The essence of good design

PSP · « **on:** 13 Feb 2004, 03:56 pm »

I just wanted to put this quote out where everyone would see it. This quote from Hugh is true wisdom:

Quote

As a general principle, and of course I should say YMMV, an upgrade needs to address component choices first, operating point second, and maybe wire last of all. I tend to tune everything by ear; most of my time goes into component choice and operating point once the topology/schematic has been figured out. I feel that the design is only mature when as many components as possible have been removed from the circuit. I was disappointed with the GK1 because there are so many caps in the front end; but Darl and I tried very hard to remove them, and every excision resulted in a one-legged presentation, so we HAD to put 'em back in! I am absolutely convinced that the best sonics come with elegant minimalism; generally I have noticed that this applies across all technologies. Complexity is easy, and often indicates incomplete or even flawed R&D. Simplicity, production repeatability, careful component selection results in higher reliability and lower cost and is the essence of good design.

TG · « **Reply #1 on:** 14 Feb 2004, 06:30 am »

Quote

... I tend to tune everything by ear ...

That's the magic mantra to memorise boys and girls, and the reason Hugh's stuff stands out from the pack.

You can't build a circuit without math, you can't get in tune without ears, and you can't make music without soul.

AKSA · « **Reply #2 on:** 14 Feb 2004, 11:26 am »

Thank you Peter, Tony,

High praise, much appreciated.

Hugh

Al Garay · « **Reply #3 on:** 14 Feb 2004, 08:47 pm »

So much love, it must be Valentine's Day!

Spend a bit of time with your friend/partner.

Al

keithvv · « **Reply #4 on:** 25 Feb 2004, 03:48 pm »

I (and my "audionut" friends) have been extremely impressed with the 55N that I built from Hugh's kit - in fact I now own six of these amps, and regularly have to un-install one so that friends can "try it" in their system.

One point that always comes up in our discussion is "why" does this amp sound so sweet, controlled, bass rich, and yet very fast; particularly when compared to many other transistor amps with high reputations? (Lest you think me just making this up, the amps that have been removed from my system and sold to make room for the 55Ns are: Classe CA-100 and McCormack DNA 0.5 revision A's.)

I have just finished reading the article in the latest (March 2004, pg. 32)AudioXpress called "Amplifier Musicality" - reprinted from a Jean Hirago article of 25 years ago - in which harmonic distortion, down to the 20th harmonic, is compared on several tube and transistor amps.

I have also read on the AKSA website - FAQ section - where Hugh states that his design for the AKSA has not focused on great THD specs, but rather on musicality of the amplifier. Perhaps his circuit topology and design implementation has hit on the component of harmonic distortion as opposed to THD?

If any of you "more technologically gifted" out there have the gear (and the desire), perhaps you might be able to measure the component of distortion in the 55N and report back? I would not be surprised to see the distortion plots have controlled decay in the second order harmonics.

Who knows, maybe Hugh could get some more advertising fodder for the masses out of this?

Thanks for reading.
Keith

Maerten · « **Reply #5 on:** 25 Feb 2004, 05:30 pm »

Quote

Harmonic Distortion: Total Harmonic Distortion at 1 kHz into 8 ohm resistive load using HP339: 0.2% at 0.1W; 0.08% at 1W; 0.007% at 10W and same to clipping
At 10kHz and 1W into 8R resistive: 0.042%; at 20kHz and 1W: 0.045%.

Not exceeding 67dB below (0.045%) at any frequency to 20KHz for 10W to 95W into 8R. Distortion profile is second and third harmonic, both 'musical', with all other distortion components, including intermodulation, immeasureable and below the noise floor.

This is for the AKSA 100W

(These results were obtained on 29 December 2002 by Robert L. Jones of California using an HP Model 339 distortion analyser on a stock 100W Nirvana routinely constructed according to the instructions. They are quoted here with his kind permission.)

keithvv · « **Reply #6 on:** 25 Feb 2004, 06:39 pm »

Maerten,
Thanks for the data on THD... I was hoping to find the lcomponent evels of the 2nd, 3rd, 4th, ... 20th harmonic relative to the primary signal level.
If you know Mr. Jones well enough, could you ask him if he by chance has that data?
Much appreciated,
Keith

Maerten · « **Reply #7 on:** 25 Feb 2004, 10:41 pm »

I'm sorry the bit of Mr Jones should've been in the quote, its from the site. I've decided to put that in later because i think if someone took the effort, he also deserves the credit

Since its from the site i assume mr Dean knows mr Jones... but if he could've mesured the specific value's you stated with his device, i think he already would've.

Just a wild guess!

I think you have to contact mr Dean for more info, since he's on the forum alot i think it would be ok to send him a pm

Maarten

AKSA · « **Reply #8 on:** 26 Feb 2004, 01:41 am »

At the time, Robert assured me that the distortion was all H2 and H3, and in fact that had been my precise aim during the design phase.

However, he did not mention the proportions. I will email him, but I doubt that he has the amp now, since he went on to build his own design to see if he could better the results!

I'll do my best, Keith. Thanks for your post, and Maerten, for yours too!

Cheers,

Hugh

TG · « **Reply #9 on:** 27 Feb 2004, 03:06 am »

Quote from: AKSA

At the time, Robert assured me that the distortion was all H2 and H3, and in fact that had been my precise aim during the design phase.

Ah, ha! - that's why AKSA amps sound so "familiar" to us tubeheads. Banishing the higher order odd harmonics from solid state is no mean feat.

ginger · « **Reply #10 on:** 27 Feb 2004, 05:13 am »

Its good to see others reflect my own design philosophy.

You see the KISS "Keep it simple stupid" principle quoted by many.

As Professor Julius Sumner-Miller used to say "Why is it so?"

Simple designs have simple problems.
Simple problems have simple solutions.

A question I often ask some of the software engineers (although it can just as easily be applied to hardware design):

Given that its 3 times as difficult to debug as it is to design, then if you have been as clever as you possibly can be in the design, when do you reckon you will complete the debugging?

ANSWER: Never.

Cheers,
Ginger

Senior Electronic Design Engineer - Technology Park, ADELAIDE

P.S. How does this relate to the AKSA.
I have just completed 4 months of work on my 55N.
I tried deleting the bootstrap and using a current source load on the Voltage Amplifier Stage. I tried various amounts of emitter degeneration on the diff amp. I tried modifying component values and operating points, I tried most of the various "Audio Jewelery" available.

Where did I end up:
Back where I started - Hugh's original design!!!!
Except:
Electrolytic Caps all Blackgates
4R7 emitter degeneration on the diff amp- else bog (Nivarna) standard.

Feel free to experiment for yourself BUT I have been exhaustive in my efforts to "improve" the design and ended up proving the design is right "as is".

econ · « **Reply #11 on:** 28 Feb 2004, 04:04 am »

Quote

4R7 emitter degeneration on the diff amp- else bog (Nivarna) standard

Interesting journey.

Could you expand on the 4R7 bit? Have you substituted a resistor?added one somewhere ? Thanks

ginger · « **Reply #12 on:** 1 Mar 2004, 03:35 am »

econ,
adding emitter degeneration is simply adding a 4R7 (in this case) resistor in series with each emitter of the diff amp. These emitters are currently connected together and so this mod requires PCB tracks to be cut and is NOT RECOMMENDED for the faint hearted. It tends to force better DC and AC balance in the diff amp and modifies the degree of change of the of the diff amp transconductance (current out vs differential voltage in) with input voltage swing.

Adding a little degeneration seemed to clean up the top end a little while adding a lot (which is common in modern high end amps) seemed to ruin the image and make it receed.

For a discussion of the effects of emitter degeneration see the Doug Self site and read his paper on the sources of distortion in power amps. It gives some indication of some of the things Hugh most likely tried in arriving at the AKSA cicrcuit. I found it extremely interesting reading - as much for looking at the distortion cancelling techniques Hugh did'nt find sonically attractive and left out as for the ones he thought helped and therefore did include.
The site is:
http://www.dself.dsl.pipex.com/ampins/dipa/dipa.htm

CAUTION: This paper describes how to design for absolute minimum distortion - It says NOTHING about how it will sound. I have also tried many of the circuit variations described in the paper and did'nt find any that are not already in the design which I though improved the sound.

Cheers,
Ginger