|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
63.246.183.223
In Reply to: RE: Bugle is Not SRPP posted by Triode_Kingdom on August 24, 2014 at 16:38:14
I'm not sure this is the TubeCAD article I was thinking of, but it came up in a quick search. The author summarizes SRPP circuit operation without a load near the end of page 2. The paragraph begins with "What if there is no load resistor?" If you click through to the rest of the article (quite a few pages), he goes into much more detail later.
--------------------------
Buy Chinese. Bury freedom.
Follow Ups:
That's an interesting article.
"Unlike the White Cathode Follower, the limitation to the SRPP is that it can only be optimized for one specific
load impedance."
The problem I'm having is how do we define the load?
When driving the grid of a DHT the load is reactive.
There's more to it than just the value of the grid resistor.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
"how do we define the load?"
If you skip to the conclusion at the end of the first article I linked, you'll see that Broskie doesn't recommend the SRPP for driving high impedance loads. If you NEED to use the SRPP for some reason, I suppose you could use a lower value grid resistor (a load more suited to the SRPP), and maybe that would effectively swamp the reactance of the tube. In that case, the reactance wouldn't enter into the calculation.
The link below is the article I was thinking about originally. It's an easier read, and there's a bit of comedy thrown in for good measure...
--------------------------
Buy Chinese. Bury freedom.
I spent some time with SPICE this evening to see where the "crossover" from push-pull to single ended takes place in a SRPP. I plugged in a pair of 6SN7s, each with a 2.2K cathode resistor. The lower cathode is not bypassed. B+ is 400V.With these components, currents through the two tubes are in phase when a 50K or higher load is applied. They are out of phase when the load is 40K or below. At roughly 45K, there is virtually no AC current through the upper tube, so this is the crossover point. I assume this means the upper tube is acting as a constant current source with these values.
Note that this "crossover" function is gradual. At 50K and 40K, AC current through the tubes is very unbalanced, with current through the upper tube being much smaller than through the lower. The currents don't become balanced until the load falls to about 17K (push pull), or rises to roughly 4Meg (single ended). Voltage gain with a 17K load is about 75% of that produced with a 4meg load. A 220K load produces an AC current ratio of roughly 4:5 (acceptable), and gain is virtually the same as at 4Meg.
Changing Rk (upper cathode resistor) to 4.7K pushes the load up to 440K to produce the same 4:5 current ratio in single ended mode. Balanced current still occurs at 4meg. On the push-pull side of the equation, balanced current now occurs with a load of 42K, and the "crossover" is pushed up to about 95K. FWIW, the SRPP with an Rk of 4.7K clips at just over 360V PP. The SRPP with a Rk of 2.2K clips at just over 200V PP.
Just to push this to the limit, I changed Rk to 10K. This caused the "crossover" to occur with a 200K load, and balanced push-pull output with a 95K load. I suspect there's a downside to pushing Rk so far up in value, but I've run out of time to test for that. The point is that even with such a high value for Rk, any load above 200K will cause the SRPP to act as a SE amplifier. At more usual values of Rk, loads well under 100K are required.
--------------------------
Buy Chinese. Bury freedom.
Edits: 08/24/14
I've read Broskie's treatise on the SRPP before, and I re-read some of it just now. So far as I can see, he never considers the condition where the Rk for the lower tube is NOT capacitor-bypassed. (At least, in every one of his schematics, there is a bypass capacitor.) There was an article in Glass Audio years ago that attempted to prove one should not bypass the lower Rk in SRPP. (Something having to do with equalizing the current flow, which is your subject, too.) Further, you did not bypass it in your SPICE simulation. Can you say something about how the SPICE read-out might have differed if you bypass Rk? Thanks.
Lew, I can't detect any significant different in SPICE with the lower cap bypassed. Gain of the circuit increases of course, but the current balance doesn't change, nor does the "crossover" point. What's funny about this is that the circuit appears to me as though it could be better analyzed without the lower stage. The bottom half looks as though it's only providing gain and acting as a mid-Z source to drive the upper half. However, I've tried to drive the upper half directly from a generator using various series impedances and DC voltages (the upper half is where the phase change normally takes place when the load Z of the SRPP is adjusted). Once I separate the two halves, modifying the value of the load has no effect on the phase of the AC current in the upper half. It's probably just that I'm not accurately simulating the lower half with a generator, but this is bothersome. Guess I'll have to confirm all this with test equipment once my workbench and shop are back together.
--------------------------
Buy Chinese. Bury freedom.
This discussion about SRPP stages is very interesting, and Jon Broskie's articles on the subject are, as always, very informative and thought-provoking.I'm trying to understand if there is in fact any benefit to be had from using an SRPP when driving into a high-impedance load. (Which means it is definitely operating in SE mode, and not push-pull.) There are some designers who use SRPPs frequently, even when driving essentially infinite impedance loads. This clearly seems to mean that much of the original point of the SRPP topologogy is lost. Is the use of an SRPP in such circumstances "pointless but harmless," or would it be better to use some other design for the amplification stage in question? Or does it have any quantifiable benefits, even though it is not working in the "intended" way?
Chris
Edits: 08/28/14
"I'm trying to understand if there is in fact any benefit to be had from using an SRPP when driving into a high-impedance load. (Which means it is definitely operating in SE mode, and not push-pull.)"
Morgan Jones in Valve Amplifiers states that the lower half of the SRPP suffers from increased distortion compared to a common cathode stage because it's driving a relatively low impedance load. It's not clear to me if the load he's referring to is the upper tube, the load attached to the circuit, or a combination of both. In other words, his statement may or may not be applicable to an SRPP driving a high Z load. It's very unfortunate that in his description of the circuit, the load Z is never specifically mentioned, nor do his representative schematics include a value. He does however include a graph that shows THD of an SRPP declining rapidly with an increase in applied B+.
The latter might explain why my own SRPP driver with +500V does not inject significant distortion into the amplifier chain. Or, it might be that distortion is low because mine is driving a 200K load. I am unable to find anything in print that really goes into more detail on this point.
Like you, I've noticed that many designers use the SRPP where it doesn't appear useful. I sometimes wonder if they're doing this simply to make use of the other half of a dual triode envelope. It might also be that the audio community in general has so thoroughly bought into the concept of the SRPP as a topology that cancels even-order harmonics, it's employed by many designers in an attempt to reduce THD. When I look at the SRPP from a casual perspective, it also appears that it might be able to swing closer to the rails than a common cathode amplifier, or that the upper half might act somewhat like a triode CCS. In fact, I don't think either is true, but perhaps these concepts are all part of the recipe for the popularity of the circuit.
I think your question about the SRPP being "pointless but harmless" really gets to the heart of the matter. If Morgan Jones had published circuit details for the test he performed regarding THD VS B+, we might have an answer to that. As it stands, I suspect someone will need to revisit this with empirical measurements over various load impedances and voltages before we know.
--------------------------
Buy Chinese. Bury freedom.
I said above:
"Morgan Jones in Valve Amplifiers states that the lower half of the SRPP suffers from increased distortion compared to a common cathode stage because it's driving a relatively low impedance load. ... his statement may or may not be applicable to an SRPP driving a high Z load."
Well, I reviewed the Tube CAD article (SRPP Deconstructed), and it does in fact discuss the load of the lower triode. Broskie says that without a load on the circuit, the lower triode sees a load impedance that is "no greater than" the number provided by this formula: Z = rp + (mu + 1)Rk. If he's correct, this works out to about 53K ohms with a pair of 6SN7s and 2.2K Rk. That's a very reasonable load for a 6SN7, and I suspect it wouldn't change too much with an output load of 200K or so. This makes it appear that Jones was indeed referring to an SRPP loaded with the relatively low load Z for which this circuit was originally intended. It is also supported by the results I saw in my own amplifiers regarding THD.
--------------------------
Buy Chinese. Bury freedom.
"...once my workbench and shop are back together."
TK, that seems to be a theme lately.
Not sure what your reason is but I'm in the middle of a major purge as my shop had just got out of hand to the point that I couldn't work on anything.
I'm telling ya, leave a cleared flat surface anywhere around me and I'll fill it up with junk in a heart beat.
A couple of more days of cleaning and tossing and I'll be back in business.
I just have to keep telling myself "You can't keep more than you have room for".
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
A messy work bench is one thing, but I've had another toy taking up all the space in my workshop for more than a year now. Great fun, but I've finally decided it has to go, I'm too addicted to tubes. :)
--------------------------
Buy Chinese. Bury freedom.
Nice!
Something like that could make me forget about tubes (at least for a while).
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
It's always better when everyone can actually see it. Here are two crops I put together to show both modes of the SRPP. Sorry for the small print; the tubes are 6SN7, cathode resistors 2.2K. I printed the load resistance on each in larger type. That's the only difference between the two circuits. The red sine is upper tube current, blue is the lower:
--------------------------
Buy Chinese. Bury freedom.
Thank you for your effort in running and showing us the sims, they're indeed very informative.
Nice plots! They illustrate Jon Broskie's point very clearly.
Chris
Thanks, TK - your efforts are greatly appreciated!While I was aware that the SRPP load must be above some certain point, I'd not considered what happens when that load greatly exceeds that point.
Edits: 08/24/14
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: