[regal]

Has anyone built P. Millett's 307A P-P headphone amp ? I would like to start a thread for it. Assuming the 307A has an Rp=800 I'm calculating an output impedance of around 3.2 ohms, which is the best I have seen of any pure tube headphone amp. Would give Grados a damping factor of 10 vs the typical 2 for even the most expensive tube amps (excluding hybrids.) The 307A is a predecessor to the 300B more suitable to headphones but should have the good DHT sound. The DC heaters shouldn't be to bad as they only need 1amp per tube. The transformers used look expensive but I bet Electra-Print could match them for a decent price.


307A Push-Pull Amp

I noticed on his commercial version (TTVJ) he dropped the input tube and went straight to the phase splitting transformer, probably so it would be easier to have balanced inputs, all the transformers used on the commercial amp are also different than the DIY.


I am not sure about the input tube buffer+transformer choices, it would seem we could find a tube with lower gain and match it with a different transformer and get the same or better results without the NFB. I get the feeling he picked the 7W7 for the input buffer since it was the same tube model being used for the driver.

 

[tomb]

Regal,

Have you talked to Pete about this? I thought the 307A Push-Pull was a speaker amp (tubes running at 10W and 8ohm output) and purposely so, IMHO. Again, just MHO because I haven't talked to him about it either - but he probably wouldn't want DIY "competition" for the TTVJ 307A headphone amp that he actively sells commercially.

 

[regal]

I'll check with him but his first page on the DIY website says its a headphone amp.

 

[n_maher]

Quote:

Originally Posted by regal /img/forum/go_quote.gif I'll check with him but his first page on the DIY website says its a headphone amp.

I don't think he does. He references the TTVJ amp at the beginning of the article as a headphone amp but then goes on to talk about exploring the possibilities of using the 307A tube in SE and PP power amp circuits. I believe that the design posted, as Tom mentioned, is geared towards speakers and not headphones.

That said, there may be a way to adapt it for headphone use with limited modification but really it's quite different than the TTVJ 307A amp and shouldn't be confused or even closely associated with it.

 

[nikongod]

Are you planning to use transformers designed for an 8ohm load with loading resistors in parallel with the headphones? It may be the only way to get the output impedance you require.

If you do go for something with a tap designed for a 32 ohm load you will be lucky to get even a ~3ohm DCR on the secondary. Zo>6ohms. Sorry, the damping factor may not meet the standards you have set. I bet a little global feedback will bring it back to an acceptable value.

I would recommend asking Jack what the secondary DCR of a transformer he has wound dozens of in the past will come out as

He likes that.

Quote:

I noticed on his commercial version (TTVJ) he dropped the input tube and went straight to the phase splitting transformer, probably so it would be easier to have balanced inputs, all the transformers used on the commercial amp are also different than the DIY.

I see that as being balanced to only one grid. AKA a phase combining transformer. Good stuff for galvanic isolation into a single ended amp, its a shame more people dont do it.

 

[luvdunhill]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif I would recommend asking Jack what the secondary DCR of a transformer he has wound dozens of in the past will come out as He likes that

 

[the_equalizer]

Quote:

Originally Posted by n_maher /img/forum/go_quote.gif I don't think he does. He references the TTVJ amp at the beginning of the article as a headphone amp but then goes on to talk about exploring the possibilities of using the 307A tube in SE and PP power amp circuits. I believe that the design posted, as Tom mentioned, is geared towards speakers and not headphones. <snip>

It is indeed geared towards speakers, as it's easy to see from it being listed in the 'Power Amps' section of Pete's website.

 

[regal]

I'll check to see if he is OK with us making this into a DIY headamp.

 

[dsavitsk]

Quote:

Originally Posted by regal /img/forum/go_quote.gif Assuming the 307A has an Rp=800

You might as well assume it has 4 legs and a tail, as that wouldn't be right, either.

Rather than trying to ride on Pete's coattails, why not just design your own 307a amp?

Quote:

Would give Grados a damping factor of 10 vs the typical 2 for even the most expensive tube amps (excluding hybrids.)

Damping factor turns out to be worth playing with a lot with Grados -- my recent experience says it may not be as important as we all thought it might be. I built, for instance, a pentode based transconductance amp with a Zout of ~300 ohms that is turning out to be one of my favorite Grado amps I've heard. Other recent builds include one where the transformer's secondary DCR is ~40 ohms, and it works great.

Quote:

Originally Posted by nikongod If you do go for something with a tap designed for a 32 ohm load you will be lucky to get even a ~3ohm DCR on the secondary. Zo>6ohms. Sorry, the damping factor may not meet the standards you have set.

Indeed, higher Zout due to secondary DCR seems, to my ear, to be a lot more benign than high Zout from other sources -- my best guess is that I just don't like 6080 cathode followers and attributed their badness to their impedance, when in reality it is because it is a sucky tube.

Quote:

Originally Posted by nikongod I bet a little global feedback will bring it back to an acceptable value.

Heathen.

 

[nikongod]

Quote:

Originally Posted by dsavitsk /img/forum/go_quote.gif Damping factor turns out to be worth playing with a lot with Grados -- my recent experience says it may not be as important as we all thought it might be. I built, for instance, a pentode based transconductance amp with a Zout of ~300 ohms that is turning out to be one of my favorite Grado amps I've heard.

Do you have a schematic or a page written up for this? I quite enjoy reading your articles.
Quote:

Originally Posted by dsavitsk Indeed, higher Zout due to secondary DCR seems, to my ear, to be a lot more benign than high Zout from other sources -- my best guess is that I just don't like 6080 cathode followers and attributed their badness to their impedance, when in reality it is because it is a sucky tube.

I nearly spat my beverage out when I read that. Good stuff.

6080 cathode followers distort hard when driving low impedance loads. That cant possibly help the situation at all.
Quote:

Originally Posted by dsavitsk Quote: Originally Posted by nikongod /img/forum/go_quote.gif I bet a little global feedback will bring it (output impedance) back to an acceptable value. Heathen.

Could I possibly redeem myself by pointing out that this global feedback loop would also reduce THD?

 

[n_maher]

Quote:

Originally Posted by dsavitsk /img/forum/go_quote.gif Rather than trying to ride on Pete's coattails, why not just design your own 307a amp?

I think I should be offended.

Quote:

Damping factor turns out to be worth playing with a lot with Grados -- my recent experience says it may not be as important as we all thought it might be.

My experience says that it's model dependent. The HF-1s were much more tolerating of highish (~30ohm) output impedances than the HF-2s were. The HF-1s actually liked the lower damping factor, the HF-2s became bloated and didn't sound right at all. They both sound quite good out of a trafo-coupled amp though which seems to be in line with what you're saying.

Quote:

Indeed, higher Zout due to secondary DCR seems, to my ear, to be a lot more benign than high Zout from other sources -- my best guess is that I just don't like 6080 cathode followers and attributed their badness to their impedance, when in reality it is because it is a sucky tube.

Ok, now I'm officially offended.

We'll see if I let you listen to my next sucky amp.

 

[dsavitsk]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif Do you have a schematic or a page written up for this? I quite enjoy reading your articles.

Nothing written up just yet as it is still a work in progress. I did take it to the ChiUniFi meet -- here's a picture.

and a schematic

I used the same case as my SSS/Shining Path amp as that one had worn out it's welcome and got disassembled.

But, basically it is a 12GN7 hooked up as a pentode, transformer coupled, with no feedback. The glow tubes act as regulators for the screen grid. It started as me trying to figure out how to apply feedback to a pentode spud with a pot on the input (here's the thread) -- turns out it isn't quite obvious. I ended up connecting the base of Ck through the OPT's secondary winding which is 100% negative feedback leaves you with a unity gain tube stage. It is a neat trick, but it sounded like a solid state buffer with 100% feedback. On a lark I just disconnected the feedback to see what would happen, and it was shockingly good.

The lower damping factor being good for Grados thing is not entirely surprising -- my fostex based speakers sound better off the 8 ohm tap than the 4 ohm on my 300B amps, and they sound truly awful from a Gainclone.


Quote:

Could I possibly redeem myself by pointing out that this global feedback loop would also reduce THD?

Have you read the Pass article on feedback and distortion spectrum? It is one of his better articles I think.

 

[pabbi1]

Quote:

Originally Posted by n_maher /img/forum/go_quote.gif

I has a sad...

Oh, and if you guys are really serious about this, I was about 90% parts complete in this amp, but realized it is just way the hell beyond me. But, if you want a damn good start towards putting one together, I have the interstage trafos, rectifiers (and 7w7, all with sockets), James 9611 iron, and other passive parts all acquired - lots listed in the FS forum.

Sold my <ahem> 307a, but, really just choked on the DC heaters, where I just couldn't swallow the Tent lab prices. This is NOT for the feint of heart, nor shallow pockets. Notice that Nate is the only other person to tread down this dark path.

Having heard the TTVJ 307a, there is gold at the end of that rainbow, but a steep price paid getting there.

 

[regal]

I don't know where but I thought I read that the 307A had the same Rp of a 300B, dsavitsk I apologize. I can't seem to find a datasheet for this tube.

I guess based on my experiments with Grados (MS1's, HF2's, and MS-Pros) I found a strong correlation to low output impedance to less congestion. For example the 5842 SET OPT amp doesn't compare favorably to my HV-properly biased Tube + SS buffer hybrid amp that has an output impedance of 2 (with zero NFB). It isn't so noticeable until a complex passage hits and to me it is obvious that the driver is out of control. Escpecially a SET amp damping plays a critical role because the damping factor changes as the tube output voltage goes up/down with the music. For what its worth , thats been my experience. I think Grados are actually one of the toughest headphones to design a tube amp around.



So I guess I got overly enthused when Pappi mentioned this amp, thinking it could be decent damping factor Grado tube amp with no NFB.

 

[regal]

Quote:

Originally Posted by pabbi1 /img/forum/go_quote.gif I has a sad... Sold my <ahem> 307a, but, really just choked on the DC heaters, where I just couldn't swallow the Tent lab prices.

Also Pappi, check out this site: LM317 / LM338 / LM350 Voltage and Current Regulator Calculators

The current source calculator is a good start for designing DHT heaters. After the rectifier you will need 10,000 uF cap and enough voltage headroom, plus a case mount heatsink. May need to preregulate. But this may come close to the expensive commercial alternative, at least it has been reported so for preamps. Worth a try before you throw in the towel as it would be inexpensive the put one together to try.