If the load is infinite then both valves must conduct the same. One cannot increase conduction while the other decreases, hench, it cannot be push pull, and we are out of SRPP territory.
If you take the output from the lower anode then it is not an SRPP (I call it a half-mu stage). The upper valve acts simply like a resistor, and there is no push-pull action, even if the load is not infinite. The OP's question was specifically about the SRPP, so the Gomes is off topic (I haven't seen any other versions of it).
Would you have a copy of that article I could read? These numbers keep getting thrown around with NO reference to output level, load, or gain, which certainly puts me at a disadvantage since I actually specified these important variables in my arguments. What does 0.05% mean? At 1V? 10V? 1k load? 1M load? Second harmonic? Seventh harmonic?
Why would that be true? Is Vg (and hence Vp) equal for top and bottom with signal applied?
It helps to actually put some on the bench and take spectra. If memory serves, I've posted some of these results in the past- it's a very, very linear tube (excepting, of course, the notorious relabeled 6ES8s).
Hi,
They must conduct the same current. Yet voltages across them change and as their various parameters depend on voltage and current we find that both lower and upper valve in the SRPP are subject to non-linearities.
If we had a simple common cathode stage with a perfect current source, the inherent non-linearity and the "bunching up" of the curves provide the absolute limit to non-linearity.
One may be able to provide cancellation between the common cathode stage and (say) a cathode follower to further lower distortion, but that is another story and requires a different design approach.
Now let's look at the SRPP.
The lower tube operates just as above. But now the load is another identical triode. In other words our load becomes non-linear. And as it so happens the non-linearity of the load impedance introduces changes that in effect straightens out the compound anode curves.
So, understanding the SRPP we find that we can can optimise the SRPP for any given fixed load from infinity to zero (but only for that one fixed load impedance) to produce a THD basically as low as we care to tune.
A friend's tubed line stage that uses SRPP plus follower with an added load between the stages can be tuned to a THD with two zero's after the decimal point, as long as the tubes used are reasonably consistent.
The funny thing is that it sounds much better with a much higher THD and with the extra load completely removed, thus leaving only the cathode followers THD (the SRPP cancels nearly perfect).
Call it what you will. It's an SRPP.
No. It does not act like a resistor. Moreover, it cannot act like a resistor. It is a triode, not a resistor. So it does not act like one and that is precisely why the PP cancellation remains in force.
I maintain that the Gomez stage is an alternative variant of a series connection of SRPP and White Follower, with some added potential to linearise the circuit.
So it is not wholly off topic.
Ciao T
They must conduct the same current. Yet voltages across them change and as their various parameters depend on voltage and current we find that both lower and upper valve in the SRPP are subject to non-linearities.
If we had a simple common cathode stage with a perfect current source, the inherent non-linearity and the "bunching up" of the curves provide the absolute limit to non-linearity.
One may be able to provide cancellation between the common cathode stage and (say) a cathode follower to further lower distortion, but that is another story and requires a different design approach.
Now let's look at the SRPP.
The lower tube operates just as above. But now the load is another identical triode. In other words our load becomes non-linear. And as it so happens the non-linearity of the load impedance introduces changes that in effect straightens out the compound anode curves.
So, understanding the SRPP we find that we can can optimise the SRPP for any given fixed load from infinity to zero (but only for that one fixed load impedance) to produce a THD basically as low as we care to tune.
A friend's tubed line stage that uses SRPP plus follower with an added load between the stages can be tuned to a THD with two zero's after the decimal point, as long as the tubes used are reasonably consistent.
The funny thing is that it sounds much better with a much higher THD and with the extra load completely removed, thus leaving only the cathode followers THD (the SRPP cancels nearly perfect).
Call it what you will. It's an SRPP.
No. It does not act like a resistor. Moreover, it cannot act like a resistor. It is a triode, not a resistor. So it does not act like one and that is precisely why the PP cancellation remains in force.
I maintain that the Gomez stage is an alternative variant of a series connection of SRPP and White Follower, with some added potential to linearise the circuit.
So it is not wholly off topic.
Ciao T
Dear Stuart,
The article is in Sound Practices #13.
You can obtain the whole sound practices archive on CD for a very reasonable outlay here:
SOUND PRACTICES MAGAZINE ARCHIVE CD - eBay (item 330442397394 end time Aug-11-10 08:38:26 PDT)
Of course, the whole thing is full with the lunatick ravings of audio fashionistas, but it also has many extremely solid articles on fundamentals and (among other things) Gordon Rankins servo controlled cathode follower anno 94 IIRC.
I honestly believ that buying and reading the whole archive would give your views some more far-sightness.
I'f just copy the article from my own copy if I wanted merely to proove a point, but my discourses never are about making, proving or asserting a specific point.
So please buy the damn thing and read it.
Ciao T
The article is in Sound Practices #13.
You can obtain the whole sound practices archive on CD for a very reasonable outlay here:
SOUND PRACTICES MAGAZINE ARCHIVE CD - eBay (item 330442397394 end time Aug-11-10 08:38:26 PDT)
Of course, the whole thing is full with the lunatick ravings of audio fashionistas, but it also has many extremely solid articles on fundamentals and (among other things) Gordon Rankins servo controlled cathode follower anno 94 IIRC.
I honestly believ that buying and reading the whole archive would give your views some more far-sightness.
I'f just copy the article from my own copy if I wanted merely to proove a point, but my discourses never are about making, proving or asserting a specific point.
So please buy the damn thing and read it.
Ciao T
http://www.jeack.com.au/~lucas/2A3SRPP.jpg
What would happen if one were to put a plate resistor on top of that totem pole? I.e. for voltage dropping purposes.
What would happen if one were to put a plate resistor on top of that totem pole? I.e. for voltage dropping purposes.
I would increase the value of the 12k /3W resistor, instead.
300+ V seems a bit high for a 6DJ8 in that circuit.
Of course it does! If you take the output from the lower anode then the upper triode receives no input signal, and it is powerless to do anything other than act like a resistance equal to Rk(mu+1) (not a perfectly linear resistance of course, since mu is not perfectly constant). And it in unquestionably not an SRPP. It is neither shunt regulated nor push-pull, which I'd say pretty much disqualifies it...
The White cathode follower, on the other hand, is a kind of SRPP, but it is more useful to keep the names separate so everyone knows what we mean by an SRPP amplifier.
I suggest you read Valley & Wallman's Vacuum Tube Amplifiers which covers this nicely. BKAA61 - Vacuum Tube Amplifiers
Jacques, ok, but what would happen if one would still put a plate resistor on top?
In my design, there is only one B+. I have found this to improve sonics significantly, and I wouldn't do it any other way anymore. Of course this complicates things much when trying out SRPP.
Also, how critical is the balancing of the two cathode resistors, in my link they are both 820 ohm (for the 12AX7), but is it important for the function of the SRPP that they be as close as possible? I.e. hand picking resistors.
In my design, there is only one B+. I have found this to improve sonics significantly, and I wouldn't do it any other way anymore. Of course this complicates things much when trying out SRPP.
Also, how critical is the balancing of the two cathode resistors, in my link they are both 820 ohm (for the 12AX7), but is it important for the function of the SRPP that they be as close as possible? I.e. hand picking resistors.
The gain would fall somewhat. Overall not much effect in that circuit.
Nope. Not unless you have extraordinarily well matched triodes! (Like statistically impossible...)
I'll pm you my email addy and greatly appreciate it.
OK, 2V, 100k load. I'm far less impressed now.
Thanks for filling in the crucial detail.What does the SRPP do if set up direct coupled to the next stage (power triode, 1626)? Is this simply stupid, or can I get a good sound out of it? Is SRPP for cap coupled only?
I'm planning on trying out my 6SL7's on this experiment, with B+ currently around 434V, SRPP output around 180V, cathode resistors for 6SL7 maybe 820 ohm (lower cathode bypassed with 100µF cap + some film caps) - what kind of performance can be expected from this kind of setup?
I'm planning on trying out my 6SL7's on this experiment, with B+ currently around 434V, SRPP output around 180V, cathode resistors for 6SL7 maybe 820 ohm (lower cathode bypassed with 100µF cap + some film caps) - what kind of performance can be expected from this kind of setup?
Direct coupling to the power tube can give better performance if you drive the power tube to the point where it is drawing significant grid current.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.