I don't doubt what you heard, Frank. But I think it can't necessarily be generalized. It could depend on many specific details of the affected system.
Maybe in your case the RF happens to infiltrate through a particular PN junction after it enters an IC, where it gets rectified and then just happens to change a DC operating point of a particular transistor circuit, somewhere deeper inside of that IC, raising the DC operating point a little too much and causing the audible effects you described.
Maybe in some other system, there is a DC operating point that is a little bit too low and the RF fixes the problem and the system sounds better, and only sounds best when there is a certain RF level present.
Granted, certain RF modulation types could cause the rectified RF to modulate a DC operating point, or directly mimic an audio-range signal (and AM could actually easily BE demodulated into intelligible audio). But who knows?
And there are the wide/broadband RF bursts that are generated by something electrical that is very narrow in the time domain (a switch being flipped on, for example). Those can actually be manifested as DC offset step-functions or impulses that get amplified and sent to the speakers, where they sound like pops or cracks.
In terms of RF, I would probably worry most about the very subtle effects, which are just below the level of being consciously noticeable or identifiable by a typical careful listener who is not looking for anything wrong. They might not notice anything in particular being wrong, and probably wouldn't associate it with RF if they did notice. That leaves them with something like "This system is not great.". Maybe then it's viewed as a "bad design", which technically, it is, but not in the way they think it is.
The problem is compounded by the fact that the electromagnetic environment is highly variable, by location and time of day. If I were designing and selling high-end audio equipment, I would try to make sure that it was as immune as possible to RF.
By the way, there's an interesting section in the TDA2030A datasheet, at ti.com, Section 4, about a very simple method for measuring TIM (Transient Intermodulation Distortion), which also touches on limiting the input's slew rate, for an amplifier. But everything is an input, for RF. There are remedies for some of the other RF paths into an amplifier that are discussed in Chapter 7 of Analog Devices' / Walt Jung's "Op Amp Applications Handbook", at http://www.analog.com/library/analogDialogue/archives/39-05/op_amp_applications_handbook.html .
Maybe in your case the RF happens to infiltrate through a particular PN junction after it enters an IC, where it gets rectified and then just happens to change a DC operating point of a particular transistor circuit, somewhere deeper inside of that IC, raising the DC operating point a little too much and causing the audible effects you described.
Maybe in some other system, there is a DC operating point that is a little bit too low and the RF fixes the problem and the system sounds better, and only sounds best when there is a certain RF level present.
Granted, certain RF modulation types could cause the rectified RF to modulate a DC operating point, or directly mimic an audio-range signal (and AM could actually easily BE demodulated into intelligible audio). But who knows?
And there are the wide/broadband RF bursts that are generated by something electrical that is very narrow in the time domain (a switch being flipped on, for example). Those can actually be manifested as DC offset step-functions or impulses that get amplified and sent to the speakers, where they sound like pops or cracks.
In terms of RF, I would probably worry most about the very subtle effects, which are just below the level of being consciously noticeable or identifiable by a typical careful listener who is not looking for anything wrong. They might not notice anything in particular being wrong, and probably wouldn't associate it with RF if they did notice. That leaves them with something like "This system is not great.". Maybe then it's viewed as a "bad design", which technically, it is, but not in the way they think it is.
The problem is compounded by the fact that the electromagnetic environment is highly variable, by location and time of day. If I were designing and selling high-end audio equipment, I would try to make sure that it was as immune as possible to RF.
By the way, there's an interesting section in the TDA2030A datasheet, at ti.com, Section 4, about a very simple method for measuring TIM (Transient Intermodulation Distortion), which also touches on limiting the input's slew rate, for an amplifier. But everything is an input, for RF. There are remedies for some of the other RF paths into an amplifier that are discussed in Chapter 7 of Analog Devices' / Walt Jung's "Op Amp Applications Handbook", at http://www.analog.com/library/analogDialogue/archives/39-05/op_amp_applications_handbook.html .
Last edited:
Tom, I'm only generalising because of the number of times I've heard this in my own experiments. Plus, the negative qualities I've mentioned are very typical of the majority of other people's setups that I've come across - this does encourage me to put 2 and 2 together, 😉 ...
The "very subtle effects, which are just below the level of being consciously noticeable or identifiable by a typical careful listener who is not looking for anything wrong." is exactly what I'm referring to. Normal distortion is discerned by the listener, and pronounced as such; low level distortion has been discerned by the listener when he says "This system is not great." ...
The "very subtle effects, which are just below the level of being consciously noticeable or identifiable by a typical careful listener who is not looking for anything wrong." is exactly what I'm referring to. Normal distortion is discerned by the listener, and pronounced as such; low level distortion has been discerned by the listener when he says "This system is not great." ...
Last edited:
Not wishing to deny or question what you are saying, Frank, I would only point out that exactly the same thing could happen with inferior supply capacitors. The solution to this is the aforementioned RC (1 Ohm + xxx nF, just like a zobel) from power supply to ground.
The high range becomes dead, metallic sounding, harsh, and it's usually the titanium dome tweeter that gets the blame, when in fact all it's doing is reproducing what it's getting.
The high range becomes dead, metallic sounding, harsh, and it's usually the titanium dome tweeter that gets the blame, when in fact all it's doing is reproducing what it's getting.
Patent US4594561 - Audio amplifier with resistive damping for minimizing time displacement ... - Google Patents
I don't know if this is worth a read ? I skipped through it and probably wrongly said to myself that is simply the Q of a circuit . I doubt I gave it 2 minutes so forgive . It claims to have identified a new distortion mechanism .
DF 96 , I hope you see it .
I don't know if this is worth a read ? I skipped through it and probably wrongly said to myself that is simply the Q of a circuit . I doubt I gave it 2 minutes so forgive . It claims to have identified a new distortion mechanism .
DF 96 , I hope you see it .
Patents are written in a way which makes them very difficult to read. One might almost think that the aim is not to publish the idea (as intended by patent law) but encourage accidental violation and subsequent punitive action - but that would be cynical.
However, the gist seems to be that adding resistors across reactive elements stops some esoteric distortion mechanism. Pure nonsense IMHO.
However, the gist seems to be that adding resistors across reactive elements stops some esoteric distortion mechanism. Pure nonsense IMHO.
I came to that conclusion also . Resistive damping of a tuned circuit is hardly new if that is what is being said .
as I wrote before, some 2 weeks ago I had the chance to do some comparisons, including my amp vs. an entry-level Accuphase (entry-level as per their definition, e.g. 3k new). there was a definite improvement in highs reproduction with the Accu. which made me remember my amp's FR response which you can see below. -2.5 dB at 20kHz. not much, eh? I mean, we all know that not many adults are able to hear past 18kHz anyway, some don't even manage past 16k. (FWIW, I'm 33 and never had an audiogram done; I suspect a tiny bit of hearing loss in the right ear because of an otitis I suffered from a few years back).
and... after all, just look at a spectrogram of real music, you'll be able to see there's not much high-frequency content to begin with, not even instantaneous, let alone averaged. right? and, as if it wasn't enough, there's Fletcher-Munson. we're definitely less sensitive to highs, right? and... finally, ~2dB is considered the "barely detectable" threshold for freq. response variations.
but I thought I have all the tools to test just what those 2.5dB at 20kHz mean. so I generated some white noise and applied an inverse curve in order to have a near-flat frequency response up 20kHz. then A/B'd from my listening position as I have a PC remote. I could even do ABX, as pressing "next" a large number of times with the eyes closed makes it pretty impossible to tell which of the 2 files is which.
well, I couldn't believe how audible the difference was. in a way it's "barely" detectable. in another way, that "barely" means a lot. it's able to give a very distinct tonality to the hiss and make it sound, well, more "open". I'll try it with music too but I don't expect any change in perception.
next time when someone goes on about how "smooth" the NOS DACs are I'll remember to keep some salt nearby 😀
and... after all, just look at a spectrogram of real music, you'll be able to see there's not much high-frequency content to begin with, not even instantaneous, let alone averaged. right? and, as if it wasn't enough, there's Fletcher-Munson. we're definitely less sensitive to highs, right? and... finally, ~2dB is considered the "barely detectable" threshold for freq. response variations.
but I thought I have all the tools to test just what those 2.5dB at 20kHz mean. so I generated some white noise and applied an inverse curve in order to have a near-flat frequency response up 20kHz. then A/B'd from my listening position as I have a PC remote. I could even do ABX, as pressing "next" a large number of times with the eyes closed makes it pretty impossible to tell which of the 2 files is which.
well, I couldn't believe how audible the difference was. in a way it's "barely" detectable. in another way, that "barely" means a lot. it's able to give a very distinct tonality to the hiss and make it sound, well, more "open". I'll try it with music too but I don't expect any change in perception.
next time when someone goes on about how "smooth" the NOS DACs are I'll remember to keep some salt nearby 😀
Last edited:
Patents aren't that hard to read but they are a different language with a really specific grammar to make the legal challenges work in a consistent way.
I didn't go too deeply into the patent (it has expired) but the core issue is managing coupling of the power supply modulations in the the ground and input signals. That is a very valid issue. I can't say if his trick works. I doubt it would have been something that could survive a court challenge (resistors to discharge the caps were common long before it was filed).
I didn't go too deeply into the patent (it has expired) but the core issue is managing coupling of the power supply modulations in the the ground and input signals. That is a very valid issue. I can't say if his trick works. I doubt it would have been something that could survive a court challenge (resistors to discharge the caps were common long before it was filed).
@mr_push_pull
All you forgot to think about is phase response. With -2.5 dB at 20 kHz, it is not good.
I read somehwre, can't remember where, that many people can hear a 4 degree phase shift at 20 kHz as a slightly out of focus sound, not clear enough. Working backwards, you need an actual bandwidth of about 200 kHz to avoid this problem.
I push mine to 350+ kHz at -3 dB ref. full nominal power into 8 Ohms, then I install a low pass filter with -3 dB at around 200 kHz. Never fails, even if the absolute overall results do vary.
I have to say it - I am constantly amazed how in the end, I arrive at something Harman Kardon has already done many years ago. Nice, clean, simple classic circuits, no wild stuff. Though they did have their share of madness in form of incredibly complex ciruitry, especially in the up market Citation series.
All you forgot to think about is phase response. With -2.5 dB at 20 kHz, it is not good.
I read somehwre, can't remember where, that many people can hear a 4 degree phase shift at 20 kHz as a slightly out of focus sound, not clear enough. Working backwards, you need an actual bandwidth of about 200 kHz to avoid this problem.
I push mine to 350+ kHz at -3 dB ref. full nominal power into 8 Ohms, then I install a low pass filter with -3 dB at around 200 kHz. Never fails, even if the absolute overall results do vary.
I have to say it - I am constantly amazed how in the end, I arrive at something Harman Kardon has already done many years ago. Nice, clean, simple classic circuits, no wild stuff. Though they did have their share of madness in form of incredibly complex ciruitry, especially in the up market Citation series.
I believe it's called legalese. 😀 Texts written by lawyers for other lawyers to read. 😀
Aren't such resitors called "bleeders"? Or am I short circuiting?
actually, I didn't 🙂
but I have absolutely no serious reason to believe that what I heard was because of phase-related issues. yeah, I know, if you have -3dB at 20kHz you have phase issues below that. but all the studies I know about don't mention phase audibility at the top of the range. quite the contrary, audibility is said to be lower there. at least compared to midrange audibility.
and I could mention that a decoupling cap at the input of an amp, if incorrectly chosen, can have effects up to the mid-bass where it's pretty much proven since the 70s to be audible. you rarely see discussion about that, everyone cares about the magnitude response 🙂
and, with independent channel white noise there's practically no phasing to talk about.
but yes, I do agree that an DC-coupled amp having a corner frequency in the 100s of kHzs region and will shut up all naysayers 😀
Last edited:
Thats alot of rolloff for only 18K and it will absolutely affect the sound below it, you need double bandwidth(0-40K) to not have any top end issues and more has been known to "sound " better for whatever reasons.
Regards,
allow me to waste my time, money and nerves drawing that conclusion 🙂
IME, in this regard (post 20kHz response) "has been know" = guys who sell stuff say so. I'm not saying it's not true but I'll refrain from drawing any conclusions until I do side to side comparisons.
and one more thing. there are 2 distinct issues at play.
1. the pre-20kHz phase deviation caused by my amp
2. the audibility of a 2.5 dB boost centered at 20kHz. no-one asked whether my EQ was linear or minimum phase or what happened post 20kHz 🙂 even if my conclusion (the audible difference between the 2 amps being caused by magnitude response alone) is partially incorrect, the discussion of minimum vs. linear phase is still relevant.
Last edited:
Yep, all sorts of things can create "problem" sound, at least what I call such, I just group them under a generic banner of 'weaknesses'. These days there is so much RF being spewed in all directions that it's obviously a more major issue, and it's extremely easy to check whether there is an audible impact, simply by turning everything in range off.
I have zero tolerance these days for "grubby" sound, something that very expensive, and highly engineered components quite often produce, simply because the effort hasn't been made to sort out every last issue ...
Well, the sort of gear at the recent hifi show, and dealer showrooms ... 😛,😀
Seriously, I listen into the sound, rather than the broad front edge of it, to the inner detail - the quality that gives the replay either a hifi sound, or preferably makes it work as a reproduction of a musical event. By that standard much ambitious gear is grubby, it's all upfront, and there's very little past that sometimes very impressive, high impact assault ... 😀
Seriously, I listen into the sound, rather than the broad front edge of it, to the inner detail - the quality that gives the replay either a hifi sound, or preferably makes it work as a reproduction of a musical event. By that standard much ambitious gear is grubby, it's all upfront, and there's very little past that sometimes very impressive, high impact assault ... 😀
Slowly, slowly, movement forward occurs ... . These are comments by a member of another forum who has a system which most would consider very high end, who has put an enormous amount of money into setting up the listening enviroment to be just right: How important is the recording?? - Page 3 and How important is the recording?? - Page 4.
His comments mirror my findings and thoughts ... typical ambitious systems get the "spatial congestion" disease really, really badly ...
His comments mirror my findings and thoughts ... typical ambitious systems get the "spatial congestion" disease really, really badly ...
- Status
- Not open for further replies.