News to me - but for a change, good news to me.
I think it's safe to say that really good audio and valid attemps to do something new and better today lives in small companies at the outskirts of the big name industry, on the fringe, so to speak.
Perhaps I am being unfair to the big names, but as far as I can see, the vast majority of them all concerned on how to make it smaller, more portable and MP3/WiFi/MPEG compatible.
How did you measure the color?
Was it a scientfic test, or just fooling around?
Which spectrometer did you use and why?
How big are the phase shifts of the three prime colors?
Is the wife frowning or clowning?
Is price a painful memory?
And I was wondering if "one in the pocket" was a universally understood safety practice! OSHA would never approve.
Hi,
Yes.
Same Greek Tragedy about the Angel Thesis and the Dæmon Antithesis being acted out, with never a resolution, just like a stuck record.
Same Dramatis Persona as well. With a few additions, but few deletions.
Hence my regular appeal to leave all these antique arguments behind and move on.
Yet we all seem remain attached to the argument like a bunch of bulldogs to a T-Bone Steak. Even those who wish to get past this cannot, due to the Tar Baby principle.
And the warnings on the map "Yonder Way Be Monsters" is ignored too by "newbies"...
Ciao T
Yes.
Same Greek Tragedy about the Angel Thesis and the Dæmon Antithesis being acted out, with never a resolution, just like a stuck record.
Same Dramatis Persona as well. With a few additions, but few deletions.
Hence my regular appeal to leave all these antique arguments behind and move on.
Yet we all seem remain attached to the argument like a bunch of bulldogs to a T-Bone Steak. Even those who wish to get past this cannot, due to the Tar Baby principle.
And the warnings on the map "Yonder Way Be Monsters" is ignored too by "newbies"...
Ciao T
Last edited:
Dvv, you are correct in that the controversies debated here are virtually the same as the controversies we debated 30 years ago. However, let me give you just one example of how something has actually progressed (if people will allow it) that we virtually ignored (on both sides) back in 1979, for example.
This was high speed vs low speed diodes for 60Hz operation, and, or, the use of snubbers in typical audio products. In those days we selected a rectifier bridge, already potted, and thought it enough. Yet I have found that changing out the low speed diodes for fast-soft recovery diodes made both measured and listening differences. (If you know what and how to measure for the problem)
Slowly we move forward, as the inexperienced are finally being shown such a change and finally it becomes 'obvious'. I will save the 'history' of how this particular breakthrough came about for another time.
This was high speed vs low speed diodes for 60Hz operation, and, or, the use of snubbers in typical audio products. In those days we selected a rectifier bridge, already potted, and thought it enough. Yet I have found that changing out the low speed diodes for fast-soft recovery diodes made both measured and listening differences. (If you know what and how to measure for the problem)
Slowly we move forward, as the inexperienced are finally being shown such a change and finally it becomes 'obvious'. I will save the 'history' of how this particular breakthrough came about for another time.
Hi,
I also think it looks luxmannish, though they usually did a more interesting job.
Good points:
Tripple darlington output - this keeps non-linear VAS Loading down.
Hawksford Cascoded J-Fet input - rather wider range of linear input voltage than undegenerated bipolar, low noise, no non-linear input current (espeically if a volume control is before it) and very linear with high signal
Double Differential - moves the current mirror (to convert diff to se) problems to a higher level node where their relative impact is lowered.
Class A Driver Stage - at least only the output transistors switch.
Mixed Miller and lead/lag compensation - keeps the slewrate fairly high, but the heavy handed compensation of the tripple darlington put's a dent in that.
DC Blocking is implemented intelligently - allows high quality film cap's INSIDE the AC feedback loop, compared to a low quality cap OUTSIDE the feedback loop.
Not so good points:
No degeneration in the second stage - this means the VAS is not linearised, it helps slew rate (compromised by heavy handed capacitive loading) though, so may be the lesser evil. Then again, some degeneration may allow lower compensation cap's on the output and improved stage level linearity.
Primitive current mirror with minimal degeneration - underlying linearity is not good, probably, together with the un-degenerated VAS implemented also to limit drop-out voltage. A set of separate rails for the front-end a few volt higher than the output stage would introduce another degree of freedom, which may help.
Ciao T
Out of curiosity I've attached the schematic of an amplifier that I particularly like the sound of, which measures well with conventional measurements but I find sounds significantly better than many other similar spec amplifiers I've heard, and I'm quite interested in any comments from the likes of Wavebourn and ThorstenL on (a) whether they think the design is good or bad in their opinion and (b) whether its just a typical by the numbers "textbook" design or whether there is anything clever/unique/novel about it that might account for a subjectively pleasing result. (or is it all just in my head
Extra points for recognising/guessing what brand/model it is
I also think it looks luxmannish, though they usually did a more interesting job.
Good points:
Tripple darlington output - this keeps non-linear VAS Loading down.
Hawksford Cascoded J-Fet input - rather wider range of linear input voltage than undegenerated bipolar, low noise, no non-linear input current (espeically if a volume control is before it) and very linear with high signal
Double Differential - moves the current mirror (to convert diff to se) problems to a higher level node where their relative impact is lowered.
Class A Driver Stage - at least only the output transistors switch.
Mixed Miller and lead/lag compensation - keeps the slewrate fairly high, but the heavy handed compensation of the tripple darlington put's a dent in that.
DC Blocking is implemented intelligently - allows high quality film cap's INSIDE the AC feedback loop, compared to a low quality cap OUTSIDE the feedback loop.
Not so good points:
No degeneration in the second stage - this means the VAS is not linearised, it helps slew rate (compromised by heavy handed capacitive loading) though, so may be the lesser evil. Then again, some degeneration may allow lower compensation cap's on the output and improved stage level linearity.
Primitive current mirror with minimal degeneration - underlying linearity is not good, probably, together with the un-degenerated VAS implemented also to limit drop-out voltage. A set of separate rails for the front-end a few volt higher than the output stage would introduce another degree of freedom, which may help.
Ciao T
Yes, it was in fact a piece by Rees and Shaffer that started this, as the issue that had floated to the top of the take-home stack had a Reg Williamson amp on the cover, and within an article that ruminated on the letters exchange and forum I mentioned. It was titled The Saga of Golden Ear and Meter Reader (TAA 4/1979). Reading that led me back to the earlier issues.
And yes, I know that word processors (other than carbon-based lifeform ones) were nonexistent in those days, and someone had to set type. It still seems like something nearly invariant though. We now have auto-correct functions (that usually don't catch wrong words, just misspelled ones) and other utilities, but I still see gross errors everywhere. It is easy to do.
But I once read a PhD dissertation by the now-professor David Meyer, and he told me I would find no typos. He was correct! I think that's the only time it's ever happened to me.
Brad
DBMandrake;2912257 [snip said:Out of curiosity I've attached the schematic of an amplifier that I particularly like the sound of, which measures well with conventional measurements but I find sounds significantly better than many other similar spec amplifiers I've heard, and I'm quite interested in any comments from the likes of Wavebourn and ThorstenL on (a) whether they think the design is good or bad in their opinion and (b) whether its just a typical by the numbers "textbook" design or whether there is anything clever/unique/novel about it that might account for a subjectively pleasing result. (or is it all just in my head
Extra points for recognising/guessing what brand/model it is
Of course it is all in your head, but no less real
Thorsten has done all the heavy lifting here in his analysis, IMO, but I would add just a bit more about the bootstrapped-drains input stage*, namely that the JFETs have nearly constant power dissipation with signal swing, provided that there is plenty of gain following (and there is). The parts are monolithic duals (now unavailable
although some substitutes exist) which helps thermal tracking too --- but it's nice to have the dissipation constant with signal as well.Brad
*popularized by Hawksford but predated by a bunch.
A lot of dirty work in this field. Sad. I sometimes wonder how I can remain in the relative good graces of some of my friends, particularly the meter readers and double-blinders, but as well those who have an attitude that can only be described as theological, as I too try to preserve an open mind, while trying to decode what is "really" going on.
Agreed, but we also both know how few such diodes were at that time. The choice wasn't, shall we say, plentyful, was it?
Fortunately for us, the times have changed, and such diodes today are many, certainly enough to make one's choice quite reasonable in type, number and price.
Personally, having read what I read here, I am willing to give them a serious try, which I have not done so far. But I looked at the catalog of my local suppliers and I note an abundance of them - thus, I expect no problems.
You see, this is not the USA, locally I do not have available 20% of what you have over there, and when I do have it, I have at abonmibale prices. You can order Analog Devices op amps form Analog itself, and I can't because they won't ship over here. Same thing with Motorola/ON Semi samples batches, and this severely limits my choice in just about everything.
If I do find a part I'd like to have, instead of paying $3.80 you'd pay buying from AD, I am asked to pay the equivalent of $28 for it. And people wonder why are US made devices shrinking in sales volume - as opposed to them, I can buy a Toshiba 2SC5200/2SA1943 pair for almost the same as you'd pay for them in the US, maybe sa 5% more expensive.
My point is, these variances in purchase prices also influence how much work is going on around the world. If my test bed initially costs twice what yours costs, but my salary is like 1/8 of yours, how likely am I to experiment much? I believe this dichotomy is a part of the answer why there is so much less reserach and development going on - the prices have escalated to unbearable levels.
So I get sound advice (sorry) from the experienced folks here, whip open my trusty Mouser or Digi-Key catalog and find the modern parts I am looking for are not stocked, 6 month lead time, and 5000 minimum order. THAT is a solid reason it experimenting harder. A case in point was just trying to fine some of those wonderful high speed soft recovery diodes. I had to buy the 35A bridge from England for the DH-120. Including shipping, it cost almost 30 bucks. Where is the supply that I could use to replace the dozens of generic 1 to 4A generic 100x series used in everything? Not a stocked item..... Where did I have to go to get the MOSFETS for the output? Again, England.
10 years ago you could at least get overpriced bad quality from your local Radio Shack. Not any more.
10 years ago you could at least get overpriced bad quality from your local Radio Shack. Not any more.
That's plain silly... There is huge variety of fast/ultrafast/Schottky devices on avail.
Unless...
I used to know the man, who was purchasing all hardware from the country you've mentioned to support his fatherland, bolts and nuts included.
Could you be more precise, what are the parts in question?
I stand corrected.
I was obviously a little too personal. I hadn't even heard of them almost 20 years after you.
On hindsight, I just plain hate to remember how many things I should have known, but didn't. Or not fully.
Like my grandma used to say - we learn all our lives, and still die stupid.
In some ways, this strongly reminds me of a discussion I once participated in, some 6 or 7 years ago on another forum, related to op amps used for DC servos.
One group maintained it was best to use op amps with FET inputs, and said the trusty workhorse TLO series was just fine.
Another group maintained that even if the circuit was dealing with frequencies below 3 Hz, and often below 1 Hz, speed was of essence, and the TLO family was just not quick enough.
A third faction thought the op amp's own drift characteristics were of prime importance, so one had to choose the likes of OP 97, etc.
My view was, and still is, that while speed and input FET stages were indeed very useful, it was the settling time which would be best to look at. To that effect, I supported Analog Devices, who at the time was just about the only manufacturer outside the pro arena who had settling times of less than 100 nS, versus 1.5...2 uS of the pack. Just 20 times faster, no big deal. Yet they consistently offered a better sound overall, in my view, especially if one offloaded them by adding a discrete current boost stage (in other audio stages, not as a DC servo).
Because of the local parts supply problems, today I use LF411 for DIY purposes, because they are readily available and their own drift is like 2 uV per annum. In things I do strictly for myself, I'll probably use AD 818. It seems to me I get a slightly better sound that way, but it could be a placebo effect.
One group maintained it was best to use op amps with FET inputs, and said the trusty workhorse TLO series was just fine.
Another group maintained that even if the circuit was dealing with frequencies below 3 Hz, and often below 1 Hz, speed was of essence, and the TLO family was just not quick enough.
A third faction thought the op amp's own drift characteristics were of prime importance, so one had to choose the likes of OP 97, etc.
My view was, and still is, that while speed and input FET stages were indeed very useful, it was the settling time which would be best to look at. To that effect, I supported Analog Devices, who at the time was just about the only manufacturer outside the pro arena who had settling times of less than 100 nS, versus 1.5...2 uS of the pack. Just 20 times faster, no big deal. Yet they consistently offered a better sound overall, in my view, especially if one offloaded them by adding a discrete current boost stage (in other audio stages, not as a DC servo).
Because of the local parts supply problems, today I use LF411 for DIY purposes, because they are readily available and their own drift is like 2 uV per annum. In things I do strictly for myself, I'll probably use AD 818. It seems to me I get a slightly better sound that way, but it could be a placebo effect.
Hi,
I though those where essentially the same.
Blind faith in double blind tests or meters can only be considered a religious cult, complete and replete with it's own popes, anti-popes, the holy inquisition and the occasional heretic.
I would suggest that while blind in their own way, the listeners, that is the actual empiricists, are by far less religious and by far more tolerant (I may not appear tolerant, but that is strictly self- and community-defence and I am neither listener, meter reader or double blinder, I'm bit of each and grouchy in my old age.
If you snipe at me, remember I trained with .57 cal semi-automatic "anti-sniper" rifles (the russian Simonov) in my youth and Dejan may have friends that actually saw action using these. So expect a supersonic, fat and heavy slug or two with pretty good aim coming your way on short notice, if your shot kicks up the sand near me.
In principle the Church of Meter Readers and the Church of the Double-Blinders are welcome to hold their services and worship their particular totems just as any other group. I even happily accept that if I'm in their church I will worship the mighty AP2 and ABX Comparator (or just stay outside) and abide by their rules.
I do start to take exception if they insist that outside their churches, on public, communal side walks and plaza's their totems must also be worshipped and any expression of dissent must first pass their test with their totem before one is allowed to make it and in a quite uncivil way.
Ciao T
I though those where essentially the same.
Blind faith in double blind tests or meters can only be considered a religious cult, complete and replete with it's own popes, anti-popes, the holy inquisition and the occasional heretic.
I would suggest that while blind in their own way, the listeners, that is the actual empiricists, are by far less religious and by far more tolerant (I may not appear tolerant, but that is strictly self- and community-defence and I am neither listener, meter reader or double blinder, I'm bit of each and grouchy in my old age.
If you snipe at me, remember I trained with .57 cal semi-automatic "anti-sniper" rifles (the russian Simonov) in my youth and Dejan may have friends that actually saw action using these. So expect a supersonic, fat and heavy slug or two with pretty good aim coming your way on short notice, if your shot kicks up the sand near me.
In principle the Church of Meter Readers and the Church of the Double-Blinders are welcome to hold their services and worship their particular totems just as any other group. I even happily accept that if I'm in their church I will worship the mighty AP2 and ABX Comparator (or just stay outside) and abide by their rules.
I do start to take exception if they insist that outside their churches, on public, communal side walks and plaza's their totems must also be worshipped and any expression of dissent must first pass their test with their totem before one is allowed to make it and in a quite uncivil way.
Ciao T
Correct.
It's an RV-371 from, I think, around 1993 which was during the period when Alpine owned them, how much that influenced their designs at a component level compared to pre and post Alpine days I'm not sure. The only other Luxman I've used extensively was the L-309 from 1974 which I also liked a lot, although I'm sure by today's standards the design is fairly prehistoric. Since there are paralleled pairs of output transistors, (eg Q730/Q734) when you say triple Darlington are you referring to all the way back to Q722 as being part of the Darlington triple ? What are the drawbacks of so many Darlington coupled stages - mainly loss of potential output voltage swing ?
Interesting. I recognised the cascode stage but wasn't sure what its benefit or purpose might be at audio frequencies. My main exposure to cascode stages is in RF applications where its used to extend bandwidth by eliminating effects of miller capacitance.
Noise level of the amplifier is indeed very low - with 94dB/W/M speakers I have to put my ear within a couple of inches of the tweeter before I can hear any hiss at all. (The spec sheet attached claims 112dB SNR for the power amp alone, and 100dB including all preamp and input switching stages - I only use it as a 2 channel amp even though its an AV receiver, so I'm only looking at Front L/R specs)
Which transistors do you consider to be the Class A driver stage ? Q726/728 ? Q722/724 ? Or all 4 of them ?
Are you referring to C722/724 ?
Ok I'm a little bit confused as to which stage you're referring to as the second stage, do you mean Q712/714 ? By no degeneration do you mean because their emitters are tied directly together ?
So you mean Q716/718, their 100 ohm emitter resistors are too low ?
Thanks for the feedback, to my surprise I think I understood most of what you said
(I should know some of this stuff, but sadly its been a few years since I was heavily hands on in electronics...)Interesting comments on the monolithic dual jfet.Of course it is all in your head, but no less real
Thorsten has done all the heavy lifting here in his analysis, IMO, but I would add just a bit more about the bootstrapped-drains input stage*, namely that the JFETs have nearly constant power dissipation with signal swing, provided that there is plenty of gain following (and there is). The parts are monolithic duals (now unavailable
So are you saying that the 2SK389 is no longer available at all as a spare part (gulp) or just that it's not available for bulk orders for new designs ? And there aren't many modern alternatives ? So do not many audio amplifiers go to the trouble of having a monolithic dual for the first differential input stage ?
Attachments
Last edited:
"The charlatan and the boob are both intrinsically opposed to a search for truth; the last thing they want is public verifiability and controlled experimentation. In fact, when outsiders crassly insist on such tests and the results fail to support the claims- the non-scientist calls on a marvelous array of excuses as to why the uninitiated have perverted their domain."
- Fred Gruenberger "A Measure for Crackpots" Rand Corporation 1962
BTW, I remember the Rod Rees articles. When they first came out, I was a very young fellow, couldn't figure out what he was talking about, but figured that one day I would be able to figure it out. Like bcarso, I ran into them again recently and reread them. Now I know what the problem was: they were the worst sort of philosophical gassiness. There really was nothing to understand there.
- Fred Gruenberger "A Measure for Crackpots" Rand Corporation 1962
BTW, I remember the Rod Rees articles. When they first came out, I was a very young fellow, couldn't figure out what he was talking about, but figured that one day I would be able to figure it out. Like bcarso, I ran into them again recently and reread them. Now I know what the problem was: they were the worst sort of philosophical gassiness. There really was nothing to understand there.
- Status
- Not open for further replies.