I've heard a lot of good about SRPP - it's supposed to have very low distortion and (depending on where you take the output) low output impedance. However as many of you probably agree, figures don't mean good sound necessarily.
What I'm asking is how would you describe the overall musicality, detail levels, transparency and so fort (insert your favourite mojowords here) of SRPP, and how would you compare it to the traditional common cathode plate loaded topology. Is the SE magic still there in SRPP?
What I'm asking is how would you describe the overall musicality, detail levels, transparency and so fort (insert your favourite mojowords here) of SRPP, and how would you compare it to the traditional common cathode plate loaded topology. Is the SE magic still there in SRPP?
A topology does not have a characteristic sound. An SRPP built with one particular valve, with particular component values and voltages, in a particular system, will not have the same "musicility" (whatever that means) as another SRPP using a different valve under different conditions. The question has no answer.
Do you want low distortion or musicality? As you mention SE magic I assume the latter.
The SRPP does not have low impedance output, unless by low impedance you mean lower than a common cathode but much higher than a cathode follower. Typically, SRPP output impedance is about 0.3-0.5 times the impedance of the same valve used in common cathode. This is why it also has a reputation for being fussy about what follows. Too much load pushes up distortion.
I find the current popularity of the SRPP rather puzzling. In many cases a common cathode followed by a cathode follower would work much better, yet some people don't like cathode followers. Maybe this is because there are too many poorly designed ones, using the wrong valve type.
The SRPP does not have low impedance output, unless by low impedance you mean lower than a common cathode but much higher than a cathode follower. Typically, SRPP output impedance is about 0.3-0.5 times the impedance of the same valve used in common cathode. This is why it also has a reputation for being fussy about what follows. Too much load pushes up distortion.
I find the current popularity of the SRPP rather puzzling. In many cases a common cathode followed by a cathode follower would work much better, yet some people don't like cathode followers. Maybe this is because there are too many poorly designed ones, using the wrong valve type.
Sy,
Surely Sir, you jest?
Set up correctly an SRPP stage has lower distortion than ANY other form (including Mu-Follower or CCS load) of Gainstage (not follower) that can be realised using the same double triode.
Of course, not all SRPP's are set up correctly for low distortion.
Ciao T
Surely Sir, you jest?
Set up correctly an SRPP stage has lower distortion than ANY other form (including Mu-Follower or CCS load) of Gainstage (not follower) that can be realised using the same double triode.
Of course, not all SRPP's are set up correctly for low distortion.
Ciao T
Hi,
Yes. This can be fixed though, by adding another tube set up as a modified white follower. If you do that you end with something called after it's inventor the "Gomes" stage. Others have called it supertotempole.
John Broskie did an expensive expose of the Gomes stage and (as usual) found that in theory the Aikido is better at suppressing power supply noise* while the Gomes stage did rather well on other accounts.
I personally like this stage as it can be optimised for lowest distortion with no recourse to load impedance, while retaining a classic "SE" spectrum and also sounding quite "single-ended", subjectively speaking. I agree with John Broskie that it does not have a lot of power supply noise rejection, but I like to make powersupplies very quiet anyway, so I never found this an issue.
A Linestage from diyhifisupply used a 6CG7 & 5687 Gomes Stage (yes, I had a small part in the design). This in practice measured better than the Aikido version on all counts and more critically, it sounds quite extraordinarily good.
My personal biggest complaint about it was that I have absolutely no use for a linestage with 20dB Gain.
Ciao T
* I still fail to understand this quest for ultra-high power supply rejection to the exclusion and even detriment of other performance attributes, including sound quality. I always feel it is much easier to design a competent powersupply to start with, instead of trying to fix noise later with all sorts of circuit tricks that react unfavourably to non-ideally matched tubes, non-ideal capacitors and resistors with tolerance...
Yes. This can be fixed though, by adding another tube set up as a modified white follower. If you do that you end with something called after it's inventor the "Gomes" stage. Others have called it supertotempole.
John Broskie did an expensive expose of the Gomes stage and (as usual) found that in theory the Aikido is better at suppressing power supply noise* while the Gomes stage did rather well on other accounts.
I personally like this stage as it can be optimised for lowest distortion with no recourse to load impedance, while retaining a classic "SE" spectrum and also sounding quite "single-ended", subjectively speaking. I agree with John Broskie that it does not have a lot of power supply noise rejection, but I like to make powersupplies very quiet anyway, so I never found this an issue.
A Linestage from diyhifisupply used a 6CG7 & 5687 Gomes Stage (yes, I had a small part in the design). This in practice measured better than the Aikido version on all counts and more critically, it sounds quite extraordinarily good.
My personal biggest complaint about it was that I have absolutely no use for a linestage with 20dB Gain.
Ciao T
* I still fail to understand this quest for ultra-high power supply rejection to the exclusion and even detriment of other performance attributes, including sound quality. I always feel it is much easier to design a competent powersupply to start with, instead of trying to fix noise later with all sorts of circuit tricks that react unfavourably to non-ideally matched tubes, non-ideal capacitors and resistors with tolerance...
Hi Sy,
Care then to support your blanket assertion with some actual data?
Which Tube gainstage (as stage with gain) topology using a dual triode and no looped feedback can you present to illustrate that it offers lower distortion than an SRPP optimised for lowerst distortion?
Ciao T
Care then to support your blanket assertion with some actual data?
Which Tube gainstage (as stage with gain) topology using a dual triode and no looped feedback can you present to illustrate that it offers lower distortion than an SRPP optimised for lowerst distortion?
Ciao T
I'll ask you the opposite question: can you show me an SRPP stage with lower distortion and lower source impedance than a properly designed mu follower or beta follower- or for that matter, a common cathode-cathode follower combination- using the same tubes? I won't put in the new qualification that the triodes have to be dual, just to make things easier for you.
The "no looped feedback" is another new restriction that you suddenly threw in, but I won't hold you to that, either.
The "no looped feedback" is another new restriction that you suddenly threw in, but I won't hold you to that, either.
In fact, I'll make this even easier for you- can you top the distortion and source impedance performance of the ImPasse using a 6SN7 in SRPP (feel free to use another tube on top)? Or better results than the 6SN7/D3a mu follower shown and characterized in Morgan's book?
Just to make things fair, please compare at the same output voltages and load- we each showed distortion performance at 20V or so.
Just to make things fair, please compare at the same output voltages and load- we each showed distortion performance at 20V or so.
I believe Jones reported distortion around -50dB (0.3%) for that didn't he? I may have that wrong though.
If you read:
Evers, M. V. (1996). Distortion Minima Loading of the SRPP. Sound Practices. (13) pp. 40-1.
he reported distortion of less than 0.2% for apparently slap-dash SRPPs, dropping to less than 0.05% under the right (heavy load!) conditions!
The Gomes does not use an SRPP, it uses an actively loaded triode.
http://www.tubecad.com/2004/GomezAmp1.gif
Also, if you DC-couple an SRPP it becomes purely single-ended in operation, so its not an SRPP any more anyway.
Close, it was actually about 6dB better for the better 6SN7s. This is at about 20VRMS out.
In the ImPasse, with CCS and cathode follower, I managed to do a bit better than that, getting 0.04% (second harmonic dominated) at 14VRMS with a 10k load. Jack Walton independently confirmed that.
Now, how does the 6SN7 SRPP do at that output level?
Ok I looked it up: Valve amplifiers - Google Books
Varied between -50 and -55dB, and -56dB for a CV1988, but those are weighted results.
Hang on, that's not fair! That CCS adds a whole new dimension. A fair test would be a resistance loaded gain stage and CF versus the SRPP; two triodes and three resistor versus two triodes and two resistors.
Last edited:
Dear Stuart,
I observe this now quite a few diverse occasions.
You make a blanket, unqualified statement condemning something you consider "fashionable", contrary to actual observable facts on the subject.
Then I challenge you on the lack of qualification as well as accuracy in your statement and you backpedal, bringing more and more conditions and qualifications to any possible experiment, in fact already confirming that my challenge was accurate.
May I suggest that to save both of us some time you consider:
1) Be less pithy, pith is a virtue in many fields, but not in the technologickal arts.
2) Dogmatik, unqualified statements are appropriate in religion, they do not constitute proof in any sense and they do not help others to make engineering driven decisions, like you claim you wish to engender.
3) Resistors, Capacitors, Chokes, as well circuit features such Mu-Follower, SRPP, Cathode Follower, Gomes Stage, Aikido are devices/designs with diverse features, benefits and disbenefits and which partickular device/design suits a given application depends on diverse faktors.
4) If you like to help others understand (and perchance even convince them your approach is the best), giving them an expose of the different factors is a by far better way than to issue dogmatic one liners.
In the meantime, I will take the liberty to challenge any such false statements when I encounter them.
Oh, BTW, you should check out the VanEvers Article in your copy of Sound Practices. Using a non-linear valve as the ECC88 he got much lower distortion out of a SRPP than you list for your best efforts using a supposedly much more linear valve.
Of course, it should not REQUIRE the above article, if one understands the nature of the SRPP and can read the anode curves of a Triode, then, to avoid the conclusion that an optimised SRPP MUST have lower distortion than any resistive or CCS loaded loaded common cathode stage takes a lot of effort in avoiding to see the plain and obvious.
Of course, I do appreciate when people apply a lot of effort, on the other hand.
Ciao T
I observe this now quite a few diverse occasions.
You make a blanket, unqualified statement condemning something you consider "fashionable", contrary to actual observable facts on the subject.
Then I challenge you on the lack of qualification as well as accuracy in your statement and you backpedal, bringing more and more conditions and qualifications to any possible experiment, in fact already confirming that my challenge was accurate.
May I suggest that to save both of us some time you consider:
1) Be less pithy, pith is a virtue in many fields, but not in the technologickal arts.
2) Dogmatik, unqualified statements are appropriate in religion, they do not constitute proof in any sense and they do not help others to make engineering driven decisions, like you claim you wish to engender.
3) Resistors, Capacitors, Chokes, as well circuit features such Mu-Follower, SRPP, Cathode Follower, Gomes Stage, Aikido are devices/designs with diverse features, benefits and disbenefits and which partickular device/design suits a given application depends on diverse faktors.
4) If you like to help others understand (and perchance even convince them your approach is the best), giving them an expose of the different factors is a by far better way than to issue dogmatic one liners.
In the meantime, I will take the liberty to challenge any such false statements when I encounter them.
Oh, BTW, you should check out the VanEvers Article in your copy of Sound Practices. Using a non-linear valve as the ECC88 he got much lower distortion out of a SRPP than you list for your best efforts using a supposedly much more linear valve.
Of course, it should not REQUIRE the above article, if one understands the nature of the SRPP and can read the anode curves of a Triode, then, to avoid the conclusion that an optimised SRPP MUST have lower distortion than any resistive or CCS loaded loaded common cathode stage takes a lot of effort in avoiding to see the plain and obvious.
Of course, I do appreciate when people apply a lot of effort, on the other hand.
Ciao T
Hi,
This is just one of several variations of the Gomes stage. More where discussed around 12 or maybe 15 years ago on the Sound practices e-mail list (check the archives).
Also, I will still maintain that a SRPP functions as one, even if the load is infinite and the output is take from the lower anode and not the upper cathode.
The push-pull nature of the stage and the resultant straightening of the compound anode curves for a class A push pull stage (viz. Steve Bench's site for a good explanation) remains in force.
Sorry, this I may need some more explanation and maybe an example.
Ciao T
The Gomes does not use an SRPP, it uses an actively loaded triode.
This is just one of several variations of the Gomes stage. More where discussed around 12 or maybe 15 years ago on the Sound practices e-mail list (check the archives).
Also, I will still maintain that a SRPP functions as one, even if the load is infinite and the output is take from the lower anode and not the upper cathode.
The push-pull nature of the stage and the resultant straightening of the compound anode curves for a class A push pull stage (viz. Steve Bench's site for a good explanation) remains in force.
Sorry, this I may need some more explanation and maybe an example.
Ciao T
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.