Dave Osborne wrote:
The Natural Philosopher wrote:

The difference between the circuits in magazines, and the circuits you
put in production are simple.

A magazine circuit worked well for one man, once.

A production circuit has to work well for all spreads of
semiconductors under all customer conditions.

The very first circuit I ever built from a magazine failed when I left
in in the sun. It was fine when I put it back in the shade. I was IIRC
13. That's when I realised that I had better start understanding what
I was doing. Since the guys in the magazines patently didn't.

I probably built half of all the HiFi amps of the period. Pretty sure
I did a Linsley hood was it a Bailey. Did a class A as well. Worked
for Uncle Clive, then on into HiFi proper for a couple of years, then
the music business for disco and PA and guitar usage..waste of my time
really. Only one of the firms I worked for still exists. (Actually
only one of the firms I worked for in my entire career doing
electronics still exists in this country, and that's Marconi Elliot
Avionics, which lingers on as SELEX Sensors and Airborne Systems. The
other one still exists in S Africa)

It was post Uncle Clive when I realised finally that technicians could
put circuits together and fiddle with them till they worked without
really understanding why they did, or the implications of their design
choices, that I finally answered the question of how alleged experts
who wrote for magazines could produce stuff that didn't work properly.
Because they were of the 'fiddle till it works' brigade.

Now we all did that,. but some of us with more insight started from
better places that required less fiddling to get to better results.

The better results in power amps came really from better output
devices. Essentially the better those were the more feed back you
could slap round the whole thing, without going unstable, and the
lower distortion you could get, and the higher frequency the nasties
in the crossover region could be pushed.

And 99% of the circuits were in fact mild adaptations of the
application notes supplied by the semiconductor magazines. That's
where the better fiddlers started their fiddling from.

Ultimately I realised that all a PA is, is an arrangement of power
devices that is broadly symmetrical to drive a complex load. and
simply work back from there. Those devices themselves have to be
driven, safely and securely so that all devices are inside the SOAR,
and then they need phase splitting if its class B, and then some kind
of reasonably low noise input stage.

Then try and keep the whole arrangement stable into just about any
possible load, with enough feed back to get distortion down..early
stuff we could do 0.1%, pretty soon I was hitting 0.025%, and by the
time I finished it was pretty much off the scale in the noise. Even
low level crossover was getting to be almost unmeasurable, compared
with the 1-2% it had been on e.g. a Quad 303. Frequency responses
likewise went up as better devices came along.

Hitachi complementary FETS were the best devices I worked with..back
about 1980 or so.

Really, there wasn't anything left to do then. You could produce an
amplifier that was better than any valve amplifier, ruler flat and
totally distortion free, in a back shed. You couldn't sell it though,
without employing the sort of spin merchant that made Alistair
Campbell look like Honest John, Or putting valves in it.

I realised from my time building high distortion guitar amplification
that people actually like distortion. It's something they get used to
and it sounds 'right' to them. Likewise natural resonances and
colouration in Loudspeakers.

Put them in front of a flat amp and really good loudspeakers and they
feel something is missing: It is. All the distortion and colouration
they are used to. The sound becomes naked and sterile, because all
they now hear is what is actually on the CD..

Well, the CD was in by then, and the cassette tape had taken the big
money out of the music business, prejudice and religion ruled the HiFi
world, and I was still broke, and no one wanted a circuit designer
outside of Silicon Valley or Japan so I started playing with
microcomputers and turned my back on it all.

I was in fact pretty disillusioned. I'd given them 'the closest
approach to the original sound' made em rock solid, and idiot proof,
and realised that that wasn't actually what they wanted. They wanted
style, bragging rights, fashion statements, a brand..not what I could
provide really. I majored in proper engineering.

At least in the music business they were more honest. 'Just so it
sounds like a Marshall turned up to 11, I don't care what that means,
that's what I want'.

I got a Technics amp once. Looked at the circuit diagram and thought
'hey, That's the way *I* broke the patent on that clever AAB design I
stole..and they integrated the whole lot on a chip to get temperature
stability' It was a ripoff of my ripoff!!..

And I wouldn't mind betting that the output of any Peavey guitar Amp
to this day is exactly the way I designed it (but not for Peavey!). So
some of the work lives on I suppose.


That's a very interesting post and echoes a lot of my experience over
the years. Thanks for that.

I recently came across this (it's been around for ten years, but passed
me by):

http://en.wikipedia.org/wiki/Gainclone

Says it all really. HiFi as a concept is pretty dead for me. Almost
everything audio sounds 'good enough' these days (if used in an
appropriate context, natch).

well those chips are not that much good for top end applications really.

They are good enough for guitars though, and I used a few in my
time..hung some weird feedback round em to do high output impedance and
soft clipping a la vox AC 30..

But they are good enough to be called 'hifi' for sure.

And ten times better than most loudspeakers..:-)