 KB · Views: 170
What's wrong with a 6SJ7 driving a 6V6 RC coupled

Absolutely nothing. But it’s fun to go where you’ve never been.

That's where I started in the early 60's at around age 10. <snip:
Cool story ! Thanks ! I’ve started to read the thread.

The same year, mr. Bates published the non-direct coupled version of this amplifier on the june issue of Audio Engineering, page 30.

Do you have access to the orher pages of the article ?
 
Member
Joined 2015
Paid Member
The full issue of Radio News (issue 11 1949) and Audio Engineering (june 1949) are currently available for download as pdf files from the americanaudiohistory.com website. The audio engineering article is a single page one. This article has been published at least on a further magazine because I've found another scan on the web with different impagination but the exact same text and schematic; I have been unable to recognize the magazine. I plan to build in the near future a small amplifier with cathode follower output stage, and this Bates schematics is often quoted as example. It does have a few seriuous drawbacks, however. It has been obviously designed with a trial and error approach, using the parts that the author had on its drawer. This means that it probably works, unlike other designs from magazines, but it is equally probable that it may work better after optimizations and checks on a circuit simulator. The 6SJ7 / VT116 really looks to have been grabbed from a scrapped WW2 communication receiver. I have at least a dozen metal VT-116 and this tube has the important advantage to be really cheap, but I don't see a compelling reason to use it on a modern Hi-Fi build; I tried it on a test circuit with the standard component values listed on the databook and I was not impressed. As small output tube, 6V6 is ok but a tube with higher gm such as EL84 / 6BQ5 will ease the driver tube job. It would be interesting to compare on the simulator the original 6SJ7-6V6 tubes to something more recent at the same power level, such as EF86-EL84. The Tubelab cathode follower amp with sweep tubes is a far more reliable and documented design and I am certain that it work as described, but it does not meet my current need to have something small on power and size. This is why I currently have the Bates schematic around on my desk again.
 
The full issue of Radio News (issue 11 1949) and Audio Engineering (june 1949) are currently available for download as pdf files from the americanaudiohistory.com website. The audio engineering article is a single page one. This article has been published at least on a further magazine because I've found another scan on the web with different impagination but the exact same text and schematic; I have been unable to recognize the magazine. I plan to build in the near future a small amplifier with cathode follower output stage, and this Bates schematics is often quoted as example. It does have a few seriuous drawbacks, however. It has been obviously designed with a trial and error approach, using the parts that the author had on its drawer. This means that it probably works, unlike other designs from magazines, but it is equally probable that it may work better after optimizations and checks on a circuit simulator. The 6SJ7 / VT116 really looks to have been grabbed from a scrapped WW2 communication receiver. I have at least a dozen metal VT-116 and this tube has the important advantage to be really cheap, but I don't see a compelling reason to use it on a modern Hi-Fi build; I tried it on a test circuit with the standard component values listed on the databook and I was not impressed. As small output tube, 6V6 is ok but a tube with higher gm such as EL84 / 6BQ5 will ease the driver tube job. It would be interesting to compare on the simulator the original 6SJ7-6V6 tubes to something more recent at the same power level, such as EF86-EL84. The Tubelab cathode follower amp with sweep tubes is a far more reliable and documented design and I am certain that it work as described, but it does not meet my current need to have something small on power and size. This is why I currently have the Bates schematic around on my desk again.

We seem to have similar goals with this circuit. I have thought of modifying it to use an ef86 driver and el84 output stage, but since I wanted to keep the dc coupling at this moment, I decided not to do it. But now that I learned more about the limitations of the dc coupled version, it becomes tempting to try to modify the rc coupled version.

Something else that could be done would be to modify the mullard 3 watt amplifier’s output stage into a cathode follower one. I haven’t checked for the component values, but it could be a matter of adjusting some resistor values, removing the NFB loop an putting the opt on the cathode instead of the anode. The bias of the el84 might not be good with the 250 ohm DCR of the OPTs primary though.
 
You always need to remember that a cathode follower output stage needs much higher signal voltages from the previous stage. In some cases there will be inadequate gain for this, sometimes inadequate voltage swing, and almost always increased distortion. Hence such an amplifier needs careful design from the start, not merely "modifying" an existing quite different circuit.
 
You always need to remember that a cathode follower output stage needs much higher signal voltages from the previous stage. In some cases there will be inadequate gain for this, sometimes inadequate voltage swing, and almost always increased distortion. Hence such an amplifier needs careful design from the start, not merely "modifying" an existing quite different circuit.

I was mostly referring to the mullard circuit because it has the same two valves that I would like to use. But as you said, there would probably be nothing left from the mullard circuit at the end of the modifications. I just thought that it would be easier to start with a circuit that uses the same valves. Do you think that the Bates (rc coupled) circuit would "work" with the valves replaced by the ef86/el84 as a starting point, and that the components values could be adjusted afterwards ?

Btw, I have found this thread. The guy built the rc coupled version using an EL84 output and a 6AG5 driver. He surprisingly claim to be able to drive "vintage 3 way speakers" with it. Looks like that, without being perfect, the kind of modifications I’d like to do could be done.

single-ended cathode follower output stage
 
Last edited:

PRR

Member
Joined 2003
Paid Member
As for Bates: a then-new RCA 45 player, a 12-inch Jensen, a small livingroom, and unknown taste in dynamics.... maybe he could "astonish" himself. I've been "impressed" by some quite modest amplifiers in a similar situation.

Maybe I could also ask you guys what it would take to make the output stage work as it’s intended.........

An ideal device as a power amp, choke/transformer-loaded, 250V supply, will swing +/-250V (500Vpp) around the idle point.

Cathode follower is "unity gain". So the driver has to make 250V peak or 500Vpp.

6V6 Triode won't swing full supply voltage. But also won't reach unity gain. So for stick-in-sand figuring, we still need 250V peak.

As DF96 and others say-- this won't happen by happenstance.

A good rule-of-thumb for a resistance-loaded driver is: peak output at 5%THD is near 20% of supply. Or the other way: supply must be > 5X the peak out needed.

5X 250V is 1,250V supply. :eek:

We "can" cheat and bootstrap the driver (above) since we have a cathode follower. However all distortion in the CF boots-back to the driver; not a great cheat. (BJTs make better "CF"s which is why many BJT audio power amps use bootstrapping effectively.)

> has many different HV secondaries

OK, so spitball a >1,250V few-mA supply. Driver plate must sit 1/3-1/2 of supply, 400V-600V.
Ignore the fact that few small tubes enjoy such voltages (we can always stuff a EL34).

With pentode driver, 1,300V supply, 650V on plate, driving AC coupled 6L6 with 355V supply, 6k load, I'm simming 4 Watts at 22% THD. 2 Watts at 8%THD. There's an unresolved kink, which may be a model flaw, though when I isolate the cathode follower alone in DC Sweep it's dead linear.
 

Attachments

  • Dinolobe-1.gif
    Dinolobe-1.gif
    10.8 KB · Views: 209
The obvious solution to that circuit has already been posted by Tubelab.

If one wants it simpler it is possible but RC coupling, CSS for the gain stage and independent voltage supply (if the one for the output stage is not enough) are necessary.

The Bates' circuit is just wrong. Forget it, throw it in the bin!!! You will only waste your time and money. If you are happy with less than 0.5W output power there are better and cheaper solutions....it's really not worth the trouble as is in the original article.
 
I guess i’ll have to find another ocatal/two valves circuit with 5k OPTs to fill my enclosure, wich has already been drilled. The other version of the circuit, the RC coupled one, was still tempting, but the power figures are so low for such an investment in time. 1250v is just wrong for such a low power output !!

Since my chassis is ready to receive this circuit topology, and because I can’t find a suitable two valve all octal single ended schematic for a 6v6gt, I think that I will assemble the rc coupled version of the bates amplifier. That will satisfy my curiosity and i’ll be able to measure it’s output power in real world.

The other option would be to tame myself and take the time to design something else that would fit the bill. I don’t have a computer that can run the tubecad programs and don’t know how to use LTspice so I would have to do it "by hand", unless you guys know of a circuit that I could use. Thanks to everyone who answered this thread, it’s been really instructive !
 

Attachments

  • A5996493-A439-4C98-82FB-E353D651E6A1.gif
    A5996493-A439-4C98-82FB-E353D651E6A1.gif
    11.2 KB · Views: 186
For me the bottom line of this cct max output would be One Watt on a good day. It is triode connected no matter how we look at it. 4.5W is the spec for a pentode connected 6V6 on a 250V supply. Better jack that up a few volts to account for the cathode bias resister & OPT primary R losses. So a supply of 275V might do it.

But never for the triode connexion.:)

Some of the ccts published in that era are not exactly as claimed. Need to be careful. I wuz there!! Few people had any kind of reliable test equipment. I was lucky, I worked in an R&D lab.:D
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.