Sound Quality Vs. Measurements

Status
Not open for further replies.
wDX5mBf.jpg


> 10 years ago I doodled this when by a small river in Texas . The output stage although simple is reasonably free of problems . If run hot almost none . This was about the last time I could do any handwriting . Forgive any errors ( i.e R1 C1 position is inverted ) . A simple bias system would be fine if class A to drive 600R at 10 Vrms . The 47 R VAS resistor might not be my choice now .

I dare say built with BD 139/140 it would be great as a pre-amp . The parts were from the parts bin ( 0R25 not 0R22 ) .

The 0.1 uF and 220 uF were if reconnecting to another PSU . The RAW DC about 2V above the regulated at 230 V - 10% .

The sound is ultra sweet so might suit well . How you do the input stage is the fun of it . Input JFET LTP and bipolar second LTP might have virtue . What JFET is easy to get and likes 60 V ( other uses ) ?
 
PlNVXff.jpg


This is the amp most Brit amps follow . Bob ( J R Stewart ) only gave it as a generic form for discussion . The 1969 Sinclair was similar . Bob actually shows how to anticipate and cure the ills of this amp . I suspect the Philips for all of it's sophistication was not working as well as the fully refined version of this amp . I met Bob about 3 years after this . How he got so far so young I will never understand .

Bob's big mistake was talking about it whilst still with Lecson . He convinced others to get on and make it .

The diode IS920 slightly improves the compromise when compared with complimentary outputs . It is called a Baxendale diode by some . There is still some virtue in this method using all NPN's . The switching is never as clean as complimentary . That also goes for feedback pairs .
 
I thought I had made it clear, but it seems not.

Wayne, while indeed I am far from shy, I TRY to be reaslistic, with varying results.

Sure, I know a lot about power amps compared to a typical DIY hobbyist, but that pales a lot compared to people like Demian, Pavel and John. I am here to learn and, with any luck, pick up a few pointers like Demian made a few posts ago. If I was up to SOTA standards, he wouldn't have had to make them.

They are the professionals in the SOTA range, I am at best a well informed amateur.

In addition to which my requirements are vastly different to yours, far less demanding. As a matter of fact, you are the one and only person I know of who has 1 Ohm speakers.

D,

every discussion we have you revert back to my requirements, its obvious you are not getting what is being said , this is nothing to do with me , we are discussing power amplifiers in general and paper tiger amplfiers don't sound as well as none paper tiger amplifiers , regardless of 8/4/2 load .

I pointed you to an article discussing why ..
I pointed to a show situation where amps were shutting down on specific speakers.


Everything posted was only for discussion , nothing to do with me or any of my requirements, for discussion , Demian brought up a good case For sufficient vas when driving low Z loads.

So Nothing to do with your Amp philosophy or target goals , there are many guru's here who stuff i have tried and parted ways , sonics is very subjective , what isnt and is not allowed here is toys , paper tiger amplifiers loaded with dynamic killing protection BS, is shown The door , they suck all the life out of the music , nothing is right . Biggest offense, They don't size , everything is pleasant and lilliputian, toobs for example get this , really good SS does too with more energy, I know most here don't get it , most seems to listening to electronic dynamic limited music. It sounds good to them , doesn't work for me and others who understand what I'm saying .

So believe me , I'm not discussing "My" requirement D...... :D
 
Last edited:
The compl-diff was independently invented by J. Curl, as he stated. The DC servo concept as a 'solution' as we know and use it today, was invented by myself and first discussed with Walt Jung. We had just published some work on capacitors in Audio magazine and I suggested eliminating the coupling cap... I had done so and it sounded better. I used the word 'servo' to Walt as it was used and applied to other fields of science. Walt's phono stage with his dc servo design implimentation first appeared in TAA.

These are old subjects of historical interest now and well documented.

THx-RNMarsh

Thanks for the memory trip richard .....

Well, the multiplication table is simple enough. 100W/8 Ohms is 200W/4 Ohms is 400W/2 Ohms. 100W/8 Ohms is 5 A peak, is 10A peak into 4 Ohms is 20A peak into 2 Ohms.

If I want 20A peak into 2 Ohms, I would need quite a few output transistor pairs to handle that safely on any longer term basis. I agree 2 mS is ridiculously short, a literally better something than nothing "solution". As I said before, 50 mS is the shortest interval I'd consider as realistic.

Just for memory's sake, I'd like to take you back to that German made amp, LAS Mega 1, designed in 1978. The reason I mention it is because those guys had that problem licked even then. They used 1.5 Ohm/17W emitter resistors to make sure the output trannies, 4 pairs of BD 249/250 C (rated at 100V, 125W, 25 A continuous, 40A impulse) never COULD be asked to deliver more than their specs indicated. The whole amp was rated at 150W/4 Ohms and 100W/8 Ohms, with a nice 100 V/uS slew rate, and excpet for one transistor, used off the shelf parts. Unfortunately, in those days (1979), nobody thoght of testing them for 2 Ohm operation. But I bet it could stand up well to that test, and way better than most in its price class.

The point is, I've been considering low impedance aspects ever since then, and that's ben 35 years now. My views, right or wrong, have been put to quite a few tests since then, so I'm not talking off the top of my hat. I maintain that we need to have an amp capable of working into 2 Ohms, but not necessarily on a continuous basis, so long as the limitation to continuous basis are simply cooling considerations. No point chasing something that will quickly burn out one's output stage.

It is my feeling you do not quite see this and similar limitations. One needs to balance the whole design. To make the 2 Ohm continous mark, one needs to:

1. Use DAMN big heat sinks, and a lot of them,

2. Use several output device pairs, which is a hornet's nest in itself to make them work in unison,

3. Use forced air cooling (fans) to control the temperature rise, or

4. Use Class D amps.

I mentioned, and now I mention it again, some weeks ago, a friend and I really powered my system to uncomfortbaly loud levels, approximately 95-96 dB SPL at 3 m, which means 101-102 dB SPL at the usual 1m distance. The power amp PEAKED at just over 70W/8 Ohms nominal, cleanly riding out anything we threw at it. The key point - it never changed its overall tonal balance while doing so. To me, that spells a very good amp because that is a hell of hard act to pull off. Both the Marantz 170 DC and the Citation 24 did it no problemo, as did the Karan integrated, while quite expectedly, the small Philips amp choked.

Now, if my speakers were not 8 but 4 Ohms, there's no telling how it might have worked out. I suspect the Marantz would have started to seriously struggle, but not so the Citation, given that it has a dual mono topology and an effecticely better and stronger power supply system overall.

That's the problem with sweeping statements, they lose all touch with reality, which is very widely varied. In my case, this works in my favor, and in your case, this works against you because it tends to reduce your particular requirements.

Sweeping ..

While not an amplifier designer , i do have an understanding of what is necessary. My custom built SS amp in 78 could drive 2 ohm all day from it's fully regulated and separate chassis supply .

:)
 
wDX5mBf.jpg


> 10 years ago I doodled this when by a small river in Texas . The output stage although simple is reasonably free of problems . If run hot almost none . This was about the last time I could do any handwriting . Forgive any errors ( i.e R1 C1 position is inverted ) . A simple bias system would be fine if class A to drive 600R at 10 Vrms . The 47 R VAS resistor might not be my choice now .

I dare say built with BD 139/140 it would be great as a pre-amp . The parts were from the parts bin ( 0R25 not 0R22 ) .

The 0.1 uF and 220 uF were if reconnecting to another PSU . The RAW DC about 2V above the regulated at 230 V - 10% .

The sound is ultra sweet so might suit well . How you do the input stage is the fun of it . Input JFET LTP and bipolar second LTP might have virtue . What JFET is easy to get and likes 60 V ( other uses ) ?

You have this up and running Nige .....
 
Please ask WJung. I have a copy of the TAA article but it is on the other side of the planet now. As I recall, there wasnt much made of it in that article but it later produced other articles and detailed math analysis of that first one and variations of the topology. I dont remember that author. (also in TAA archives somewhere). It may helpful to you.

[J.Curl was privy to our discussions about this and has also first person knowledge of the events. John, Walt and I were all close at this period in time.]

THx-RNMarsh

Will do, Richard, thank you.
 
Wayne, so what exactly are we discussing?

If it's that amps should be able to drive 2 Ohm loads on a continuous basis, I disagree as I see no practical point to it if the amp is rated at 100W/8 Ohms nominally and if it is to be used for home listening.

It may make sense only if the loudspeaker load is criminally badly behaved, but I mean really bad and is below average sensitivity to boot (i.e. if they produce less than say 86 dB/2.83V/1m).
 
What About hi sensitivity , benign load speakers that sound criminally bad ..:)

Take a look at amplfiers at the top of the food chain , talk to your buddy Karan , they all have one thing in common, size , big psu, we can sit here all day posting about them ..


They are not door stoppers, I guess you disagree with them too .... :)
 
Last edited:
What About hi sensitivity , benign load speakers that sound criminally bad ..:)

Exactly the same as mammoth amps, with massive power supplies, which sound disintersted and dead - they are the problem of those who own them, not mine.

Take a look at amplfiers at the top of the food chain , talk to your buddy Karan , they all have one thing in common, size , big psu, we can sit here all day posting about them ..

I don't know what you mean by massive, but they certainly have substantial power supplies. I never denied that, if you megamps of current, all that juice has to come from somewhere. But I still say that just because an amp has all that, it's still no guarantee it will sound good, just loud.

They are not door stoppers, I guess you disagree with them too .... :)

I don't know what you mean by "door stopers".
 
So I spent the day looking over that Philips lookalike preamp. I’ve changed a few parts, and added a few parts, mostly as per Demian’s suggestions.
Input FET buffer now has a trimmer to trim its DC component down to around 1 mV.

High frequency funnies appeared at above 500 kHz. The problem was eventully traced to FET buffer, and resolved by adding a simple 100 pF cap, so that in conjuction with R16 (2,2k) it acts as a first order filter. This eliminated high freqiency funnies.

The bis circuit was changed altogether. The multiple diodes gave way to a classic single transistor adjustable circuit. I have seen that same humble transistor (BC182B) used in that same place in Amcron’s pro power amps. The reason why I did that I am unaware of how diodes and fixed bias ciruits generally behave over time, does aging affect them, and if it does, how and to what extent. Besides, I always like to have my options left open, who knows, somebody may prefer still more bias, etc. I see no particular point in driving it over its present 20.9 mA, but that’s just me.

A speedup capacitor was added to the drivers.

A few compionent values have been changed, but basically, it’s still the same circuit with a little bit higher global NFB.

Mesurements were once again performed. For a 3V output, open loop, R source 600 Ohms, both pots at 99%, I got:

0.058% for 25 mV input into 600 Ohms, 20-20.000 Hz
0.007% for 20 mV input into 10k, 20-20,000 Hz
0.007% for 20 mV input into 20k, 20-20,000 Hz

Still open loop, fro a 10V output, R source 600 Ohms, both pots at 99%, I got:

0.067% for 77 mV input into 600 Ohms, 20-20.000 Hz
0.054% for 72 mV input into 10k Ohms, 20-20.000 Hz
0.052% for 71 mV input into 20k Ohms, 20-20.000 Hz

One wonders why one needs global NFB at all with this kind of performance, but still ...

Now, closed loop, R source 600 Ohms, both pots at 99%, I got:

0.011% for 445 mV input, 10V output, 600 Ohm load, 20-20.000 Hz
0.021% for 408 mV input, 10V output, 10/20 kOhms loads, 50 kHz
0.041% ditto, but at 100kHz

For a 3 V output, all measurements into 600 Ohm loads are below 0.007% and for higher loads below 0.005% at 100 kHz.

Global NFB has been raised to about 5.5:1, or 15 dB. Hardly unseen before, my own Citation 21 preamp works with a global NFB factor of 3:1, or 12 dB.

All in all, not bad. With a bit more work, I suspect it could be a very solid preamp. It might be a good future project.
 

Attachments

  • Preamp No.2.pdf
    41.3 KB · Views: 69
Last edited:
wDX5mBf.jpg


> 10 years ago I doodled this when by a small river in Texas . The output stage although simple is reasonably free of problems . If run hot almost none . This was about the last time I could do any handwriting . Forgive any errors ( i.e R1 C1 position is inverted ) . A simple bias system would be fine if class A to drive 600R at 10 Vrms . The 47 R VAS resistor might not be my choice now .

I dare say built with BD 139/140 it would be great as a pre-amp . The parts were from the parts bin ( 0R25 not 0R22 ) .

The 0.1 uF and 220 uF were if reconnecting to another PSU . The RAW DC about 2V above the regulated at 230 V - 10% .

The sound is ultra sweet so might suit well . How you do the input stage is the fun of it . Input JFET LTP and bipolar second LTP might have virtue . What JFET is easy to get and likes 60 V ( other uses ) ?

Nige, I don't think I'll overdo it if I say that a vast number of amps were made much along these lines in the second half of the 70ies. From Nikko, Toshiba, Panasonic/Technics, Fisher et al.

If you like, I'll send you some service scematics.
 
AX tech editor
Joined 2002
Paid Member
There is a note in W. Marshall Leach's first power amp project published in the Audio magazine, in which he says that to the best of his knowledge, the first fully complementary topology he is aware of was launched by James Bongiorno. Hence the misunderstnding.

Glad to set this right.

I do remember the review by Walt Jung in Audio Amateur of JB's Great American Sound Ampzilla which had a fully balanced/differential front end. I do remember because Walt decribed what he felt was a design oversight by JB (unbalance in the stage) and got into a shouting match with him through letters to the editor.
The impression I got at the time was that it was the first commercial product that had this configuration.

Edit: I looked it up, issue 2/75. Then I stumbled over a 1971 ad for SWTP's Tiger .01 which had the same input config, from 1971.

Jan
 
Last edited:
I do remember the review by Walt Jung in Audio Amateur of JB's Great American Sound Ampzilla which had a fully balanced/differential front end. I do remember because Walt decribed what he felt was a design oversight by JB (unbalance in the stage) and got into a shouting match with him through letters to the editor.
The impression I got at the time was that it was the first commercial product that had this configuration.

Jan

Jan, my remark predates that by several years. I was referring to your plain fully complementary design without the balanced output stage as JB did for the Ampzilla.

BTW, I used to call JB "Old Grumpy", he could be difficult at times, even if he was very kind and generous to me personally, for which I remain grateful.
 
You have this up and running Nige .....

Yes I did . It was rather nice . A friend asked me to give him something not too precious as it might get cloned ( 1000's of times ) . So I gave him the Texas amp . The real Texas amp was bassed on 1964 parts as in Leak ( all NPN a la Bob Stewart ) . 50 were made , it was abandoned ( QC problems ) . My one never arrived . I heard it for 1 hour .

Inca talon integrated

My point is a simple design sometimes is more guaranteed to work .
 
Nige, I don't think I'll overdo it if I say that a vast number of amps were made much along these lines in the second half of the 70ies. From Nikko, Toshiba, Panasonic/Technics, Fisher et al.

If you like, I'll send you some service scematics.

Like an op amp it is 10% of the picture . I will have to photograph the PSU drawings as they are not PDF . When you see the PSU I think everyone will say that's extensive . The VAS PSU is complex discrete shunt regulated . The volume is by relays and resistors like an old fashioned telephone exchange .


1973 cum Denon PMA 250
 
Dvv . When the shop was closed we had 1000's of service manuals . I suspect you and I have duplicates . Mostly mine are Sony , Kenwood , Arcam . Quad .

This amp was not about what I could do as much as how I could do something good that wouldn't benefit the Chinese company too much . Like yourself everyone would say old hat . Hiding things in plain sight it is called . One of my friends says it has done well on the Chinese domestic marker . They make mostly class A above that . The original Chinese version was dreadful . Thing is 90% of it was well engineered . It's defect was class AB with no loop feedback ( yes ) with an inadequate driver stage . All the transistors you see were the ones they used . Would you believe they didn't understand how a typical Brit amp works ! I said like an op amp . Their question was " how do we set 0V " .

The analogy would be the new Norton motorcycle . It looks exactly like the old ones . That is until small details are looked at . For example the pistons 90 degrees apart in a parallel twin ( should be 360 or 180 ) . A small balance shaft to cope with the distortion if you like . That is a radical difference .
 
Is this supposed to be the Texas amp ..........


I was in the Huston suburbs . I was the only person outside at 100 F . Mad dogs and Englishmen go out in the mid day sun etcetera .

BTW . I adore Texans . So easy to know . I think there is something of Gone With The Wind about me and it is respected there . The Texans you meet outside of Texas are another thing . Texas itself isn't that wonderful and was very pouted . The guy I was staying with had Bose 901 and was alcoholically addicted to them . He never gave up saying about them like I wasn't in the industry . Yes they sounded mute as they usually do . Remember I like them .

I would love to tell you about NASA . It would be pages of stuff we mostly know . Huston Zoo . Is that steam loco still there ? One enters NASA past the loco . KFC is better in England , how weird .
 
Last edited:
Status
Not open for further replies.