The mu-follower and SRPP can be regarded, and analysed, as the same circuit. The difference is that in the mu-follower the upper valve cathode resistor is so large than different bias arrangements are needed, to separate AC and DC connections to the upper grid. The SRPP can be used in a balanced configuration, with upper and lower valves being the same with the same resistor value - this can reduce even-order distortion. The mu-follower is inherently unbalanced, so instead low distortion is obtained by raising the anode load of the lower valve. Both of these circuits have an optimum load, but this is usually ignored and the alternative near-infinite load used instead.
PSRR depends largely on the ratio of cathode resistors, assuming the same valve type top and bottom.
PSRR depends largely on the ratio of cathode resistors, assuming the same valve type top and bottom.
This is an particularly important point with broad implications across high-end audio, and which I too have been recently contemplating. It seems there are two parts to this issue.
Part one: why do most home audio systems, even though they may cost thousands of dollars, fall so short of providing a live sounding (not necessarily pleasurable) auditory experience? By live sounding, while I don't mean 100% totally convincing, I do mean where there is that recognition of a real event that is quite hard to put in to words, but you know it when you hear it. What strikes me as especially intriguing is how live stage events using amplifiers and sound reinforcement quality speakers often produce a 'live' sound that most home systems do not. Even when the great majority of the sound is coming from the event's sound system. Perceived dynamic freedom seems at the root this, at least to my ears, however, it also seems there is more than a simple SPL difference at work.
Part two: must those home systems which do reproduce a live sounding auditory event cost upwards of as much as the cost of the average home!? My scientific brain agrees with yours, it doesn't accept such a notion. It seems to me that the real R&D challenge which is before the audiophile industry is to discover why a run-of-the-mill concert PA system can sound live in a way that most home systems don't. Then to translate that knowledge in to truly affordable home use systems which do not cost as much as the home in which they would be placed.
Last edited:
Because of the amazing amount of data compression inherent to stereo recording and playback. That renders everything else secondary.
Indeed live recordings are mixed on the hoof to sound "live". Studio recordings are carefully crafted to sound acceptable on almost any system - which inevitably involves peak compression and other modifications.
I do tend to agree that very good systems can be made from ordinary components. Ordinary components will take you 80% of the way given a well designed circuit, but the remaining 20% is hard won with many many expensive components. For me that remaining 20% is not worth it and refining the actual circuit is the better way to go.
It seems that many DIYers go from a single fixed (old) circuit and try to achieve their 80% through expensive components - when in fact all that is available to them is a 20% improvement.
Shoog
I do tend to agree that very good systems can be made from ordinary components. Ordinary components will take you 80% of the way given a well designed circuit, but the remaining 20% is hard won with many many expensive components. For me that remaining 20% is not worth it and refining the actual circuit is the better way to go.
It seems that many DIYers go from a single fixed (old) circuit and try to achieve their 80% through expensive components - when in fact all that is available to them is a 20% improvement.
Shoog
Member
Joined 2009
Paid Member
Then I was kidding myself that I understood the essence of this issue. I don't get what the issue is with the iron-domains. I thought in PP trafo the primary dc currents balance and there is nothing strange happening with the iron ???
Or, are you talking about how the core magnetization has to flip direction and traces out a hystereses curve which means nasties happen at the zero crossing with each cycle of music ? something avoided with SE. If so, how does you design address this ?
I agree, it is the great crime in audio. I notice is very clearly with my FM tuner. Most of the 'pop' stations are compressed but there's one local French language channel that sounds as if they are doing something so 'right' and it sounds wonderful.
Last edited:
SY, tell me more. Is the compression of which you speak inherent in the mediums, or more, studio practice and proceedure?
As far as playback equipment limitations, this gets us back to the question of home system cost, and whether spending big bucks is a prerequisite to achieving that live sound quality to which I've been referring. It would seem not, since even modest sound reinforcement systems, with all of their other sonic limitations, often sucessfully transfer that 'live' quality intact.
Last edited:
My concern is that while I agree there is an audio cost/benefit version of the Pareto rule, this rule presently lies on a curve which is asymptotic. Meaning, that while exponentially increasing expenditure brings us only diminishingly nearer to the live experience, not only will we never reach the live experience via the current curve, but will remain quite some distance from convincingly live home reproduction. Why? What is it, exactly, that makes reproducing a live event, even an fundamentally electronic event such as a rock concert, seem to forever remain a pipe dream? Is the live physical auditory event magical insome way? As Nazaroo says, and I concur, I don't believe in audio magic.
Last edited:
As has been discussed - achieving a live sound is physically impossible for the reason that even live recordings are exposed to post production mixing to achieve that all important compression for standard equipment. You cannot put back what has been removed in the studio,
Having said that the best system I have heard is a friends normal HIFI amp driving a pair of 15inch PA cabs. PA speakers have the significant advantage of high sensitivity which tends to convey microdetail really well whilst retaining the ability to deliver convincing bass kick effortlessly. These are qualities which are difficult to achieve with conventional sealed or ported box HIFI speakers.
Shoog
"Or, are you talking about how the core magnetization has to flip direction and traces out a hystereses curve which means nasties happen at the zero crossing with each cycle of music ? something avoided with SE. "
The nasties from hysteresis occur near the peaks, not at the zero crossing. SE OT's still have magnetic domains flipping around, just fewer of them, since they are mostly pinned down by the DC, hence the 3X bigger core to compensate. Same net flux change per volt in either case.
The nasties from hysteresis occur near the peaks, not at the zero crossing. SE OT's still have magnetic domains flipping around, just fewer of them, since they are mostly pinned down by the DC, hence the 3X bigger core to compensate. Same net flux change per volt in either case.
Last edited:
Inherent to the medium. Studio practice can improve it by clever signal processing or (more frequently) make things less "real" by layering heavy dynamic compression and EQ onto a fundamentally flawed medium.
I'm limiting myself to the capture of live sound events which are not PA systems. The process of creating music via recording (e.g., multitracking, overdubbing, DI, backward recording, etc, etc, etc) is a different thing and beyond considerations of "fidelity" or "reality," since there never was a "reality" to reproduce.
OK, with that limitation in mind, let's consider a live musical event. You're sitting in a room/hall/club immersed in a three dimensional soundfield. Your ears sample that soundfield at many points, since your head isn't locked into a vise. So there's a three-dimensional field with the pressure being time and space variant. Take that complex 3d sound field and sample it at two and only two points with transducers having who-knows-what properties (which are strongly frequency variant) regarding polar response. Now we have compressed a huge amount of information into a pair of time variant one dimensional functions gathered by those transducers.
Fast forward to your listening room. Those two single-dimensional functions are fed into two loudspeakers with who-knows-what properties (which are strongly frequency variant) regarding polar response. The output of those transducers interacts with your listening room. Do you have any confidence that the 3d soundfield where your ears are (not two points, but a set of volumes, since your head is not locked in a vise) created from the transducers and the room has any relationship to the originally samples 3d soundfield?
Headphones and binaural recording don't help- they remove the variables of polar pattern and the listening room, but the mikes still only sample at two points rather than through the volume subtended by your ears/head-motion, and in reproduction, the soundfield moves with your head, a very unnatural experience. There's still a huge loss of information.
The stunning thing is that stereo recording and reproduction works as well as it does.
I agree. For those who are skeptical, try making or procuring an uncompressed piece of music and playing it on your home system. Be careful though, because it may cause damage! Most home systems aren't capable of reproducing the dynamics of live music. In fact, some instruments naturally have too much dynamic range and musicians frequently use real time compression to even out the dynamics in live shows. But as a whole, live music is simply less compressed.
As far as sensitivity, I have also found that high sensitivity speakers and low power amps are the way to go for the reasons you stated. I can't see owning speakers that require more power than an average SET amp can produce, say 5 wpc. Large PA speakers can work here, but I prefer a small single driver with no crossover network at all, effectively making an active speaker system with the drivers directly connected to the SET amp's OPTs. I do run a subwoofer for bass reinforcement, a small cone/sealed box arrangement is fast enough to keep up with a single driver system if it's x'ed over low enough..
Which means that while I am enjoying this thread, I have no use for such a power amplifier as Nazaroo has made, although I could see a SE mu-follower driver stage feeding a power triode, the solution my Simple SE uses presenting the driver with a CCS load seems to work well too, and is very simple.
So now we see the real reason why low feedback tube amplification sounds more "alive", the 3/2 power law natural tube response (with low feedback) acts as a de-compressor. Add more feedback and it goes "lifeless".
This may be the real reason Nazaroo's amp sounded so good, no global feedback. (and no power supply noise near full power output, where it has to run due to the efficiency problems, giving maximum tube de-compression)
This may be the real reason Nazaroo's amp sounded so good, no global feedback. (and no power supply noise near full power output, where it has to run due to the efficiency problems, giving maximum tube de-compression)
Last edited:
This thread is rather long but it lacks very important piece of real technical information - THD/IMD numbers with spectrum analysis.
I loaded PP amp with 20 Hz sinewave at max power (60W) and have not observed any problem described by original poster. There are no visible 100 Hz hum modulation and vice versa.
I must admit design idea is really very original, and finished product may sound fine and find its customers, but it is unnecessary complex and I see no proof of its outstanding characteristics, or advantages over classical PP.
I loaded PP amp with 20 Hz sinewave at max power (60W) and have not observed any problem described by original poster. There are no visible 100 Hz hum modulation and vice versa.
I must admit design idea is really very original, and finished product may sound fine and find its customers, but it is unnecessary complex and I see no proof of its outstanding characteristics, or advantages over classical PP.
That would depend upon what transformer you use.
I suggest using a 10-20 uF or larger cap to avoid bass issues.
I have used various transformers.
I have not built a pair in over 12 years.
simlar to a 6L6 pair, about 20-40 watts per channel,
depending upon operating voltages chosen.
- Status
- Not open for further replies.