Sound Quality Vs. Measurements - Page 720 - diyAudio
Go Back   Home > Forums > Member Areas > The Lounge

The Lounge A place to talk about almost anything but politics and religion.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 18th September 2012, 08:09 AM   #7191
diyAudio Member
 
Join Date: Mar 2007
Location: Oxfordshire
DVV . I am saying feedback is always good and global the very best . Now I know that will ruffle a few feathers . The reason " none " of us use large amounts of global is we find it passes an optimum point .

A power amp mostly these days is an op amp . Why should a power op amp be different ?

I have found having very high feedback around what is very like an op amp with a current buffer outside the loop works well . The output stage would have a reasonable damping factor as you say with a few output pairs . I realize you prefer a slightly higher output emitter resistor than Douglas Self . He says 0R1 is not dangerous . Myself I would use 6 x 0R1 and 0R47 output . The 0R47 would also drive my protection circuit .

I noticed sometime ago that MOSFET's ( Exicon 10N/P20 ) are happy with very high feedback factors . This helps them produce lower distrotion than some think . They have high Ron which can be reduced by multiple pairs . These often need no source resistors for balance ( Exicon supply matched sets free of charge ) .

I remember years ago discussing racing car engines with a Williams engineer . He said the Honda engines were fitted with a handle . Each passing engineer was asked to turn the handle . This prevented the valve springs from taking a set . We discussed Desmodromic valves as these would go to perhaps 14 000 RPM . Now the F1 cars use Air driven Desmodromics ( Ducati had them for years ) . Old Mr Honda thought any engine that revved to 20 000 RPM past a magic point where the air entering the engine would almost seem as in a jet engine . I think that is where we are with feedback . The valves are too slow ?
  Reply With Quote
Old 18th September 2012, 09:14 AM   #7192
dvv is offline dvv  Serbia
diyAudio Member
 
Join Date: Sep 2004
Location: Belgrade, Serbia
Disagreed on global feedback. I would never call it the very best. Necessary - yes, beneficial if moderate - yes, but "best"? Never! I think it was pronounced that way by factory engineers who simply had no time and no resources for research into the local vs global debate, so the easiest method available to them was to use high amounts of global feedback. In its heyday, I have seen global NFB figures of 60...80 dB. To me, that's called rape.

On the other side of the spectrum, we have the zero global NFB camp. In all my years in audio, I have yet to hear an amp, no matter whether solid state or tube, which does not sound a bit loose and which does not leave me with a feeling of a job not quite finished.

Also, over the years, I have closely followed Harman/Kardon's work and models and I find they sometimes get the balance between local and global feedback just right. The problem why so many of their otherwise very good designs fall short is their philosophy of using unbelievably poor quality pots, which seem to gobble up any and all benefits of otherwise good designs. I still have trouble believing just how much better my own HK 6550 sounds once I dumped their awful volume pot and substituted it for a regular ALPS Blue pot - believe me, it's really hard to believe, but it amply proves the point that the devil is in the details. EVERYTHING matters.

Anyway, that's a nominally 50/70 WPC 8/4 Ohms SEPP design, which uses only 17 dB of global NFB, yet manages to box way above its price class. Now, after the pot change (and capacitor change, using HK's Korean made caps sold as regular spare/service parts), it truly majors on ambience and ambient detail as I have never heard a zero global NFB device do. Of the ones I have heard, of course, which - oddly enough - were all to the one more expensive than that HK, some by several times over.

I believe that everything in life, including audio, is best served by a balanced approach between the two extremes. Even our own history teaches us that once balance is lost, odd and unpleasant things start to happen. Entire societies go sideways.

My own 26 dB global NFB mark is a very subjective one, I have built models which worked better with both less and more, but the excursion sideways was never more than +/- 6 dB, so I take 26 dB as the central hinge point, not as God's own truth carved in stone. Audio electronics are fun precisely because there are no design absolutes, it always boils down to try it and see/hear. Of course, you can sometimes bend some rules, but mostly the basic rules hold true each and every time.

This applies to output impedance as well. It's not how big it is - I think once you go over 20:1, there is little (if anything) to be gained in real world terms, but much in commercial terms, because the consumers have been brainwashed over the last few decades to believe in extremely high damping factors.

But, here we come to a crossroads. Two things. The first is HOW was that DF obtained, by making the amp with a good DF in open loop or by feeding it inordinate amounts of global feedback. The second is how does it stand in comparison with what you see as the minimum load impedance your amp must be able to negotiate without losing control. In my case, I see 2 Ohms as the minimum it must be able to work steadystate into, but 20:1/2 Ohms is 40:1/4 Ohms and 80:1/8 Ohms. Arguably, one could halve all these figures and still be just fine, but as ever, I like working with a healthy safety margin. With Wayne, double all the figures, others may be quite happy with half values.

I use DC Servo for the simple reason that it allows me to avoid capacitors in the signal path, and I do my best to get rid of that output inductor as well. In short, I try hard to make my signal path as clean as possible. And now I should put in series resistors? Granted, there may well be loudspeakers out there which will perform better with those resistors, and while I will include the space and the holes for them, my sample will NOT have them.

As for emittor resistors, as far as I know they are mostly determined by temperature considerations and serve to avoid thermal runaway. At least, that's what National Semiconductor "Audio Cookbook" says. And I actually wore out my initial copy from the early 80ies to shreds, but fortunately, I recommended it to an acquaintance from Seattle, he bought it and loved it, and then bought a second copy which he mailed to me - good man, I'm very grateful. I imagine just about everybody here has at least seen it, and probably owns a copy.

The point is, D. Self's recommendation for 0.1 Ohm resistors is downright suicidal unless he expects the amp to work at low level power only, has incredibly large heat sinks well above the 20 lb mark for relatively low power levels, or has an aggressive protection circuit.

The lowest value I have ever seen in a mass produced product is 0.15 Ohms, and that in a Dual CV1700 integrated amp only - BUT, in it, a SEPP design using Motorola 250W TO3 output trannies for nominally 80/120W into 8/4 Ohms, EACH transistor has its own dedicated heat sink, or in other words, it has an unusually beefy cooling system for a series product (and while it wasn't cheap in its time, it wasn't very expensive either). Typically, TOP-3 modern power transistors use emmitter resistors with values ranging from 0.22 to 0.47 Ohms, but I have NEVER seen anyone dare to go below 0.22 Ohms. For why, well, look at their thermal data and the heatsinks they use and you can work it out yourself. I find that many Japanese manufacturers tend to be VERY optimistic in this respect; all too often, push their amps a little harder and the overheat protection triggers.

So, good luck to D. Self with his logic, but he has been involved mostly with highly commercial products, where every penny counts, and that logic is obvious here and there in his reasoning.

MOSFETs are a completely different story, and I am not the one to tell it.
__________________
Per Aspera Ad Astra.
  Reply With Quote
Old 18th September 2012, 12:06 PM   #7193
diyAudio Member
 
Join Date: Mar 2007
Location: Oxfordshire
I think 26 dB feedback is a very good level .

I have not seen less than 0R22 is any amplifier I have repaired .

I still feel that global feedback is best . I agree it has to be used with caution . I have to say my oscilloscope spots problems before my ears or speakers do thank goodness .

I think given a damping factor of 6 there is an opportunity to use high global feedback to the driver stage and no global feedback to the output current buffer . I have built this amplifier and enjoyed it . In the end I had a little global feedback ( 1/7 of the total ) . It started life as a OPA 604 and Complimentary Feedback Pair output stage . I then designed a driver stage . The OPA 604 can be configured as a unity gain stage without problems . I didn't scale this up above 8 V rms 4 ohms . I see no reason why it shouldn't be universally true . 16 Vrms 4 ohms is possible in bridge .

BTW , no criticism of usual practice . One day devices might come along to change how we think . As said MOSFET's do seem happier with high feedback ( and nice ) .
  Reply With Quote
Old 18th September 2012, 03:44 PM   #7194
dvv is offline dvv  Serbia
diyAudio Member
 
Join Date: Sep 2004
Location: Belgrade, Serbia
Although, I must say, my experience shows that what Lavardine stated in their patent application, is generally (mostly) quite true.

In essence, they claim that NFB is good and you should apply it liberally, provided you previously did everyting you could to minimise distortion, have at least some local degeneration and that your open loop bandwidth is 20 kHz or more.
__________________
Per Aspera Ad Astra.
  Reply With Quote
Old 18th September 2012, 03:49 PM   #7195
dvv is offline dvv  Serbia
diyAudio Member
 
Join Date: Sep 2004
Location: Belgrade, Serbia
BTW Nige, was it myself who talked you into using Burr-Brown's OP604/2604?

I ask because that's the only BB op amp the sound of which I find very pleasing, the rest which I have tried I find either screechy, or dull, including their most touted models, like the 627/637. To me, they either screech at you, or are very bland and sound totally uninterested in working at all.

I could almost say that the 604/2604 is a misbegotten renegade in their ranks.
__________________
Per Aspera Ad Astra.
  Reply With Quote
Old 18th September 2012, 04:04 PM   #7196
bcarso is offline bcarso  United States
diyAudio Member
 
Join Date: May 2005
Location: Canoga Park, California
Quote:
Originally Posted by dvv View Post


The point is, D. Self's recommendation for 0.1 Ohm resistors is downright suicidal unless he expects the amp to work at low level power only, has incredibly large heat sinks well above the 20 lb mark for relatively low power levels, or has an aggressive protection circuit.

The lowest value I have ever seen in a mass produced product is 0.15 Ohms, and that in a Dual CV1700 integrated amp only - BUT, in it, a SEPP design using Motorola 250W TO3 output trannies for nominally 80/120W into 8/4 Ohms, EACH transistor has its own dedicated heat sink, or in other words, it has an unusually beefy cooling system for a series product (and while it wasn't cheap in its time, it wasn't very expensive either). Typically, TOP-3 modern power transistors use emmitter resistors with values ranging from 0.22 to 0.47 Ohms, but I have NEVER seen anyone dare to go below 0.22 Ohms. For why, well, look at their thermal data and the heatsinks they use and you can work it out yourself. I find that many Japanese manufacturers tend to be VERY optimistic in this respect; all too often, push their amps a little harder and the overheat protection triggers.

So, good luck to D. Self with his logic, but he has been involved mostly with highly commercial products, where every penny counts, and that logic is obvious here and there in his reasoning.
Just to throw in an anomaly here: my ancient NAD 3020 afaik has NO emitter resistors, just big heat dissipators for each TO-3 device (2N3055, MJ2955?) for the very low rated power. There is a thermal cutout switch, which has never tripped for me. And the only thing that has gone wrong with the power amp section over the substantial number of years is one of the auxiliary rail supply bulk caps failed open. The main bulk caps are still fine, despite my being assured by someone in here that the adhesive that affixes them to the PCB is corrosive and must be removed as soon as possible.

Of course the cheap switches and pot started failing ages ago, so I just bypassed them using the convenient direct inputs in the back. And I use the piece nowadays to drive low-sensitivity headphones, connected directly to the speaker outputs.

It sounds good, although my expectations are low
  Reply With Quote
Old 18th September 2012, 04:36 PM   #7197
dvv is offline dvv  Serbia
diyAudio Member
 
Join Date: Sep 2004
Location: Belgrade, Serbia
Quote:
Originally Posted by bcarso View Post
Just to throw in an anomaly here: my ancient NAD 3020 afaik has NO emitter resistors, just big heat dissipators for each TO-3 device (2N3055, MJ2955?) for the very low rated power. There is a thermal cutout switch, which has never tripped for me. And the only thing that has gone wrong with the power amp section over the substantial number of years is one of the auxiliary rail supply bulk caps failed open. The main bulk caps are still fine, despite my being assured by someone in here that the adhesive that affixes them to the PCB is corrosive and must be removed as soon as possible.

Of course the cheap switches and pot started failing ages ago, so I just bypassed them using the convenient direct inputs in the back. And I use the piece nowadays to drive low-sensitivity headphones, connected directly to the speaker outputs.

It sounds good, although my expectations are low
You said it yourself - anomaly.

That amp was made to work at very low power levels indeed, at least relative to its output power transistor capabilities. I am willing to bet that if it was asked to work at any greater power level for any period of time, it's thermal protection would soon cut it off.

Think, Brad - if you want your amp to be capable of large power delivery into lower impedances, say sub4 Ohms, you will necessarily have to assume pretty large power dissipations, like it or not. After all, say 10 W contiuous into say 3 Ohms is about the equivalent of 3.7W into 8 Ohms, same voltage but more current, and hence more heating of the output stage. Where is this heat to go? Into flimsy U shaped heat sinks? Not for long, trust me.

There was a time, say late 70ies and early 80ies, when the industry worked on the assumption of sub 1W continuous power dissipation in most homes, and thus skipped the tedious work of installing anything that looked half serious in terms of heat sinking. Also saved costs. Even the presumably better makers, such as for example Kenwood, who used simple U shaped aluminum tunnels for heat sinks, right after installing twin power transformers for a true dual mono operation (e.g. models 6100, 7100, etc), played the low power average continuous game.

No offence Brad, but NAD have always gone for the El Cheapo approach in those days, so I am not surprised at such a turn of events. But try finding a Sansui, or a Pioneer amp from those days, including the lowest models in the range, and you will see proper heat sinking and proper emitter resistors without fail.
__________________
Per Aspera Ad Astra.
  Reply With Quote
Old 18th September 2012, 06:21 PM   #7198
bcarso is offline bcarso  United States
diyAudio Member
 
Join Date: May 2005
Location: Canoga Park, California
Quote:
Originally Posted by dvv View Post
You said it yourself - anomaly.


No offence Brad, but NAD have always gone for the El Cheapo approach in those days, so I am not surprised at such a turn of events. But try finding a Sansui, or a Pioneer amp from those days, including the lowest models in the range, and you will see proper heat sinking and proper emitter resistors without fail.
No offense taken of course. I was frankly shocked when I saw the schematic. And it was indeed cheap, but a rather clever cost-optimized design for which the designer, Bjørn Erik Edvardsen, was justly celebrated. And of all the unlikely people to recommend it, the inveterate innovator Keith O. Johnson of Reference Recordings, Spectral, et al., who told me he was tired of not hearing any music since his own amplifiers had died and he didn't have the time to repair them. So he got one, used, from a friend, and in KOJ's words, "I didn't have to do anything to it to make it sound right".
  Reply With Quote
Old 18th September 2012, 07:42 PM   #7199
dvv is offline dvv  Serbia
diyAudio Member
 
Join Date: Sep 2004
Location: Belgrade, Serbia
Quote:
Originally Posted by bcarso View Post
No offense taken of course. I was frankly shocked when I saw the schematic. And it was indeed cheap, but a rather clever cost-optimized design for which the designer, Bjørn Erik Edvardsen, was justly celebrated. And of all the unlikely people to recommend it, the inveterate innovator Keith O. Johnson of Reference Recordings, Spectral, et al., who told me he was tired of not hearing any music since his own amplifiers had died and he didn't have the time to repair them. So he got one, used, from a friend, and in KOJ's words, "I didn't have to do anything to it to make it sound right".
With that, and in those days, I have no argument.

However, that model was NAD's star product, which propelled them to glory and vast sales. For the time, it WAS better than most, and for the price, it ruled.

Let me put it this way - it was a great deal for the price.

Soon enough, others came along (e.g. Rotel, Proton, Vector Research, etc) and did similar things, but on a higher level, both in terms of power output AND general performance, so it was eclipsed soon enough. To be sure, NAD could always fall back on its receivers, which were also rather good at the time (e.g. model 160), but their first foray into higher performance, higher cost products, in form of a 90 WRMS/8 Ohms integrated amp, with Dolby coding/encoding included on board, didn't outlast the year it appeared in.

There was nothing wrong with it, it had no obvious or even less obvious faults, the problem was it had no obvious or less obvious aspects to recommend it either. A very much "me too" product, about the same as the average of its power and price group.

To be fair to NAD, others were trying hard with the same formula as well, for example, Leak tried with a biggie integrated amp as well (actually custom manufactured by Rotel), I think it was model 3900 or some such, but didn't make it either. The Japanese were far too deeply entrenched.

For the next giant killer, we had to wait another 13-14 years, until Pioneer and Marantz did a similar minimalist design in the early 90ies. Again, both were good for the time and price, but were still quickly forgotten anyway.
__________________
Per Aspera Ad Astra.
  Reply With Quote
Old 21st September 2012, 06:42 AM   #7200
diyAudio Member
 
Join Date: Mar 2007
Location: Oxfordshire
Quote:
Originally Posted by bcarso View Post
No offense taken of course. I was frankly shocked when I saw the schematic. And it was indeed cheap, but a rather clever cost-optimized design for which the designer, Bjørn Erik Edvardsen, was justly celebrated. And of all the unlikely people to recommend it, the inveterate innovator Keith O. Johnson of Reference Recordings, Spectral, et al., who told me he was tired of not hearing any music since his own amplifiers had died and he didn't have the time to repair them. So he got one, used, from a friend, and in KOJ's words, "I didn't have to do anything to it to make it sound right".
Strange thing is Proton who made them tried to market an improved version when their contract was up . It was a pale imitation . 3020 would drive 2 ohms and attempt 92 watts briefly ( Hi Fi Choice test ) . I was told 3020 is full of junk that was selected carefully . Other amps are filled with treasure that often doesn't work . 2 x 3020's is ideal . One as pre and the other as power amp .

A new Germanium device !
Contouring Vest With Germanium Fabric 82% off
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Quality Control differences = variations in sound quality? KT Class D 3 4th June 2014 12:02 AM
Sound Card for Measurements Marik Solid State 2 2nd January 2012 08:59 PM
Sound Card Recommendations (For Audio Measurements) dchisholm Equipment & Tools 5 16th July 2011 09:40 AM
How to protect sound card during amp measurements? okapi Everything Else 13 2nd September 2008 03:06 PM
Sound cards - test and measurements jackinnj Everything Else 2 5th July 2003 03:02 PM


New To Site? Need Help?

All times are GMT. The time now is 06:53 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2