I think the SRPP has become something of a tube audio design fashion statement, and possibly an irrational one at that.. I used them a lot in some of my older designs, and perhaps (definitely) a bit too widely. To my chagrin I found some real practical downsides to its use including excessively high output impedances with low transconductance high mu triodes like the 12AX7A/ECC83. (Much worse incidentally that a straight CC stage driving a CF in this case.)
About 11 yrs ago I designed a phono stage that used one in the output stage and lost the possibility of a very lucrative distribution agreement because it could not effectively drive the capacitance of the most popular interconnects of the day. It bombed in a critical 3rd party system evaluation for that reason, despite my verbal warning about its inability to drive very capacitive cables.
There are other shortcomings including linearity and noise where it is generally inferior to the mu follower, and a CCS (or resistively) loaded common cathode stage.
It is quite good at driving the specific load impedance for which it was designed and for delivering some appreciable amount of power into a load. It is not a great choice where the primary goal is voltage gain, and the wrong choice in tube type will result in a stage that can't drive even reasonable loads. (Usually the result of needing voltage gain)
I still admit to liking the way the 6SN7 sounds in SRPP configuration, and use it that way in one of my old line stages, I also use it to drive the output stage in my SE amplifier where it works pretty hard, and gives a good account of itself. Certain tubes like the 6021 IMHO sound dreadful and produce bad measurements when used in SRPP. I haven't designed anything using this topology since about 1999..
How about a 5842 with a CCS load, and LED bias, output impedance will be <2K, (or you could configure as a mu-follower using a DN2540 mosfet) gain is high enough that a very small I/V resistor could be used for good dac linearity, and it is extremely low noise.
About 11 yrs ago I designed a phono stage that used one in the output stage and lost the possibility of a very lucrative distribution agreement because it could not effectively drive the capacitance of the most popular interconnects of the day. It bombed in a critical 3rd party system evaluation for that reason, despite my verbal warning about its inability to drive very capacitive cables.
There are other shortcomings including linearity and noise where it is generally inferior to the mu follower, and a CCS (or resistively) loaded common cathode stage.
It is quite good at driving the specific load impedance for which it was designed and for delivering some appreciable amount of power into a load. It is not a great choice where the primary goal is voltage gain, and the wrong choice in tube type will result in a stage that can't drive even reasonable loads. (Usually the result of needing voltage gain)
I still admit to liking the way the 6SN7 sounds in SRPP configuration, and use it that way in one of my old line stages, I also use it to drive the output stage in my SE amplifier where it works pretty hard, and gives a good account of itself. Certain tubes like the 6021 IMHO sound dreadful and produce bad measurements when used in SRPP. I haven't designed anything using this topology since about 1999..
How about a 5842 with a CCS load, and LED bias, output impedance will be <2K, (or you could configure as a mu-follower using a DN2540 mosfet) gain is high enough that a very small I/V resistor could be used for good dac linearity, and it is extremely low noise.
Good suggestion - a 7119 with a 1:1 transformer is my chosen solution...for now at least
FWIW: Our SRPP experiences are very similar, Kevin.
Jim
Myself:
SY:
kevinkr:
I would go for a 6C45 with anodechoke, that is close to the above mentioned solution and if I am not mistaken this has been used years ago by some guy called Lesha.
Also why a PCB when these circuits are simple enough to hardwire on the tubesocket?
No cap necessary at the input. But the problem is that you will get at least100Vrms out(if the tube wouldn´t clip before) with a typical V-DAC.
SY:
kevinkr:
I would go for a 6C45 with anodechoke, that is close to the above mentioned solution and if I am not mistaken this has been used years ago by some guy called Lesha.
Also why a PCB when these circuits are simple enough to hardwire on the tubesocket?
No cap necessary at the input. But the problem is that you will get at least100Vrms out(if the tube wouldn´t clip before) with a typical V-DAC
.
Good suggestion - a 7119 with a 1:1 transformer is my chosen solution...for now at least
FWIW: Our SRPP experiences are very similar, Kevin.
Jim
Hi Jim,
I'm using transformers in my dac and sacd player as well, post I/V conversion (all currently SS) to convert the ground referenced differential outputs to floating or grounded unbalanced or balanced as needed. They also provide some additional LPF filtering beyond what the I/V stage themselves are designed to do. (I did several conventional tube based cd and dacs with current output dacs, was never completely satisfied with their performance. No high transconductance tubes or transformers in these designs.)
IMO the SRPP became very popular in Japan and that popularity spread rapidly to this side of the pond sometime in the late 1990s. It then became the universal panacea for those wishing (irrationally?) to avoid the use of the much maligned CF. Most of us had not yet discovered high transconductance triodes or pentodes, probably the one of the more ideal solutions in many applications requiring some voltage gain and low output impedances - they were rarely if ever used in Hi-Fi with the notable exception of Stuart Hegemon's Citation II and Citation V amplifier designs.
But note that you are making the same basic mistake as the OP and that SY mentions. A CCS load presents a constant and very high impedance to the tube. A choke load presents a not as high load that varies with frequency (Z=2piFL) leading to distortion higher distortion.
The choke makes sense in positions where you need a large signal swing (above B+), such as on a driver tube or on a parafeed output stage. In those cases, it makes sense to trade linearity for swing. But here, where the swing is quite modest and where linearity is most important, the CCS is the better solution.
As SY might say, if your goal is a tone control of some sort, then by all means use the choke if you think it sounds better. But if your goal is a linear amplification stage, use a CCS.
Hello
I saw this old post and if it could help I would like to share experiences with TDA1541 and 6021 submini tubes.
I'm not a big fanatic of Philips DAC, I prefer AD1865 in NOS or PCM in standard. But I tried a TDA1541S1 in NOS just to have a TDA in my collection.
For decoupling caps, I obtain the best with 1µF Panasonic ECQ-P (mkp) on each pin with 0,47µF old mkp on 2 last pins on each side.
For I/V resistor, I saw everywhere that 32 ohms is the best, I tried with a 32 ohms riken . The sound is good, but with not a lot of dynamic and you need a buffer with a lot of gain. So I added a 22 ohms riken resistor to have 54 ohms. The improvment to my ear is clear, more dynamic with the same detail than with 32 ohms. No audible distortion arrived with this added resistor (don't forget that an ear is not an oscilloscope). The more difference to my ear with dynamic difference, is that with 54 ohms, treble sound less metalic and we have more presence.
For 6021 submini. In a first time I tried grounded cathode 5703WB, 6021WA, 6DJ8, 6n23p. After that I tried 6DJ8 in totem pole, and at the end 6021WA totem pole (SRPP) with 160V, Ra=Rk= 200 ohms, 220µF cathode cap, 1,5µF output cap and 100K load resistor at output.
I can say that the sound is very very good, one of the best to my ear. And no, 6021WA in totem pole (SRPP) is not a bad solution.
With TDA1541 like 6dj8 in audio note ad1865 dac, the SRPP give a sound "analog like". In fact the sound is more sweet, more like an LP. I think that audio note used totem pole in the AD1865 DAC for these reason, it's not for a technical reason, but just because with a totem pole output stage the sound is more "analog". When you create something in audio, you don't create the perfect technical object, but something that you will listen, so your taste or the taste of the final listener is in what you create.
At the end, after a lot of tests, I stayed with 54 ohms and a 6021WA totem pole (SRPP). I play piano and guitar since 25 years, and I stayed with this config because to my ear, piano and guitar sound the more natural with the better spatialization. One of my favorite test is to test with "Wild Horses" and "Love In Vain" from Rolling Stones "Stripped" album. The guitar of Keith sound really like a Martin. I made other tests with "koln concert" of keith jarrett to verify that the piano sound like a piano. End at the end I verified with some symphonic records that the sound is not a melting pot.
So the only thing I can say to resume, is, don't hesitate to experiment your ideas, sometimes the result is marvelous, sometimes the result is a s#$%,.....human ears and brains are not oscilloscopes.
Cordially
Pascal
I saw this old post and if it could help I would like to share experiences with TDA1541 and 6021 submini tubes.
I'm not a big fanatic of Philips DAC, I prefer AD1865 in NOS or PCM in standard. But I tried a TDA1541S1 in NOS just to have a TDA in my collection.
For decoupling caps, I obtain the best with 1µF Panasonic ECQ-P (mkp) on each pin with 0,47µF old mkp on 2 last pins on each side.
For I/V resistor, I saw everywhere that 32 ohms is the best, I tried with a 32 ohms riken . The sound is good, but with not a lot of dynamic and you need a buffer with a lot of gain. So I added a 22 ohms riken resistor to have 54 ohms. The improvment to my ear is clear, more dynamic with the same detail than with 32 ohms. No audible distortion arrived with this added resistor (don't forget that an ear is not an oscilloscope). The more difference to my ear with dynamic difference, is that with 54 ohms, treble sound less metalic and we have more presence.
For 6021 submini. In a first time I tried grounded cathode 5703WB, 6021WA, 6DJ8, 6n23p. After that I tried 6DJ8 in totem pole, and at the end 6021WA totem pole (SRPP) with 160V, Ra=Rk= 200 ohms, 220µF cathode cap, 1,5µF output cap and 100K load resistor at output.
I can say that the sound is very very good, one of the best to my ear. And no, 6021WA in totem pole (SRPP) is not a bad solution.
With TDA1541 like 6dj8 in audio note ad1865 dac, the SRPP give a sound "analog like". In fact the sound is more sweet, more like an LP. I think that audio note used totem pole in the AD1865 DAC for these reason, it's not for a technical reason, but just because with a totem pole output stage the sound is more "analog". When you create something in audio, you don't create the perfect technical object, but something that you will listen, so your taste or the taste of the final listener is in what you create.
At the end, after a lot of tests, I stayed with 54 ohms and a 6021WA totem pole (SRPP). I play piano and guitar since 25 years, and I stayed with this config because to my ear, piano and guitar sound the more natural with the better spatialization. One of my favorite test is to test with "Wild Horses" and "Love In Vain" from Rolling Stones "Stripped" album. The guitar of Keith sound really like a Martin. I made other tests with "koln concert" of keith jarrett to verify that the piano sound like a piano. End at the end I verified with some symphonic records that the sound is not a melting pot.
So the only thing I can say to resume, is, don't hesitate to experiment your ideas, sometimes the result is marvelous, sometimes the result is a s#$%,.....human ears and brains are not oscilloscopes.
Cordially
Pascal
Last edited:
Hi Pascal,
Really informative, thanks for sharing your experiences. I have some questions, just want to know what your ears "tell" you:
Have you tried 6DJ8 (E88CC)? How could you describe the sound re. the 6021WA? You can perhaps raise the V+ to 250...300V. Also increase the Rk up to 1.5K in case of 6DJ8.
Have you tried the totem pole circuit omitting the shunt capacitor at Rk? Gain will be less (about 16 so I use a transformer in front), but you'll get better linearity in return. Also the type, value, make etc. of the cathode cap will not play any role.
What power supply do you use? SRPP has poor PSRR so it is sensitive in this respect.
Thanks,
Laszlo
Really informative, thanks for sharing your experiences. I have some questions, just want to know what your ears "tell" you:
Have you tried 6DJ8 (E88CC)? How could you describe the sound re. the 6021WA? You can perhaps raise the V+ to 250...300V. Also increase the Rk up to 1.5K in case of 6DJ8.
Have you tried the totem pole circuit omitting the shunt capacitor at Rk? Gain will be less (about 16 so I use a transformer in front), but you'll get better linearity in return. Also the type, value, make etc. of the cathode cap will not play any role.
What power supply do you use? SRPP has poor PSRR so it is sensitive in this respect.
Thanks,
Laszlo
Hi Laszlo
in the past, my method was to create with a real respect of electronic theory, and after conception and realisation I listened.
But I noticed some things that shock me. Some projects had splendid curves on an oscilloscope, but with listening, were *****.
Sometimes you spend hours to modify and to complicate PSU and design, and at the end, when you listen in blind test, you hear no change, no improvment; or unfortunately sometimes a modification of the sound which is not liked at all.
So now, I tried to make a link between taste and technical theory. I try to know that if I make this way, that will sound this way.
I if i have an idea, I tried, and after I listen and analyse how it sound.
I tried 6DJ8 (and my favorite 6n23p) in different configurations. I like it for DAC like CS4397 (you can see how on my site, sorry it's in french, but my french it's better than my english CS4397 DAC ) Of course with more voltage (from memory I tried 6DJ8 in SRPP with 220 or 250V
If I compare 6DJ8 to 6021WA, the submini 6021WA sound more like an octal tube, more or less like a 6SN7.
I have not tried yet, but I have 6K4 which seems to be more a submini 6SN7 than 6021WA.
I will try these 6K4 in a future DAC project. This project is a DAC with a switch which can permit to switch between two cards, one with two PCM1794 and one with two WM8740 (PCM and WM are in mono mode, one per channel). But with one shared input and one shared output.
6DJ8 is a real good tube, but i'm not convince that it is the better for a totem pole. It sound smooth to my ears in totem pole, with a lot of clarity and detail, but sometimes too much slicing.
A 6SN7 is more groovy, with the same detail and dynamic but with more texture. It's difficult to find good words to describe, but if you have ever heard Gibson and Martin acoustic guitars, 6DJ8 is like Gibson and 6SN7 like Martin, more "charnel" in french, i think "carnal" in english ?
Or an other comparison to compare 6DJ8 and 6SN7/6021WA in totem pole, could be that 6DJ8 is Beatles, and 6SN7/6021WA is Rolling Stones
For cathode cap, I use only Elna Silmic II, Panasonic FC or Nichicon Muse ES with MKP or paper oil decoupling cap. I prefer with cap because, without, I never find the same quality in bass impact and groove. It's typical with violoncello of "Suites pour Violoncelle seul de J.S. Bach" or on "Dark Side of The Moon" (of course this album is better in LP, but it could be used to test DAC).
Power supply is a real good debate. I tried a lot of PSU types. I noticed (or perhaps it's just a question of taste), that the more sophisticated is the PSU, the less you have dynamic and live in the music, and more you have compressed sound.
My favorite PSU for tubes is a simple tube rectifier (sometimes silicon if I have no choice), a good MKP or paper oil capacitor, a good choke, a medium capacitor, a resistor or an other choke and a big capacitor. Sometimes I stabilise with an IRF830 like on the CS4397, if I can do without, I prefer.
But I use a Schaffner filter on primary of the transformer, and very good transformers with electrostatic screen linked to ground.
It's the same for DAC PSU, a lot of people use complex PSU, and I tried these sort of PSU, I never heard a difference with a good LM350 or LM317 PSU. For me the real good solution is to use quality choke between 2 big quality capacitors before LM, and to use a choke between the LM supply last capacitor and the last capacitor at the chip.
I compared this sort of PSU with friends who have made same projects with complicated and sophisticated PSU, and in blind test, no audible difference, just that in lot of cases my PSU gave more dynamic.
I test all what I create with the same system, my 2A3 SE amplifier with Klipsch 98db speakers and my ecc99 totem pole preamplifier (or sometimes other amplifiers). I can put my ears on klipsch drivers and horns and I hear nothing without signal, it's enough for me ;-)
Cordially
Pascal
in the past, my method was to create with a real respect of electronic theory, and after conception and realisation I listened.
But I noticed some things that shock me. Some projects had splendid curves on an oscilloscope, but with listening, were *****.
Sometimes you spend hours to modify and to complicate PSU and design, and at the end, when you listen in blind test, you hear no change, no improvment; or unfortunately sometimes a modification of the sound which is not liked at all.
So now, I tried to make a link between taste and technical theory. I try to know that if I make this way, that will sound this way.
I if i have an idea, I tried, and after I listen and analyse how it sound.
I tried 6DJ8 (and my favorite 6n23p) in different configurations. I like it for DAC like CS4397 (you can see how on my site, sorry it's in french, but my french it's better than my english
CS4397 DAC ) Of course with more voltage (from memory I tried 6DJ8 in SRPP with 220 or 250VIf I compare 6DJ8 to 6021WA, the submini 6021WA sound more like an octal tube, more or less like a 6SN7.
I have not tried yet, but I have 6K4 which seems to be more a submini 6SN7 than 6021WA.
I will try these 6K4 in a future DAC project. This project is a DAC with a switch which can permit to switch between two cards, one with two PCM1794 and one with two WM8740 (PCM and WM are in mono mode, one per channel). But with one shared input and one shared output.
6DJ8 is a real good tube, but i'm not convince that it is the better for a totem pole. It sound smooth to my ears in totem pole, with a lot of clarity and detail, but sometimes too much slicing.
A 6SN7 is more groovy, with the same detail and dynamic but with more texture. It's difficult to find good words to describe, but if you have ever heard Gibson and Martin acoustic guitars, 6DJ8 is like Gibson and 6SN7 like Martin, more "charnel" in french, i think "carnal" in english ?
Or an other comparison to compare 6DJ8 and 6SN7/6021WA in totem pole, could be that 6DJ8 is Beatles, and 6SN7/6021WA is Rolling Stones
For cathode cap, I use only Elna Silmic II, Panasonic FC or Nichicon Muse ES with MKP or paper oil decoupling cap. I prefer with cap because, without, I never find the same quality in bass impact and groove. It's typical with violoncello of "Suites pour Violoncelle seul de J.S. Bach" or on "Dark Side of The Moon" (of course this album is better in LP, but it could be used to test DAC).
Power supply is a real good debate. I tried a lot of PSU types. I noticed (or perhaps it's just a question of taste), that the more sophisticated is the PSU, the less you have dynamic and live in the music, and more you have compressed sound.
My favorite PSU for tubes is a simple tube rectifier (sometimes silicon if I have no choice), a good MKP or paper oil capacitor, a good choke, a medium capacitor, a resistor or an other choke and a big capacitor. Sometimes I stabilise with an IRF830 like on the CS4397, if I can do without, I prefer.
But I use a Schaffner filter on primary of the transformer, and very good transformers with electrostatic screen linked to ground.
It's the same for DAC PSU, a lot of people use complex PSU, and I tried these sort of PSU, I never heard a difference with a good LM350 or LM317 PSU. For me the real good solution is to use quality choke between 2 big quality capacitors before LM, and to use a choke between the LM supply last capacitor and the last capacitor at the chip.
I compared this sort of PSU with friends who have made same projects with complicated and sophisticated PSU, and in blind test, no audible difference, just that in lot of cases my PSU gave more dynamic.
I test all what I create with the same system, my 2A3 SE amplifier with Klipsch 98db speakers and my ecc99 totem pole preamplifier (or sometimes other amplifiers). I can put my ears on klipsch drivers and horns and I hear nothing without signal, it's enough for me ;-)
Cordially
Pascal
Last edited:
With the 28R resistor it is hard on the edge, yes. But it works well
But that´s why i started the thread, to talk about things in the circuit.
Every opinion is welcome.
Well, if so: use only one channel per DAC. I am suspicious the two channels of the 1541 or not the same when they heat up and one is faster to break down (heat ...). At least I have seen several chips fail on continuos 1 KHz 0 dB files. One channel starts to 'break down'.
It would mean inverting the channel, and even if possible keeping one channel empty per chip.
I am hard pressed to find out which channel though, sorry, I'd have to go see, I didn't register it in a Note.
albert
Hi Albert,
thanks for your recommendation.
In the meantime i have finished my new Reference TDA1541A DAC project and i am running the Tube-I-zator stage with "only" two DAC modules.
So the R-I/V value of 28R is perfect.
Best regards,
Oliver
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.