mikelm, the 'resonances' you describe are simply wonky stability. Many (all?) Golden Pinnae power amps suffer from this.
If you test them with real speakers, you often see bursts of oscillation on part of the waveform. This will be dependent on the level and thermal & signal history of the amp. Easily visible on a scope and audible too. No ABX bla bla needed.
If you measure distortion under those conditions, you'll see distortion MUCH greater than on the usual 8R loads.
No wonder Golden Pinnae amps sound different but in my book, these are NOT good amps.
BTW, many Golden Pinnae amps 'respond' (peaky response, oscillation, stability bla bla) to exotic cables too. Unlike boring amps used by us unwashed masses that don't. I'd quote my AES paper on this but I'm nervous of being shot at by Wave.
_______________
An important take on this for preamps is that decoupling & stability are VITAL for good sound.
The physical PCB layouts are an important part of this and few people understand what is needed.
Many people substituting HP OPAs for cheapo in circuits are simply hearing really bad oscillation & distortion cos these high strung racehorses dun like the mundane environments they are suddenly shoved into.
Frank, Microwave Office might sim. what such PCB layouts do but IMHO, the design of PCB layouts for good sound is, if not an Art, still a Craft, requiring both experience & good insight.
_______________
But to get back to Blowtorch ...
JC, what results did you get with the Hirata tests you conducted on Fri?
Quan's patent is US20110002472A1.
If someone can build a Quan box and bring it to JC, we'll see how a first class preamp performs on this really important test.
Not coming in on ac line -
BTW - the upper right photo is the incoming ac line when DUT is turned off. Lower right is an avaiable ac line iso-filter -RNM
BTW - the upper right photo is the incoming ac line when DUT is turned off. Lower right is an avaiable ac line iso-filter -RNM
Last edited:
mmmm - I think we're picking at nits here . . . it all depends upon the magnification factor - if it high enough almost any stimulus will set it off but it's not important . . . . .
I just had hoped that one or some you professionally trained guys could explain to me how a circuit tending towards resonance at 15Mhz for example ends up making an amp sound rough in the audio band - that's something I experienced but never really understood. If no one understands it that's fine I'll stop thinking about it.
In the meantime I'll start reading the thread from the beginning to see what I can learn
cheers
Last edited:
As a demonstration of how the Pioneer should have done it, there's this: Rey Audio RM-7V in Kenrick Showroom *Request track* - YouTube. Note in particular that the mic can be moved practically anywhere in relation to the drivers and the sound stays consistent, it's only when the camera moves completely to the side of one of the speakers that the sense of the sound being in front of you changes, but even here the naturalness of the sound remains. Unfortunately, yesterday, my YouTube feed was incredibly slow for some reason and that was disturbing the sound playback; I was hearing sibilance in this clip, a classic distortion problem, but today it's fine.
I've got nothing PC speakers to listen to this, but I can wind up the volume and the sense of the quality of this playback comes through very clearly; note especially that the room is a terrible audiophile's room, just about everything you can do wrong is part of the equation, no treatments, masses of other speakers around, you name it ...
Where I part way with many of the people here, is that I don't believe in bolting the gate after the horse's bolted - I don't fiddle with frequency responses or room treatments. I would see the Pioneer setup as suffering from a poor quality feed, the audio signal is defective at the point where the driver is rendering the sound, so either some of the internal electricals of the speaker are lacking, or more likely, the actual signal has been sufficiently damaged by the time it gets to the speaker.
Frank
my post #29861 As I said, it's called wonky stability on real life loads. Sounds REALLY rough too.
It's a real phenomena. 'Tending towards resonance at 15MHz' on the usual 8R loads is a warning that you will probably get this on real speakers.
_______________
JC, I hope the firestorm didn't destroy the Hirata results on Blowtorch you did on Fri.
Last edited:
There is, in this forum, a thread about J.m. Le Clear'ch constant directivity horns. Those horns witch give similar results. As well as mine.
Circular horns are better, on my point of view, because, if they are of the same diameter than the Bass unit, they have the same directivity curve at the Xover (according the bass driver is cut before fractionnating) in all directions. This help a natural crossing. Appolito's basses arrangement help to simulate the source coming from one point.
I hope all those examples, and comparisons with cone enclosures, will help to convince people talking of horns distortions or aggressive sounding.
Not because i'm convinced of that, but for their future pleasure.
Last edited:
Part 5 for FETs , part 6 for BJTs.
Mind also part 1 third page .
http://www.audiofaidate.org/it/articoli/Baxandall%20power%20amp%20design.pdf
George
My rationale on the FB issue is that for that mechanism to work correctly, the circuit has to be of very high performance already; in one sense the "FB amplifier", considered for a moment as a separate structure, has to be, say, of an order of magnitude better in performance that the the underlying circuit it's attempting to always correct: follows from the old adage, two wrongs don't make a right ...
My suggestion for a test to really nail what's going on is to use two quite low level, appropriate sine signals and do an IMD test. Then, mix in a significantly lower frequency at a much higher level with frequency chosen which will make sense after reading the following: do that same IMD test, but take absolutely no notice of any IMD products derived from the presence of the 3rd sine wave - hence choice of 3rd frequency will be crucial. Bump up the level of that 3rd frequency in increments to just under that for the combined signal to clip at any point, and do those IMD measurements at each point.
My belief is that the "good" circuits will show far less variation of the IMD spectrum only resulting from the 2 test signals interacting - in other words, the 3rd frequency is stressing the circuit, including its power supply, to see if that disrupts the behaviour when only looking at the context of the low level signals ...
Frank
My suggestion for a test to really nail what's going on is to use two quite low level, appropriate sine signals and do an IMD test. Then, mix in a significantly lower frequency at a much higher level with frequency chosen which will make sense after reading the following: do that same IMD test, but take absolutely no notice of any IMD products derived from the presence of the 3rd sine wave - hence choice of 3rd frequency will be crucial. Bump up the level of that 3rd frequency in increments to just under that for the combined signal to clip at any point, and do those IMD measurements at each point.
My belief is that the "good" circuits will show far less variation of the IMD spectrum only resulting from the 2 test signals interacting - in other words, the 3rd frequency is stressing the circuit, including its power supply, to see if that disrupts the behaviour when only looking at the context of the low level signals ...
Frank
You are welcome Brad. I consider it a “must” read.
I will ask. The BH curve nonlinearity and the permeability’s dependence on H field strength are phenomena known on x-former cores.
Assuming audio x-formers (input, coupling, output) have been used, studied and optimized for decades, where is the data quantifying the harmonic distortion generation due to these nonlinearities of ferromagnetic materials? (links, reference)
Partly answering my own question, here is a good old study (theory & measurements)
http://www.alcatel-lucent.com/bstj/vol14-1935/articles/bstj14-2-322.pdf
George
Your work is right on the money, mike, especially that related to eliminating spurious RF around the bridge rectifier.
Like you, I've spent a great deal of effort on getting rid of that nasty stuff coming in from bad behaviour in the mains, and agree that the rubbish can be tenacious, depressing as you say. A very simple RF injection device of gross proportions I use is a simple high power incandescent bulb in socket, connected to the system mains spur via an extension cord. I hold the socket and plug in the line where the cables join and deliberately pull them apart just enough so there's ferocious sparking, which I can delicately manipulate. The speakers of a normal system go mad with the sound of that interference, but it is possible to get to the point where the sound of the system remains completely mute, no matter how much you provoke it in this fasion.
Frank
Also, nothing stopping you paralleling any number of runs of the cable to get sufficiently low resistance - just more soldering required! Plus, then you can have fun with different twisting and braiding strategies ...
Frank
My take would be that the FB mechanism is particularly sensitive to those instabilities; it has to react extremely fast in comparison to the signal it's attempting to correct, so would tend to provoke mild, momentary, bad behaviour ...
Frank
Our guests tonight called my woofers in concrete boxes, "Washer and drier".
An externally hosted image should be here but it was not working when we last tested it.
audio power amps, even with modern transistors seldom have 5 MHz unity gain intercept – in all of the Cordell circuits I've simmed only measuring inside the error correction local feedback loop managed that high an gain intercept frequency
so for a hypothetical 15 MHz “feature” the global loop of typical audio power amps probably has << 0.3 “gain”, quite a few amps don't even have 1 MHz global loop gain intercept
The 'sound' of RF -
Oscillations - Can they be heard if above the hearing freq range. yes. Sort of... indirectly. Another one of those "its As If" things ---
RFI and oscillation - the tweeter will not behave well. Heating in the tweeter which will indirectly cause the tweeter to distort. Which then results in the audio signal sounding really, really bad just before the tweeter goes silent. It need not be 15MHz but often lower freq above hearing range is also common source of this indirect distortion. Switching amp artifacts in the output to the speaker will have similar affect if not heavily attenuated. -RNM
Oscillations - Can they be heard if above the hearing freq range. yes. Sort of... indirectly. Another one of those "its As If" things ---
RFI and oscillation - the tweeter will not behave well. Heating in the tweeter which will indirectly cause the tweeter to distort. Which then results in the audio signal sounding really, really bad just before the tweeter goes silent. It need not be 15MHz but often lower freq above hearing range is also common source of this indirect distortion. Switching amp artifacts in the output to the speaker will have similar affect if not heavily attenuated. -RNM
Last edited:
- Status
- Not open for further replies.