In article , Serge
Auckland
wrote:


In a way I do, almost every time I use my Hi-Fi. My Meridian-based
system has the volume control as part of the loudspeakers, and the
steps are 1dB at the levels I listen at.
[snip]

Adjusting the volume by 1 step isn't readily perceptible. I think I can
hear a difference, but as I've just made the change, that's not
surprising. I certainly couldn't step out of the room and return and
tell a 1dB difference on programme. Two steps seems to be the minimum I
can tell readily, although even here, I don't think I could step out of
the room and return and say the volume has changed. 3dB seems the
minimum to tell readily there's been a change.

Similarly, with tone changes, the Meridians have treble and bass
shelving in 1dB steps, and a 1dB change is imperceptible, 2dB in the
treble and 3dB in the bass becomes noticeable, boost more readily
noticeable than cut.

It is possible, however, that I am rather less sensitive to level change
than others, so I would not be dogmatic that because I can't hear it,
others can't either.

Well, the systems I use employ use a mix of Quad 34 and Armstrong 700
preamps. These both use the Alps stepped attenuators. My experience is much
like yours.

On the other hand, I'm extremely sensitive to stereo positioning, a 1dB
difference in loudness between L and R moves the image enough to be
irritating.

I find that somewhat less than a change of 1dB in balance can be audible. I
suspect this is a strong function of factors like how well the stereo works
in the first place in terms of audible symmetry, direct/indirect sound
ratio, etc.

Slainte,

Jim

--
Change 'noise' to 'jcgl' if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

Jim Lesurf wrote:

Eeyore wrote:
Jim Lesurf wrote:

More to the point, the original IV limiting using on the 405 was
particularly severe for reactive loads - rather common for
loudspeakers!

I could write an article on the subject.

You mean something like the SCAMP (Society for Cruelty to AMPlifiers)
article in the Hi Fi News section of audiomisc.co.uk ? :-)

Very much so.

Graham

Jim Lesurf wrote:

Eeyore wrote:
Jim Lesurf wrote:

More to the point, the original IV limiting using on the 405 was
particularly severe for reactive loads - rather common for
loudspeakers!

That's a very interesting subject in its own right.

Back then, designers tended to be 'over-protective' of their output
devices. No doubt influenced by the fragility of early parts.

A few decades ago the power devices were rather prone to secondary
breakdown, and had quite modest IV handling.

Things have moved on a bit. I tend no to look mainly at average
dissipation as opposed to instanteous V/I limiting now.

Can't say I've seen many failures.

Opps, giving my secrets away here.

Not sure what 'secrets' remain. Even by about 1980 I had no problem
designing a 200+ Wpc amp that used no IV SOA limiting for protection. Just
needed power line fuses. Just a matter of designing for the task. I'd
always thought that IV limiters were a dubious idea. The original 405 just
confirmed that for me. Given how much device technology has moved on, I'd
expect it to be a trivial to make a safe amp with no explicit SOA
protection these days if the designer knows what to do.

Might need rather more output devices than the bean counters would be happy
with though.

Graham

Jim Lesurf wrote:

Eeyore wrote:
Geoff Mackenzie wrote:

I do like some of their definitions - a good amp being "straight wire
plus gain".

But can it drive pure inductive or capacitive loads ?

Unless you qualify that by giving values, frequencies, etc, then no
amplifier could be said to return an unconditional 'yes'. But if you set
plausible values for domestic audio, then various amplifiers would return
'yes'. So your question is rather too vague and sweeping to be useful.

The inevitable problem.

Graham

In article , Eeyore
wrote:


Jim Lesurf wrote:

Eeyore wrote:
Jim Lesurf wrote:

More to the point, the original IV limiting using on the 405 was
particularly severe for reactive loads - rather common for
loudspeakers!

I could write an article on the subject.

You mean something like the SCAMP (Society for Cruelty to AMPlifiers)
article in the Hi Fi News section of audiomisc.co.uk ? :-)

Very much so.

Keith Howard followed up 'SCAMP' with an article that established what the
worst choice of load phase angle would be for o/p device dissipation in
class A/AB operation. He tends to do the measurements for the HFN speaker
reviews. This was an issue we'd be discussing for a while.

Since then, he has tended to check each speaker to find at what frequency
it behaves 'worst' as a load, and then mention the results in the review.
This in terms of the resistive load that would place the same peak power
dissipation demands on the o/p devices. I don't think he always does this
as some speakers are 'gentle' loads, so their worst behaviour isn't a
concern. FWIW He also wrote about this for Stereophile IIRC.

Harder to assess the effect in secondary breakdown terms as that will vary
from design to design, so no single conclusion could be drawn about a given
speaker. Fortunately, the o/p devices these days should be less prone to
breakdown nowdays than back in ye days of yore...

It isn't a new problem. But speaker manufacturers simply tend to lob the
ball into the court of the amplifier designer and smile sweetly. I can't
complain too much as I've always favoured QUAD ESLs. :-)

Slainte,

Jim

--
Change 'noise' to 'jcgl' if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

In article , Eeyore
wrote:


Jim Lesurf wrote:


Not sure what 'secrets' remain. Even by about 1980 I had no problem
designing a 200+ Wpc amp that used no IV SOA limiting for protection.
Just needed power line fuses. Just a matter of designing for the task.
I'd always thought that IV limiters were a dubious idea. The original
405 just confirmed that for me. Given how much device technology has
moved on, I'd expect it to be a trivial to make a safe amp with no
explicit SOA protection these days if the designer knows what to do.

Might need rather more output devices than the bean counters would be
happy with though.

The trick is to show them that it ends up being cheaper than having to
repair the damn things when they keep winging back with blown devices. Plus
pointing out that you save cash by not having to faff about including IV
limiting arrangements that also require components, board space, assembly,
etc. Been there. Have the tee-shirt. :-)

And, of course, to satisfy them that the higher reliability and performance
translate into more sales.

Slainte,

Jim

--
Change 'noise' to 'jcgl' if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

Jim Lesurf wrote:

Eeyore wrote:
Jim Lesurf wrote:
Eeyore wrote:
Jim Lesurf wrote:

More to the point, the original IV limiting using on the 405 was
particularly severe for reactive loads - rather common for
loudspeakers!

I could write an article on the subject.

You mean something like the SCAMP (Society for Cruelty to AMPlifiers)
article in the Hi Fi News section of audiomisc.co.uk ? :-)

Very much so.

Keith Howard followed up 'SCAMP' with an article that established what the
worst choice of load phase angle would be for o/p device dissipation in
class A/AB operation. He tends to do the measurements for the HFN speaker
reviews. This was an issue we'd be discussing for a while.

Since then, he has tended to check each speaker to find at what frequency
it behaves 'worst' as a load, and then mention the results in the review.
This in terms of the resistive load that would place the same peak power
dissipation demands on the o/p devices. I don't think he always does this
as some speakers are 'gentle' loads, so their worst behaviour isn't a
concern. FWIW He also wrote about this for Stereophile IIRC.

Harder to assess the effect in secondary breakdown terms as that will vary
from design to design, so no single conclusion could be drawn about a given
speaker. Fortunately, the o/p devices these days should be less prone to
breakdown nowdays than back in ye days of yore...

It isn't a new problem. But speaker manufacturers simply tend to lob the
ball into the court of the amplifier designer and smile sweetly. I can't
complain too much as I've always favoured QUAD ESLs. :-)

How far do they swing ? Which model even ?

Years back I did the crossover design for a modest 'bookshelf' monitor style
speaker using Mathcad.. It's all 'impedance corrected'. I haven't actually run
a test myself as it happens (although my colleague did some 'empirical ones'
and found the impedance curve very flat using spot test tones but aside from
the usual LF peak, it should be as close to resistive as possible.

As a result it sounds sweet even on cheap amps.

Graham

Jim Lesurf wrote:

Eeyore wrote:
Jim Lesurf wrote:

Not sure what 'secrets' remain. Even by about 1980 I had no problem
designing a 200+ Wpc amp that used no IV SOA limiting for protection.
Just needed power line fuses. Just a matter of designing for the task.
I'd always thought that IV limiters were a dubious idea. The original
405 just confirmed that for me. Given how much device technology has
moved on, I'd expect it to be a trivial to make a safe amp with no
explicit SOA protection these days if the designer knows what to do.

Might need rather more output devices than the bean counters would be
happy with though.

The trick is to show them that it ends up being cheaper than having to
repair the damn things when they keep winging back with blown devices.

Ah yes, and let's hope they can understand the concept !

ISTR a 'rule of thumb' that fixing faults gets TEN times more expensive every
step the product goes down the line, never mind reaching the customer !

Warranty repair costs can simply KILL a company.


Plus
pointing out that you save cash by not having to faff about including IV
limiting arrangements that also require components, board space, assembly,
etc. Been there. Have the tee-shirt. :-)

Mosfets are so simple like that !


And, of course, to satisfy them that the higher reliability and performance
translate into more sales.

One hopes.

Graham

In article , Eeyore
wrote:


Jim Lesurf wrote:
[snip SCAMP comments]

It isn't a new problem. But speaker manufacturers simply tend to lob
the ball into the court of the amplifier designer and smile sweetly. I
can't complain too much as I've always favoured QUAD ESLs. :-)

How far do they swing ? Which model even ?

In my case I used to use an early pair of the original QUADs. These weren't
even a stereo pair. One had been bought in mono days, and the other to go
with it when stereo arrived. They were then used as test loads, etc, at
Armstrong before I took them home for domestic used. I did once set fire to
one of them whilst testing an amplifier. Acoustical phoned me up when it
was sent for repair and asked me in a suitably astonished voice just what
the deleted I'd been up to. :-) Sounded fine again after repair.
Curiously, it also sounded OK for a while when flames were coming out of
it. 8-] I had been deliberately using its transformer saturation as a
nasty load to see the power amp I was working on would survive. It did.

Later on I changed to ESL63s for the main hifi, one of the early issue
boards, but I can't recall the iss number off hand. IIRC one of the
versions where the LF impedance is signal level dependent. Plus I now also
have a pair of 988s for the living room AV system.

Years back I did the crossover design for a modest 'bookshelf' monitor
style speaker using Mathcad.. It's all 'impedance corrected'. I haven't
actually run a test myself as it happens (although my colleague did some
'empirical ones' and found the impedance curve very flat using spot test
tones but aside from the usual LF peak, it should be as close to
resistive as possible.

As a result it sounds sweet even on cheap amps.

It can be done. Alas, most speaker manufacturers don't bother. Simpler from
their POV not to. Just dump the problem onto the amp, and let someone
else's product require more work and cost.

Slainte,

Jim

--
Change 'noise' to 'jcgl' if you wish to email me.
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Armstrong Audio http://www.audiomisc.co.uk/Armstrong/armstrong.html
Audio Misc http://www.audiomisc.co.uk/index.html

Eeyore wrote:

My friend's response? "Oh, you're using those fancy Chord interconnects
on the Arcam, you're using a £10 Cambridge Atlantic on the Technics."
He's not a City Banker stockbroker / commodities broker is he ? They're
pretty gullible.

No... he's a builder!

--
Squirrel Solutions Ltd Tel: (01453) 845735
http://www.squirrelsolutions.co.uk/ Fax: (01453) 843773

Registered in England: 05877408