First foray into SET, design for your comment

[zigzagflux]

Hello fellow DIY'ers:

I have built a number of PP amps, EL34 and 300B primarily, along with a few similar design PP preamps. Few were copies of designs found here, few were my own design (but obviously nothing new).

Happy with what I have, but with the single-digit temperatures making me a hermit, the tube bug has hit me again. I have a shelf full of tubes, capacitors, and test equipment that has been calling my name.

Thought it would be time to try out a 300B SET, just for kicks and to be able to audition the differences between PP and SE here at home. I've heard some Cary amps that weren't too bad, but the construction quality was pathetic, and the noise floor was harsh (above 120Hz noise, real nastiness).

OPT not purchased yet, but will probably be Lundahl. Power transformer is One Electron (only shielded for me from now on). Tubes I would consider (not currently in use elsewhere) are 7119, 6922, 6SN7, 6W6. I have attached a schematic for consideration. My goal is to run 92dB speakers with no need for a rock concert.

A two stage amp is probably not where I would lean; I am a believer in having sufficient drive for the 300B, so I won't do a 6SL7 into a 300B. I have considered PowerDrive, but will leave that off for the second project if this first one works out. Those are the basic parameters.

Played around with trying to get at least one stage direct coupled, but found I just didn't have enough B+ available to get the operating points I wanted. Either that or I would have had to resort to a dual power supply configuration. Potentially possible with the One Electron transformer, but I don't throw huge capacitors right after my rectifiers. Tradeoffs, you know.

Thoughts ?

[kevinkr]

Hi Zigzag,
I can't quite make out what the second tube is in your design, (6N6?) but I would use resistive loading on the 6SN7 (27K or so) with conventional cathode bias and a CCS (or choke) on the second stage with fixed bias (-22V?) where linearity with big signal swings is what is needed.

[dsavitsk]

Kurt, Curious why the second stage wouldn't get the CCS/diode bias (zener?) treatment, too?

[Arnulf]

Originally Posted by kevinkr

Hi Zigzag,
I can't quite make out what the second tube is in your design, (6N6?) ...

Right click the image and select "Open in New Window" (or whatever this option is called in your web browser) to get to the original image.

It says 6W6, but this is probably a mistake as that is a beam tube, not a triode.

[kevinkr]

Originally Posted by Arnulf

Right click the image and select "Open in New Window" (or whatever this option is called in your web browser) to get to the original image.

It says 6W6, but this is probably a mistake as that is a beam tube, not a triode.

6W6 is fairly commonly triode connected for driver use, and I think fairly good in that connection. The 6Y6 is very popular as a driver triode connected in Japan..

[kevinkr]

Originally Posted by Arnulf

Right click the image and select "Open in New Window" (or whatever this option is called in your web browser) to get to the original image.

It says 6W6, but this is probably a mistake as that is a beam tube, not a triode.

That of course is what I did initially, but I still could not read it. (I use FF in Linux) If I squint hard enough I can almost tell that is a "W."

[andyjevans]

If you want to go all-DHT you will get superb delicacy and transparency at the expense of a larger PSU to encompass the filament supplies - you'll need 6 in all. And easiest solution here is a seperate PSU box with all the transformers and chokes in it. So only DC in the signal chassis.

To keep it cheap and simple use DHTs with low filament current, but nevertheless enough to keep them out of the range of microphony. As I write I'm listening to a 300b SET with 30sp (VT-67) into 31 driver. Cheap and just a lovely sound in all ways. To take it up a notch go for 01a into 71a - but that's more expensive. In any case, in my opinion having tried all this out, all these all-DHT options will put you way ahead of conventional tubes - 6SN7 or whatever. There's a "26 pre amp" thread recently which will give you just about all the information you need for the first two stages. In fact a 26 is great as an input tube, and even marginally usable as a driver.

If you listen to classical music a lot or acoustic jazz and vocals, you'll notice the difference in the timbre of acoustic instruments, which to my ears is significantly more lifelike with DHTs. With rock and electronic music I can't really comment since that's not my thing.

andy

[zigzagflux]

Thanks guys for the quick responses.

First, Andy: I am with you, DHT territory might be the way to go. However, I already have a PP amp where 46's drive 300B's, so in some sense, been there done that. Lots of interstage transformers (2 per monoblock plus Tribute input splitter). If I like SET (jury is still out), I may make a second amp with DHT's in the input sections, go for the gusto.

The idea here is to use the tubes I have, which is very little selection, I admit.

Kevin, better schematic attached.

Kurt, Curious why the second stage wouldn't get the CCS/diode bias (zener?) treatment, too?

Wow, what an obvious option ! Might even give me the ability to direct couple. I'll play around with this; anyone used the mu output of the DN2540 cascode with good results? I see Gary Pimm uses it often.

[Eli Duttman]

ZZF,

Your will not get the full μ of 20 from the 1st gain block. The 221 KOhm 6W6 grid leak resistor is very much part of the 'SN7 triode's NET AC load. You will come close, but less than 20X is the result.

[zigzagflux]

Originally Posted by Eli Duttman

Your will not get the full μ of 20 from the 1st gain block.

Agreed, I'm just sketching out a gameplan at this point. I'll test each individual stage for performance and distortion at the o/p. In the past, I've just built my amps after straining over the design. This time, I'll build on a large piece of plywood; if I don't care for what I hear, I may just disassemble.

Next version: I think this might have some potential. $10 in LED's per channel. I read from Gary Pimm's site the mu output has about a 500 ohm output impedance. Looked like DC coupling stages still wasn't very obtainable with one power supply; nbd.

Use of a 5.1V zener diode looks like a good choice, too; impedance around 7 ohms is comparable to the LED string. Anyone compared the two? 5.1 to 6.2V zeners are somewhat quiet with good TC.