Winter is coming... Time to start GM70 SE parafeed amp

[djn]

I am certainly looking forward to getting mine up and running. I had to step back from all my projects while I started a business but that is going well so I don't have to spend every waking hour at work.

My 12GN7 will be cap coupled to the 70 using a big *** 1.5uf PIO cap. WOW, I am having a hard time remembering exactly what the design is. I know I am sending the B++ to the 12GN7 through a transformer but I can't remember the specs.

 

[digitrax]

I hope to drag my GM70 project out this winter and finish it. All it needs is to be wired. I am going two stage using the 12GN7 as the driver tube. I "think" it will have enough swing to drive the 70.

12GN7 is certainly capable of enough swing. You won't be disappointed in the GM-70; I've built SETs with several DHTs (as well as a few tetrodes and idht's) and I think it's about the best audio output tube there is, bar none.

Also, I think you picked a good time for your build. Judging by the drop in temperature over the last few days, I think it'll be a much colder winter than last year. – (I'm in Ann Arbor, not far from your neck of the woods).

 

[djn]

I agree Digitrax. For the most part, I don't spend time in front of the system during the warmer months, so heating the room with 120 watts of GM70 glow is fine by me.

I literally only have to wire it up. I have all the parts, I've assembled the Rod Coleman CCS GM70 heater boards, and the chassis is all prettied up. Now that I've jumped back into this post, the urge is getting stronger.

 

[kevinkr]

My GM70s combined filament and plate dissipation is 185W, total power consumption per channel is 277W - all of which practically speaking ends up as heat..
Surprisingly in the dead of winter it barely makes a dent in this cold basement, in summer some assistance from an air conditioner is needed periodically during listening sessions.

 

[yagoolar]

Understood.
So, 2 or 3 stage? (and do we care about being able to get to A2?)
Also, what kind of input sensitivity do you need?

Other candidates you might consider:
- 6Н6П
- 6Н30П/6Н30П-ДР
- D3a
- 6AU5
- 6SN7/12SN7
- 6Э5П triode
- 826

I forgot about 4P1L DHT pentodes I had bought! All DHT GM70 amp!

 

[digitrax]

I forgot about 4P1L DHT pentodes I had bought! All DHT GM70 amp!

There ya go!
Aleks Belkanov (R.I.P.) was a big advocate of using triode-strapped 4П1Л – here are his triode curves:
4П1Л triode.JPG

 

[mogliaa]

Both 6e5p and 4P1L are superb drivers. Here is my improved version of the 4P1L triode curves and model including A2 grid current:
4P1L model improved | Bartola Valves

I've done the crazy thing of using a 46 in filament bias DC coupled to drive a 4-65a. Outstanding results though however it's a massive and heavy amplifier

Cheers,
Ale

 

[yagoolar]

Ale,

Great news a 4P1L A2 can deliver 5W output power!
Slightly OT - do you think 4P1L PSE is worth considering?

Thanks
Pawel

 

[mogliaa]

Hey Pawel,
With Andy Evans we have done side by side listening tests to 300B versus 4P1L PSE and have to admit I preferred 4P1L. Not same amps though but either way given is so easy to match 4P1L pairs in general, it's worth considering given price difference. 4P1L is less demanding in drive and also can easily work in filament bias. No cathode cap so happy days.

Back to my 46 driver in filament bias is probably the best I've achieved so far, but is highly demanding in terms of power, so would probably suggest 4P1L (both pentode or triode) or 6e5p instead...
Looking forward to see your progress on this project. I do have some GM70s around but with my 4-65a SE amp I'm done with transmitter valve amps for the time being...
Cheers,
Ale

 

[yagoolar]

If I used one 4P1L to amplify and drive GM70, I would have to wire one in pentode mode.

 

[digitrax]

If I used one 4P1L to amplify and drive GM70, I would have to wire one in pentode mode.

Quite possibly. I'm going to have to refer back to the questions I asked a couple pages ago:

...do we care about being able to get to A2?
...what kind of input sensitivity do you need?

It'll be a lot easier to establish what will (and will not) work once we know the answers to those.

 

[yagoolar]

Q1. A1 is enough for a start, so - no, I don't.
Q2. Input sens. 2V rms

 

[kevinkr]

Before proceeding down the pentode path you might PM Wavebourn for his opinion on pentode operation with this tube.

I have a quartet for an eventual SE project, but my understanding is that the great performance is in triode and that it does not necessarily translate to pentode mode.

 

[digitrax]

Before proceeding down the pentode path you might PM Wavebourn for his opinion on pentode operation with this tube.

I have a quartet for an eventual SE project, but my understanding is that the great performance is in triode and that it does not necessarily translate to pentode mode.

Indeed, the 4П1Л is sometimes referred to as "the last DHT" primarily because of it's triode performance – (it's actually more linear than 2A3) – and the fact that it's directly heated. If you want to go pentode, you'd be just as well (or better) off with 12GN7, which can provide 240Vpk-pk without breaking a sweat.

To stay triode in a 2-stage, you could use 6С45П, but you'd have to "hand select" one – (its µ is typically around 40-50, but if you pick through a few, you can find instances of µ>60) – and run it a little beyond it's max. plate voltage if you want "full drive" of the GM-70 from 2V.
I have seen this done with good-sounding results, but can't vouch for the tube's life under such circumstances.

 

[kevinkr]

The D3A/7721 in triode and the E810F/7788 also in triode would be worth considering.
Perhaps the 6J52P which supposedly similar to the D3A and E810F might be worth looking at, they're cheap. I have some that I will eventually evaluate.

Here is some information on other Russian pentodes and their performance triode connected: http://www.bartola.co.uk/valves/2013/02/23/russian-pentodes-in-triode-mode/

 

[kevinkr]

Here is a Russian language data sheet for the 6J52P which I have converted from djvu to pdf.. The original file can be found at www.rutubes.com. This tube is considered in some quarters as the Russian equivalent of the D3A. (Apparently good ones sound quite good based on the comments I have found)

 

[yagoolar]

I was taught Russian in 70s and 80s at school. Some of it is still left in my memory. No problem reading it. Understanding, well...
I know Wavebourn's threads and posts, they are valuable to me.

I think 6j52p (6ż52 in pl.) is very cheap Lampa 6?52P - TEST - sprawna (3651290225) - Allegro.pl - Wi?cej ni? aukcje.

 

[digitrax]

The D3A/7721 in triode and the E810F/7788 also in triode would be worth considering.

D3a would work; E180F is a little slim (gain-wise) to get full output from a GM-70; E280F probably could, though.
Pentodes are tricky beasts. The ones I've particularly liked as triodes were EF86 (which won't give you enough voltage swing), 4П1Л and 6Э5П (which won't give you enough gain). There are, of course, a great many I haven't tried.

Perhaps the 6J52P which supposedly similar to the D3A and E810F might be worth looking at, they're cheap. I have some that I will eventually evaluate.

I've heard good things about that tube in triode; not sure about gain though – (although I know 6Ж51П has plenty).

Perhaps Ale could chime in here?

 

[kevinkr]

Yes, it is inexpensive and has also been used to drive GM70 with apparently above average results: 3x Monica on a Triode festival.

I believe this design has three stages based on the pix. The guy behind this amp also has a two stage design, but the links are expired.

 

[kevinkr]

I just ran a curve trace on the 6J52P, and while it's linearity is good I don't think it is a good choice to drive the GM70 directly. The curves will tell the story.

I attempted to trace it pentode connected, but did not find a good combination of fixed screen voltage and grid bias in the time I had available. (It would have helped to look at the data sheet.)

Yagoolar perhaps you could provide a partial translation of the data sheet? There are some parts I don't follow.