Well we should settle this like gentlemen. Either pistols at 10 paces or rapiers, or last one still standing at the bar after the trade show!
I might point out what I attempt to contribute here, on this specific thread, when I get the chance. I am most interested, HERE, on what makes 'good' sounding audio equipment. Is it design sophistication, low distortion, good parts, etc?
From the Bajazzo example, it is not just modern design that wlll do it. In fact, I suspect that many later examples of similar equipment , with quasi-complementary, full complementary, all silicon, etc, would not impress me much, subjectively. If it were true, that modern design fixed all problems, it would have been said and done, many years ago. Of course, I am a 'holdout' for better sound quality, and some older equipment 'has it' and most modern stuff doesn't. Finding out WHY is the point of this whole thread.
From the Bajazzo example, it is not just modern design that wlll do it. In fact, I suspect that many later examples of similar equipment , with quasi-complementary, full complementary, all silicon, etc, would not impress me much, subjectively. If it were true, that modern design fixed all problems, it would have been said and done, many years ago. Of course, I am a 'holdout' for better sound quality, and some older equipment 'has it' and most modern stuff doesn't. Finding out WHY is the point of this whole thread.
It might be fun to run a good old-fashioned frequency response curve on a cherished radio, just to see what we like in that respect (for radio use, anyway). The guys building those radios had complete control of everything from airwaves to air waves, and surely knew something about tone balance.
Thanks,
Chris
-------------------------
Well, let me ask the obvious question.
Looking at some of the older equipment that was readily available at the time and produced in a reasonable quantity, what is your choice of a half dozen good amplifiers? That is, good out of the box and not requiring modifications.
I don't want to turn this into a top ten thread, since those are tedious. But some of my own favorites (selected Marantz, McIntosh and Harman Kardon from yesteryear), sometimes sounded very good, but if you asked me why it would not be because they sounded especially neutral or accurate. In fact they were not neutral, but they had a particular sound that was pleasing (but not necessarily accurate).
Last edited:
Hi Ed,
I'm trying to square your graph here with the resistor distortion results I showed in my book in Figure 13.2. I think we both used essentially the same measurement technique.
It looks here like the third order THD is at about -140 dB, but it is unclear what the power level is in this resistor. If I recall, in your excellent Linear Audio article you were running 62.5 mW through a 1/4W resistor. I'll assume for the moment that is what you are running here. It also looks here like the 9th is about another 35 dB down.
In my book Figure 13.2 I show THD vs frequency for an ordinary 1% metal film axial leaded 1/4W resistor dissipating 1/4W of signal. My curve projects to about -130 dB at 1 kHz (a value below the measurement floor of my setup, so your setup may be more sensitive.
Anyway, the THD I saw was strongly dominated by 3rd, as yours were. If I dropped my voltage level by 6dB, then I would also be at 62.5 mW and my 3rd would be expected to be down in a relative sense by 12 dB, getting me to -142dB, which, if I've done my math right, is pretty close to the result you show (assuming you ran the resistor at 62.5 mW. Does this reasoning and similar result look right to you?
One thing I'm having trouble figuring out is when I work backwards from the result you show for 9th and estimate what level of 9th I should have seen if the 9th was coming from the same mechanism (but maybe it wasn't). If relative 9th goes as 8X the dB change in voltage (did I do this right?), then my 9th should have been 48 dB higher than yours, and it would have been REALLY visible in my test. I know there are probably a lot of holes in my swag reasoning here, but where am I going wrong?
It just seems like you should have seen a lot less 9th than you did. Maybe this is where the issue of a different mechanism comes in (e.g., maybe a bi-metallic interface thing rather than the thermal mechanism that appears to be responsible for the 3rd). What do you think?
Cheers,
Bob
> Finding out WHY is the point of this whole thread.
One of the major 'WHYS' is that most earlier circuits
are capacitively coupled between stages.
You and your cronies diss capacitors , but the advantage
(being able to bias each stage for optimum fidelity without
concern for DC in or out) outweighs the shortcomings of
capacitors.
One of the major 'WHYS' is that most earlier circuits
are capacitively coupled between stages.
You and your cronies diss capacitors , but the advantage
(being able to bias each stage for optimum fidelity without
concern for DC in or out) outweighs the shortcomings of
capacitors.
Yes. It would be interesting to know the actual frequency response of the acoustic output. I had a suspicion that the old radio I had rolled off a lot at low and high frequency ends and it had a HF cut switch, as well. There was very little non musical noise produced. Very pleasant.
I'll bring a copy. The company is no longer around so I think you could just pdf it. The front and back covers are covered with Jim Williams style freehand cartoons, things like a Weston Photocell and meter counting cases of beer as they pass.
I agree. Audio components (radios) from this time were basically designed much more carefully (even by ordinary kitchen mono radios; particulary by the German brand "Telefunken"). Especially the avoid of the so called "true complementary" power output contributed to the exceptional sonic quality, additional the design of the tone controls in combination with the loudspeaker in the associated envelope.
In the german magazine "Funkschau" was released several articels of Telefunken's research results concerning home audio-, portable and television components (mainly between the years 1950 and 1980).
By post 13-16 about
http://www.diyaudio.com/forums/parts/167680-vintage-transistors-2.html
you will find various cover pics of Telefunken books.
By post #12, #22 and #13 about
http://www.diyaudio.com/forums/soli...-state-audio-amplifier-components-want-2.html
is to find the first solid state products from AR ( also a circuit diagram).
Last edited:
Bob,
No arguments here. The results you, Demian and I have track reasonably well. The resistor I showed is the only one to generate ninth. Next time I am set up I may run a few more power and frequency curves to see how it tracks. All of the other more common metal film resistors had no ninth which is most likely why you didn't see any!
I also found one of the miniature metal film resistors that showed more second harmonic than the others. That one is an audiophile favorite!
My intent in showing resistor distortion was a bit for the users, but more for a simple way for manufacturers to adjust their magic formulas to get better results.
The common assumption that the third harmonic issue is just tempco is not quite right. The assumption on tempco is that it is a straight line resistance change vs temperature. I suspect for some of the formulas such as this one the line is curved! So low classic tempco may be achieved but distortion will be higher. The low cost good sounding resistors may also have much greater thermal mass so the tempco does not come into play as much as on other types.
Scott,
Accidents sometimes have useful results. You have something, ratio between claims and performance, dumping factor!
ES
That might also apply to those 1980s vintage Quad current dumping amplifiers, and other things..
even some Audio engineers have been smarter than that Strawman for decades now
as all to boringly usual "standard measurements" seems to be anything you want to beat up somebody with??
what if your standard measurement were as simple as plotting THD vs level - bad things happen when the "dumper" just starts to supply output - likely a IMD test signal that just crosses the dumper threshold would be even better for seeing the flaws
engineers, circuit designers have been known to use their knowledge of the circuit to pick tests/levels, signals, frequencies that reveal circuit operation weakness - marketing people will select anything that they can to sell product - these are completely different motivations/goals - I think here we should concentrate on tests, measurements that do use circuit operation knowledge to illustrate strengths and weaknesses, please drop the strawman attacks
the deadzone of the Quad unbiased "dumper" is simply stupid in a linear amp - any minimal AB bias at all hugely reduces distortion
placed around a "not bad" Class AB amp a Black's Feedforward stage can improve distortion # - at high frequencies - but since accurate, reliable matching of the correction and main amp gains for a deep error null pretty much has to rely on both being high negative feedback amps most of the more sensitive audio frequency range is most simply controlled to low distortion by the high feedback rather than the error correction
as all to boringly usual "standard measurements" seems to be anything you want to beat up somebody with??
what if your standard measurement were as simple as plotting THD vs level - bad things happen when the "dumper" just starts to supply output - likely a IMD test signal that just crosses the dumper threshold would be even better for seeing the flaws
engineers, circuit designers have been known to use their knowledge of the circuit to pick tests/levels, signals, frequencies that reveal circuit operation weakness - marketing people will select anything that they can to sell product - these are completely different motivations/goals - I think here we should concentrate on tests, measurements that do use circuit operation knowledge to illustrate strengths and weaknesses, please drop the strawman attacks
the deadzone of the Quad unbiased "dumper" is simply stupid in a linear amp - any minimal AB bias at all hugely reduces distortion
placed around a "not bad" Class AB amp a Black's Feedforward stage can improve distortion # - at high frequencies - but since accurate, reliable matching of the correction and main amp gains for a deep error null pretty much has to rely on both being high negative feedback amps most of the more sensitive audio frequency range is most simply controlled to low distortion by the high feedback rather than the error correction
Last edited:
The Quad 'current dumping' example is significant to show what engineers will try to get away with, IF they can use lots of feedback to make 'spec'. The Quad design is almost an affront to 'hi end' design, and long term listening tests have confirmed that it didn't work very well. However, the CONCEPT of current dumping is a very interesting one, and I developed some circuits based on a WW article in the 70's, that I offered to Studer, (for free), in order that they could have better measuring line amps, and probably better sounding, too. Of course, they ignored me.
- Status
- Not open for further replies.