No
But I was very impressed back in '85 or so, when the AP System One was launched. Very low distortion with a transformer, even av 10 Hz. That was new --- to me at least.
We don't need a transformer between pre and power amp. Do we even need balanced? And is the way high-end audio use "balanced" the right way to do things? History is important here, there is a lesson to be learned.
SM, from an academically viewpoint, most of what you write is probably correct.
But 30 years of experience tells me that an academic approach alone is just not good enough. There are so many missing links between what we hear and what we can measure. I find subjective feedback from professional users far more valuable than what my current SOTA measurement setup tells me. But I'm lucky, my clients are professionals.
If we just measure, we get lost in the trap of getting the best possible measurement. So we push everything to the extreme and optimize our design to measure the best at that condition. In audio, these extremes are very rare, say 0.1% of the time. But what really matters is audio quality in the 99.9% of the time when there are no extremes. A very good example of this is the "first watt" philosophy by Nelson Pass. One watt equals typically equals 80-90 dB SPL listening level. That's a comfortable listening level. If it's louder, our ears distortion increase and our ability to detect distortion is lowered. So my concern then is my power amps performance at 1W, not close to clipping.
On the other hand, I totally agree with you that using resistive loads for testing power amp performance is just silly. This is the blind eye approach to amp design. Actually, having studied the power amp and speaker evolution for the past three decades, things are just getting worse. Speakers are no longer "8 ohms" and amps cannot longer cope with complex and generally low impedance. The most tragic example is todays surround amps, with 7 channels and a power supply barely sufficiant for one channel.
No. When my sense of humor is showing, you'll know it.
Really?
So you're claiming to be able to read my mind, Ed?
Well I'm afraid you're doing a horrible job of it.
The measurements I linked to were specifically made at my request using cables that I had supplied to Bruno.
The genesis of this was some distortion measurements John had made using a variety of cables which he was claiming was due to microscopic "diodes" in the wire itself.
Hate to be the one to tell you, BUT there are diodes in your metal wires. More than you will ever bother to measure. I have measured them.
When the graphs of these measurements were finally published, myself and others had serious doubts that the distortion products shown were created by the cables themselves let alone microscopic "diodes" hiding inside the wire.
Then a short time later, I came across a post by Bruno Putzeys on one of the rec.audio newsgroups where he'd mentioned that he'd done some distortion testing on some cables and nothing had showed up, even though he was measuring around 25dB below where John was measuring.
John's argument was that Bruno wasn't measuring the same cables that he was.
So I arranged to send the same cables to both John and Bruno. A freebie giveaway, an older used Radio Shack Gold interconnect, a brand new Radio Shack Gold interconnect, and some interconnects made by Dan Banquer of RE Designs which were made using RG-174 coax which had a copperclad steel center conductor.
For all but the RG-174 cables, I took the cables in question and split them down the middle like zip cord and sent one half to John and the other half to Bruno.
Bruno's measurements turned up nothing beyond the residual distortion of the AP rig however John's measurements were still showing distortion products even though again, Bruno was measuring well below where John's equipment was capable of. This indicated that the distortion John was measuring wasn't being produced by the cables themselves.
That was the sole purpose of the measurements in the PDF I linked to.
So stick to audio measurements and leave the mind reading to the Amazing Kreskin.
se
That's because AP uses a technique by which the output transformers are driven by a negative output impedance equal to the DC resistance of the primary windings.
se
Last edited:
Thanks for the background.
I think we agree the micro diode claim is not valid. I have actually tested for the effect and not found it.
But we disagree on the test results. If one method showed nothing and the other something, more experimentation is needed to see why. I see you used the word "indicated" not "proved" so I really have no argument with that.
It could be the poltergeists bother John more or there really could be a difference in what the instruments are actually looking at. I don't know.
I have some feel for the AP as I use one and as with all equipment it has limits. I understand the ST measurement technique is different.
But how did you manage to get RS cables that had only one side bad?
But I must protest your denial of my ability to read minds. Think of a number, any number...
Yes, more experimentation on John's end I would say. SY went over to his place to help him work it out, but that never really panned out as far as I'm aware.
It's essentially the same. Spit out a test tone, notch out the fundamental, do an FFT on the residual. John fed the residual from the ST to an HP spectrum analyzer.
But how did you manage to get RS cables that had only one side bad?
Ancient Chinese secret.
Ok, let's put you to the test. We'll make it simple.
I'm thinking of a number between one and ten.
What number am I thinking of?
se
Last edited:
FYI, here are John's measurements of a JPS Labs cable, a Radio Shack cable and a van den Hul cable respectively.
se
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
se
I guess the "tall" component just below 16 kHz is a nearby computer screen?
Doug Self did some cable measurement some 20 years ago and published the result in JAES. His conclusion was that IF there is a diode effect in cables, the distortion components are below -140 dBu or so. He even made a special very low noise amp for his measurements.
Last edited:
I believe so, yes.
Yeah. And I'll leave it for the mentally ill to lose sleep over what's going on below -140dB.
se
No it won't oscillate. When the compensation is too large, the Zout becomes negative indeed and what you'll see is that Vout increases when you increase the load. Of course it also increases distortion.
I have a vintage AP S1, I've tried it many years ago when I was experimenting with a similar design.
jan didden
How much is the fundamental tuned out, in other words what does the noise floor correspond to in the ral world, it is obviously not a -70 floor. I am curious about the level they correspond to.
It could easily be an rf issue causing problems if it wasnt seen to a much lower level when an AP was used?
OT:
Some years ago I bought a cheap hifi in a hurry to fit with the sitting room, so there were other parties to satisfy with the size and look. It turned out to be absolutely pants performance wise. So I borrowed a 1 GHz scope and spectrum analyser from work over the weekend. It burst into song on just a small fraction of a cycle when larger lf signals were playing, and it was clearly audible. I printed off the results and took it back to the shop as not fit for purpose.
Wrinkle
Transformers in audio
Using negative impedance to linearize output transformers is an old and well known principle. And patented as well......
Walt Young had some example circuits in his latest (?) book. Here is an example from Lundahl Transformers: http://lundahl.se/pdfs/papers/feedbck.pdf
And if my memory serves me right again, JBL used negative output impedance in power amps to reduce distortion in LF drivers by an order of magnitude. This was some 40 years ago...
On the input side you can make a transformer very linear by placing it in the summing node of an inverting amplifier. I did that 25 years ago with great success. In pro audio you sometimes just need galvanic isolation...
Using negative impedance to linearize output transformers is an old and well known principle. And patented as well......
Walt Young had some example circuits in his latest (?) book. Here is an example from Lundahl Transformers: http://lundahl.se/pdfs/papers/feedbck.pdf
And if my memory serves me right again, JBL used negative output impedance in power amps to reduce distortion in LF drivers by an order of magnitude. This was some 40 years ago...
On the input side you can make a transformer very linear by placing it in the summing node of an inverting amplifier. I did that 25 years ago with great success. In pro audio you sometimes just need galvanic isolation...
Yes, the AP circuit was patented by Bruce Hofer, then (and still) AP's president. Probably expired by now; the S1 came out around 1984 IIRC.
But the methods you mention are not all the same concepts; of course wrapping a feedback loop around it decreases nonlinearities. But the AP circuit compensated exactly (in theory) for the DC resistance of the primary as that was the component causing the distortion. So it was different from a "brute force" global nfb loop.
jan didden
Last edited:
Actually, it did. We spent a little time denoising the measurement with aluminum foil and clip leads. Indeed, we could plug and unplug cables and get consistent results. However, the results were NOT the same as his previous ordering, and of course, they were at extremely low levels.
It would take more experimentation to determine the mechanism- my opinion is that it's purely a matter of interaction between the lumped impedances and the test rig's circuitry, but that's only an opinion since we didn't take it any farther.
I'm wondering how isolated from the RFI/EMI environment the measuring equipment was. Were Faraday cages used? How about the cables themselves?
What noise/hash producing equipment was nearby? Cell phones, computers, fluorescent lighting, furnace, a/c, microwave, clothes dryer, dishwasher, etc.........
Since we've gone down this road, again, with John's blessing, some answers to my previous questions would've been appreciated. Even if the answers would appear to be speculation on John's part I would still be interested in hearing them.
Last edited:
Once upon a time there was a rational goal and yardstick by which to judge the performance of audio equipment. Once the novelty of merely being able to record and reproduce sound had worn off, become taken for granted, engineers, musicians, and music lovers couldn't help but notice the vast gulf between what they heard live and what their recording technology produced. And so the concept of high fidelity was born. The goal was to reproduce the sound of some of the finest musical instruments played by the best musicians peforming the best written music in the best sounding music venues at will from a recording played at home. It was a worthy goal. It was worthy of a first class effort in 1930, it would be worthy of a first class effort today. But not only has a satisfactory solution to that problem never come close to being achieved, it beat the best minds that took it on and today, there isn't even a pretense that this is still the goal. In fact from the way I see it, there is no longer a rational goal to efforts in this industry today. There doesn't seem to be a goal at all except to make money. Small wonder it's an industry that is dying. That may be one reason, there are probaby a lot of others. Lack of knowledge of even what real music, one of the fine arts is probably one of them. It's amazaing how many young students we get here who want to learn to play an instrument who have never heard a symphony or a concerto, not just a live performance but even a recording of one. Perhaps our much faster paced lives with so many activities precludes the time it takes to sit and passively enjoy music for half an hour or an hour at a time. Or perhaps the average attention span has become so short that anything taking more than 3 minutes, coincidently the time for a pop song between commercials is too long. Even here, few if any memorable tunes have been turned out by the usual hack pop song writers in recent times.
When you watch a color TV transmission or video recording, you adjust the color intensity, tint, brightness, and contrast based on variables in each program content to match what you remember as accurate to get the best picture. Often you use skin tones as the most sensitive criteria because you are most familiar with it and subtle errors that would be inconsequential for many other objects such as being too green or too purple are annoying for skin. This was especially true for NTSC transmissions viewed on CRT sets. You also adjust the picture for the prevailing ambient light. But ultimately you judge the performace of the set by a large number of measured criteria that demonstrates which set has the best capabilites. (Those criteria can also include the availabe range and types of user adjustments.) It is the same with audio equipment. The program material varies in many ways, spectral balance being most important among them. So do the room acoustics. Both drastically affect the ultimate result of what you hear just the way signal quality and ambient light affect what you see from a TV set. Only the audiophiles and "engineers" who cater to them have removed nearly all of the adjustments to compensate for these variables or never incorporated them in their designs in the first place. And so what we are left with is no way for even those experienced listeners who are familiar with the sound of live music to make critical adjustments or even judge which equipment performs best.
If you were to give me a list of all of the equipment on the market and ask which combination will sound like live music, I'd have to say none of the above. So far the problem has beaten everyone who tackled it and it doesn't look to me like matters will change anytime in the forseeable future, at least not from the efforts I'm aware of. That's why my own efforts are largely DIY.
- Status
- Not open for further replies.