The 26 is probably the most superb sounding tube in its class, with a tendency towards being microfonic. The generously measured filament current (1 A) to produce 6 mA anode current is one reason for its greatness. Shunt regulation is nice, maybe somewhat problematic for 180 V B+. In fact, tubes are happier with constant filament current than with constant filament voltage.
I have a shunt regulator working well on my 6S45P-E headphone amp with an output voltage of 150V B+ and 120mA idle current. Each tube draws about 25mA and the rest of current is readily available on tap for any draws. Search the Salas HV shunt regulator in this forum and you'll find examples running even at higher voltages/currents. Works very well.
For #26 it will be much easier to get a shunt regulator working, because of its lower biasing current.
This was recommended by JB as a headphone amp/buffer.
Single ended Mofset - big nastly electrolytic coupling cap on the output, push pull SS buffers without caps sound much better than the 470uf electrolytic coupling cap.
The only single ended class A buffer I have seen for headphone that sounds nice is the BJT sohaII http://www.cavalliaudio.com/soha%20ii/images/SOHAIIOneChannel.gif.
There is the skerzazea mofset buffer which is a single ended class A with three channels to allow DC coupling but the mofset ground stage has too high an output impedance for good bass response. Would love to come up with a single ended class A mofset with DC coupling to headphones but have had no such luck thus far.
regal,
you are right about electrolytics far from being ideal capacitors, but in practice it`s not that bad, the advantages of single-ended topology supersedes the disadvantages and I would not prefer that ugly IC. The power supply electrolytic capacitors are in series with the load as well. A high output impedance gives less firm bass response, but strongly improves linearity at higher frequencies.
Merry Christmas.
you are right about electrolytics far from being ideal capacitors, but in practice it`s not that bad, the advantages of single-ended topology supersedes the disadvantages and I would not prefer that ugly IC. The power supply electrolytic capacitors are in series with the load as well. A high output impedance gives less firm bass response, but strongly improves linearity at higher frequencies.
Merry Christmas.
My direct experience is completely different, a big electrolytic coupling cap at the output of a headamp overwelmingly dominates the sound signature of the amp. I have done much experiementation with this, headphones are very sensitive and just don't sound right to my ears with a big electrolytic coupling cap. But we all key in on different aspects of the sound so I am not saying you are wrong, just wondering if you have done any direct comparisons.
There's also euvl's (Patrick) source follower, and it's DC coupled. Have you considered it, regal?
http://www.diyaudio.com/forums/pass-labs/128571-some-other-source-follower-configurations.html#post1592820
I'm waiting for the power jfets to arrive, and will give it a try.
http://www.diyaudio.com/forums/pass-labs/128571-some-other-source-follower-configurations.html#post1592820
I'm waiting for the power jfets to arrive, and will give it a try.
Here you made an important and correct statement - not the output cap is a problem, but a bad big electrolytic.
For a headphone amp, put at the output 300uF Black Gate N + 30uF polypropylene shunt, satisfaction will be guaranteed.
Caps in the power rails could have lower quality, provided that a stage consumes constant current even with signal, i.e. it is loaded by a close to ideal current source. However, usually output stages are not like this, except, for example, the Power Follower 99. So, one should take care and spend a money.
Last edited:
All is correct, including robust PS. I use 1 kW trafo per channel in my 50W Zen9. Trafo acoustic noise is also an issue with class A.
I wonder to guys who still point out an output cap as a big problem.
Some advise regarding big electrolytics: if one attempts to use it in a signal path, including PS, substitute one big with many parallel small Low-ESR and Low-DA. And I also use big polypropylene shunts. Some parasitic wire inductance in this case is not an issue, but signal modulation due to dielectric absorbtion - really a problem.
Last edited:
The prototype (for measurements) will probably be just p-p. Final version I'll probably etch my own layout.
@vladimirk: regarding the psu I agree with you. But these days where can one get Black Gate caps at reasonable prices?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.