Sound Quality Vs. Measurements

Status
Not open for further replies.
diyAudio Member RIP
Joined 2005
Zetex do mention low impedance and how it is compromised by even the shortest wiring . That suggests it was in the authors thoughts . It looks a very good risk .

To put it into perspective Rotel recommend upgrading from 78 series regulators to a larger number of LM317 . As the comparison here is with LM317 we might have without too much complexity a sort of bronze , silver and gold . The circuit looks simple enough to have a small PCB made ( yes please anyone ) . I think I remember Naim using 24 regulators in their middle range tuned up Philips CD player . If I understood correctly they selected LM317 and had an outlet for the rejects ( the majority , 90% ? ) . He was unwilling to say what the rejects were rejected for . Speed and noise perhaps ? I do know they said it was just looking for the best of devices , all theoretically were OK . My suspicion is as they tested everything they used it was no big deal if someone would buy the rejects to do it . If you think about it they buy lets say 1000 pieces at $100 . Then sell the rejects for $50 and end up with something they feel to be special at less than 50 cents each ( as they test everything anyway that is not a cost ) Recently they have made their own regulator which seems very complicated by comparison with the Zetex .

BTW . Even though 95% convinced as I was at the beginning I will try RC and perhaps LC alternatives to a crystal . It is not about being right or wrong . It is about being honest and saying I genuinely know . I would not do it with many things as the complexity and cost would make it stupid . In this case it is 5 minutes work to have something working . Do I expect it to be better ? No . Will I enjoy doing it ? Yes . Everyone told me idler driven turntables were a waste of time . Sadly for 60% of my life I didn't question it . They rumble so you don't want one , end of ...... Do they ? I have the world record on that at - 79dB weighted ( shared with Martina Schoener as it's both of us who created it ) . The irony is nothing different to a Garrard 401 except attention to detail . All parts of mine fit a 401 . As no one believed I had no books to read . My interest in general electronics came from building it's PSU which is no small part of the - 79 dB . I feel I could have taken 3 dB off of that figured it the motor had been properly in the feedback loop . There was an output transformer in the way . The next step a bespoke motor coil as did Thornes on the TD125 . The TD 125 wasn't feedback as such , it did at least have the lowest impedance . Also that type of motor steps so feedback will do very little to reduce vibration . The Garrard has a more elastic coupling of rotor to stator field . It's vibration is mostly rotational speed plus 100/ 120 Hz . There is chaotic vibration ( measured and considerable, jitter ? ) from the motor thrust bearing which I eliminated by an air bearing of sorts .

With Digital it is like a vegetarian eating meat for me . I find I can these days . CD ......?
 
Nige, with everybody and their dog making electronic components these days, the need for incoming quality control is an absolute must for anyone half serious.

After stressing myself out with 317/337 woes, like exchanged pin places and similar ridiculous things, I just dropped them and went all discrete, all shunt. I still use them, but in non-critical places only, like providing power for LEDs and similar highly intellectual jobs.
 
diyAudio Member RIP
Joined 2005
Nige, with everybody and their dog making electronic components these days, the need for incoming quality control is an absolute must for anyone half serious.

After stressing myself out with 317/337 woes, like exchanged pin places and similar ridiculous things, I just dropped them and went all discrete, all shunt. I still use them, but in non-critical places only, like providing power for LEDs and similar highly intellectual jobs.
My "favorite" not too long ago was discovering two different pinouts for the ubiquitous TL431 in SOT-23 and I think SO-8.

And no, it wasn't an audio application, rather a detector of automotive filler, typically used to conceal damage.
 
A friend was putting together "introduction to soldering" LED flasher kits for Atlanta Mini Maker Faire and was having a problem with a batch of transistors, something like a 2n2222, he got from Digikey. They just weren't working in the circuit, whereas his previous ones did. I told him "check the pinout." He eventually looked at the datasheet for the exact part he ordered and found the pinout different from the previous transistors, but what I MEANT was for him to stick the thing in the transistor HFE connector/setting of his DMM using different pins for E B and C until it came up with a reasonable (50 to 200) value, in other words, check which pinout actually works as a transistor.
 
Frankly, if I could, I'd ban such products. To me, this is a swindle, if the original has a certain pinout scheme, anyone producing replacements should HAVE to adhere to it no matter what, or you can't put the original's designation, because it is not a pin-for-pin replacement.

When it's a small signal TO-92 or some such, it's not too hard to get around it, but if it's a power device, then it's not nearly so easy.
 
Does anyone here have the service manual, or just the schematics, of Harman/Kardon's legendary Citation XX?

That's one whopping legend of an ampifier. It's said that it's the best power amplifier ever made, and judging by its specs, like 500 A peak current output, it certainly looks a winner.

I know for a fact that it changed ecerything in HK, their complete topology logic changed radically. Up until it came along, they used more or less similar topologies like everybody else, just more refined; after it, they had topologies no-one else, as far as I am aware, used. Starting from HK 870 power amp, they were into sub-20 dB global NFB.

Citation XX is said to be the brainchild of three people - a Japanese engineer working for Shin Shirasuna (the company which actually manufactured HK gear), Richard Miller and Matti Otala.

I have a partial schematic, with hand jotted comments of Richard Miller, but unfortunately, it was drawn from memory (Richard Miller doesn't have its schematics ???), and has his hand written comments here and there. What is missing is the current gain stage, unfortunately, the most interesting part.

BTW, it's classic measurement specs are, shall we say, mundane, THD being given as 0.1% - hardly ground breaking. Yet, its aura is exceptionally strong. Perhapbs because (at least in part) it's so rare, it seems H/K and Shin Shirasuna had issues, like Shin Shirasuna offering exactly the same products under its own monkier, at, of course, a better price.

Has anyone here even seen it? Heard it in action?

John, I would imagine you took the time and trouble to check it out? Any views, comments, impressions?
 
diyAudio Member RIP
Joined 2005
I contacted Martin Zanfino, who was VP Engineering in those days for h/k, sent him a link to your post, and he said I could share this:

"This comment is at least partially correct. The Citation XX was already in the final pre-production prototype stage when I was hired to run H/K's engineering and product development in mid-1981, and it went into production a few months later. Due to the $7500 retail price (in 1981 dollars!), the lifetime global sales forecast was in 3 digits and the intent was to repair them all in-house. That is one reason for the lack of a published schematic or service manual.

The other reason is that all of the signal path (except for the power devices) was in a hybrid module that wasn't repairable and had to be replaced if it failed. The module included thick-film printed resistors in the differential stages that were laser-trimmed to minimize harmonic distortion. The automated trimming process was developed by a technical research center in Finland with which Matti Otala was associated, and all of the hybrid modules were manufactured in Finland. I'm not sure of the exact date, but the last batch of hybrids was probably manufactured in 1982 or 1983. I doubt that any formal documentation still exists.

Feel free to post this on the diyaudio website."

Brad
 
Member
Joined 2004
Paid Member
If its the amp I think it is I attended the initial CES presentation with Matti and he explained several of his ideas. The high current came from measurements he made of speakers under actual operation. Its the combination of peak current and phase angle that required the huge current he claimed. He also claimed that any frequency non-linear element in the chain, like a speaker crossover, would cause IM distortion if I remember this correctly. He also said that digital was really poor in TIM, however this was very early in the history of CD's.

He escorted several of us through a somewhat formal listening test, very different from the usual audiophiile A-B with its "this bloatamp amp crushes that wimpamp which doesn't need to exist" commentary. I learned a fair amount from those few hours.
 
@bcarso
Brad, many thanks for the supplied comment. I appreciate your time and trouble. Please thank the author for me.

It is quite possible that my info is not accurate, since it was derived from several magazine texts at the time, and it was 30 years ago.


@1audio
Not surprised to hear this, that was the end of an era, and the manufacturers were beginning to feel a pinch coming up. If anything, then H/K was a little late with it, I suspect that if it had appeared say 2 or 3 years before it did, it would have made a much bigger splash.

On the other hand, EVERYTHING takes time to mature, and let's not forget that Otala's theories had appeared in print merely 6 or 7 years beforehand. The impact does take time to sink in and develop.

To H/K's credit, about a year later, late '83 or early '84, they did have their model 870 power amp on the market, which was said to be a "poor man's Citation XX" in many ways, but obviously made much more traditionally.

I have never even seen a Citation XX, but I am very well acquainted with the 870; if half of those rumours are true, than the XX must have been a true sensation then, and would likely still be one.

In all these years, I have never seen one being offered on sale nywhere, for any amount of money, which appears to confirm that the lucky few who have it have no intention of selling it.

@Wavebourn
Let's put it this way - having a tremendous current capability can't hurt, so long as it is not achieved at the express cost of something else.

Exactly how much is enough is a more interesting question.

All too often, we behave somewhat irrationally (normal for the big children that we are) and look at absolute maximum ratings. I find this silly, because in all my years, I have never once heard of anyone playing full power sine waves into loudspeakers in a room.

If we are to play music, then we would have to look at our effective -3 dB power level, and at peak power only in short peaks. In reality, even much less than that, because even today's extra loud mastered CDs do have dynamics of over 3 dB, and it is assumed we do NOT want any clipping.

I have read somewhere that on average, over 95% of us spend our listening time at AVERAGE power levels of less than 1 W, with peaks going up over that. This would imply that a man buys a say 100 WPC amp to make sure those peaks of say 20 Watts are reproduced faithfully, or more faithfully than a say 50 WPC amp could do (not necessarily true, but on average).

The simple fact that many people use relatively low powered tube gear and are happy with it seems to corroberate this view, with full recognition of the fact that tube electronics in general demonstrate a much bigger headroom margin than most solid state gear. On the other hand, in my experience, these tube gear peaks, while impressive in voltage terms, all too often sound quite emasculated in current terms.

Personally, I'm into H/K's idea of solid current capability, although I doubt 500 Amps really means anything to anyone, unless one is driving something like the old Inifinity Refernce speakers, which were not very efficient (app. 86 dB/2.83V/1 m), with impedance dips down to below 2 Ohm and with a solid phase shift of +/- 40 degrees and more. Or Apogees, which also have a vile impedance modulus, no wonder they were called "The Amp Killers".

One last point - these high current claims made by H/K may look impressive, but are in fact at best half baked, because I have never seen any explanation by H/K as to exactly HOW these measurements were made. Most obviously, what excatly does "instantaneous" mean? 1 mS, 5 mS, 10 mS, 20 mS as in IEC standards, what?

Electrocompaniet has also made some wild claims, but in VERY small print, at the end of the ad, they did say for 1 mS only, which I find to be uselessly short, and thus consider the claim a marketing stunt.
 
Last edited:
...................I have read somewhere that on average, over 95% of us spend our listening time at AVERAGE power levels of less than 1 W, with peaks going up over that. This would imply that a man buys a say 100 WPC amp to make sure those peaks of say 20 Watts are reproduced faithfully, or more faithfully than a say 50 WPC amp could do (not necessarily true, but on average).
20W peak:1W avearge is +13dB peak:average ratio.

A +20dB peak:average ratio would ended up being severely clipped with a 50W amplifier, whereas just reaching clipping with the attendant increase in distortion using a 100W amplifier.
A 150W amplifier would pass the +20dB peak from an average level of 1W hopefully without any audible effect since the amplifier should be well within it's linear range for both current and voltage.

If we were listening at 2.4m with those 86dB/W @ 1m speakers then our average level would be ~ 81dB ref audibility @ 0dB.
 
Last edited:
And here's where the relativity of systems kicks in full scale - my speakers do 92 dB/2,83V/1m, are exceptionally easy to drive and operate in a room which is 15 x 12 ft (i.e. really small).

So, my job is relatively and absolutely easy, as opposed to John and other production engineers, who have no idea what their amps might be driving.
 
I recently have been building a SE tube amp of DIN 45500 standard . Well I've made a pigs ear out of it and now it only has 0.1 % distortion when it is supposed to have 1% . With care I can get it back to 1% and even 5 % is possible . I did have 13 Watts . That's a bit unlucky for us Brits so I settled for 8 watts . It is so minimal as to look impossible . Hum and sensitivity is descent ( - 87 dB @ 1 W ) . It is so unspeakable in places as to get me thrown out of any orthodoxy of design .

The point is that this amplifier can test on equal ground all the subspecies of common SE tube amps . It was inspired by the PYE Mozart yet now has no great resemblance to it . It can have a damping factor not too far off of ideal if I choose . Equally it can have the damping factor of typical SE tube amps . It only has two amplifying devices in the serial signal path and both are tubes/ valves . Sensitivity ranges from 0.5V to 1.2 V depending .

Now to the crunch point . My usual amplifier is loosely a Hitachi of 100 W . When running loud 5 watts is all I can measure . I just have to say the 5 watts of my Hitachi clone is somehow more powerful sounding than my SE amp . My felling is that bass is always present in music and my Hitachi clone will do 5 Hz and much better if asked . I suspect this matters greatly .

An example . Decca stereo demonstration record . The train whistle when in the station is magical via a 211 SE valve amp , however on my Hitachi the roof of the station can be heard and is a truly beautiful sound ( bass is very deep in the whistle ) . The link will give you no idea really . The old dark blue label is a valve cutting lathe and more collectable . The light blue reissues is a transistor lathe and to my ears better . The Fire-bird extract is fantastic .

Headroom both in voltage and current somehow can be sensed I suspect .

A Journey Into Stereo Sound - Side 1, Selections 1 - 5 - YouTube
 
Status
Not open for further replies.