Tube amplifiers
 

In article , Patrick Turner
wrote:


Stewart Pinkerton wrote:


Your Beeb pal might like to consider that many modern speakers have
impedances which dip to 3 ohms or so, somewhere around the crossover
frequency, which is often in the region where the ear is most
sensitive. A difference in Ro of 0.5 ohms will give a change of volume
of about 1.5dB in the area of that dip, which most speaker engineers
would consider very likely to be audible.

A tube amp designed for 8 ohms with Ro = 0.5 ohms has a good DF.

Your use of the judgement "good" here is not one I would share as a
sweeping statement. In some cases it might be fine, but in others it might
not.

So use 8 ohms, eh.

Alas, many real loudspeakers are not so obliging, and have impedance that
vary from well above 16 Ohms at some frequencies to well below 8 at others.

If the designer has any brains, he will have provided alternatively
arranged secondaries on the OPT so that the turn ratio can be reduced by
a factor of 0.67, ( two windings of N turns instead of three windings of
N turns in series )

Alas, the same designer will also be aware that if they simply leave the
user to select a lower tap in order to get the effective impedance down
from about 0.5 Ohms to less than 0.1 then the result may be distinct drop
of the available power into speakers which have a nominal impedance of 8
Ohms, but have a lower impedance at some frequencies.

The problem here is that many real loudspeakers are not so obliging as to
have an impedance that is resistive and frequency-independent across the
audio band. Must admit that when I designed amps I used to moan at speaker
designers about this, but alas, this had no effect... :-)

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

Tube amplifiers
 

In article , Patrick Turner
wrote:


Jim Lesurf wrote:


The effect will obviously vary from one case to another, and often may
be small enough not to really matter. But I would suspect that rather
more than 1% of people would find the effect of series resistances of
this order would produce an audible change. However it easy enough for
people to try this if they wish, and then judge for themselves.

Unless ppl are told a 1 ohm R is connected, they rarely realize anyhing
is different.

You may be able to make similar comments about many effects in 'hi fi'.
:-)

However the statement you make above is quite different to the one you made
some postings ago. Hence I take it that you now accept that an output
impedance of the order of an Ohm can indeed often make an audible
difference. Hence it may have some effect on the sound, and hence people
may form a preference which will depend on the circumstances.

My point here is that with a power amp with inherently low o/p impedance a
user can fairly easily choose to add a series resistor if they so desire
and then leave it in place if they like the results.

Conversely, if they have a power amp with an inherently high o/p impedance
they they either have to go to a different tapping (if valve) and sacrifice
power (and possible change other factors), or change the amp entirely.

Thus my view is that a low o/p impedance is a better option as it accords
better with the nominal assumption that the speaker expects a voltage
source, and makes for an easier change if the user desires.

It also helps avoid effects of the kinds I've examined in detail in the
'303 and 57' page where the o/p impedance as well as the load have
significant reactance. (Since the 303 is a transistor design and I suspect
PJW did this deliberately to mimic the Quad II/57 this also shows that this
is not inherently a valve versus transistor point. Some SS amps may have
high output impedances. Some valve amps may have a low output impedance.
Matter of design.)

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

Tube amplifiers
 

In article , Stewart Pinkerton
wrote:
On Wed, 22 Dec 2004 03:08:55 +1100, Patrick Turner
wrote:


Then if you connect the 8 ohm speakers to your 3.6 ohm outlet, the
classA % of the total AB power increases, and thd is halved, DF is
obviously so much better, all at the cost of slightly less maximum
power.

Slightly less? you mean 3dB, *half* the power.

My concern here is that if you wish to go from Ro = 0.5 down to Ro 0.1
simply by changing what tap the speaker is connected to, then the reduction
in available power may be more than just 3dB. However this will depend on
the details.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

Tube amplifiers
 

In article , Iain M Churches
wrote:

"Patrick Turner" wrote in message
...


A tube amp designed for 8 ohms with Ro = 0.5 ohms has a good DF.

Stewart has misquoted my figure which was Ro=0.4 Ohms, so DF=20 I have
been given to understand by others that this is a good figure, and that
increases above DF 12 do not seem to yield audible improvement.

The difficulty here is with judgements like "audible improvement" and with
assuming this can be taken to always be so simply on the basis that you
have been "given to understand" this.

The problem is that the change may be modest in many cases. But may be
audible, and I'd suspect that if people simply were given a speaker, amp,
and room acoustic, and asked to judge the effect of switching in/out a
series resistance of this order, that there would be no reason to think
most people would clearly prefer an effective source resistance of 1 Ohm to
one 0.1 Ohms.

However if people have an amp with a low output impedance they can try this
and judge for themselves. The snag is that if they have an amp with a high
output impedance they may find this more difficult to do...

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

Tube amplifiers
 

In article , Iain M Churches
wrote:

Please refer to the AES recommendations for speaker impedances (which
are followed by reputable manufacturers) They do not allow a speaker of
nominal impedance of 8 Ohms to fall to 3 Ohms.

As an aside he Can you give me a reference for the AES standard? I don't
have this to hand and I'd be interested to see what they specifically say.

I am also curious as to what value they would accept for 'unusual' speakers
like the old Quad 57's whose impedance varies all over the shop. :-)

[snip - leave SP to reply as he chooses as the questions were directed at
him. :-) ]

So your "change of volume of about 1.5dB" is no longer valid, being
reduced to a value which could possibly be perceived on sweep tones
(this will be an interesting thing to test in the New year) but
probably not with music.

I am less sure than yourself that an o/p impedance of 0.4 Ohms would only
be noticable on sweep tones. However the only tests I've done on this sort
of thing in the past involved music, so I can't comment on the relative
audibility music versus tones.

Perhaps if Arny is interested we could make up some 'ABX' type tests that
apply the kind of modification a 0.4 Ohm (or 1 Ohm) o/p would impose on
some typical speaker(s) and then invite people to judge for themselves? It
seems to me this would be an interesting experiment.

Slainte,

Jim


Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

Tube amplifiers
 

"Jim Lesurf" wrote in message
...
In article , Iain M Churches
wrote:

Stewart has misquoted my figure which was Ro=0.4 Ohms, so DF=20 I have
been given to understand by others that this is a good figure, and that
increases above DF 12 do not seem to yield audible improvement.

The difficulty here is with judgements like "audible improvement" and with
assuming this can be taken to always be so simply on the basis that you
have been "given to understand" this.

Both Olson and Tremaine have documented test results with DFs up to 100.
I think earlier in the thread I gave the page in Tremaine's book.
These test were carried out using panels of listeners drawn from varying
age groups. Is there any reason to think that they cannot be relied upon?


When I write "given to understand", this is my way of stating something
which I have been taught, and believe to be true, but have no way of
proving for my self. Actually, in the New Year, I have the opportunity
to do some testing in a broadcast lab. This will make an interesting
experiment.


The problem is that the change may be modest in many cases. But may be
audible, and I'd suspect that if people simply were given a speaker, amp,
and room acoustic, and asked to judge the effect of switching in/out a
series resistance of this order, that there would be no reason to think
most people would clearly prefer an effective source resistance of 1 Ohm
to
one 0.1 Ohms.


Values quoted here seem to vary with the direction of the wind:-)
From where to do we get this 1 Ohm now?
I talked originally of Ro=0.4 Ohms, and DF=20, so a series resistor
of 0.33 Ohms would be more appropriate.

Cordially,

Iain

Tube amplifiers
 

In article , Patrick Turner
wrote:


Jim Lesurf wrote:

In article , Patrick Turner
wrote:




Lemme see now.

I do about 40 amps per year, nearly all SS. I am a very low volume
repairist,

OK. When you say "nearly all", how many of the amps you repair are *not* SS
during a typical year?

and if 50 of us have an average of 60 per year that's 3,000 SS amps per
year need an average fix of $80 each.

The snag here is that you have altered your own experience (40/year) to
being 60/year for others. Do you have actual figures for that? And as
above, how many of the repairs others do don't fall into your "nearly all"
category and are not SS, but valve?

There are 120,000 households for 300,000 ppl, and with an average of 1
amp per household. So 2.5% fail each year, with many failing before the
warranty period expires. Ppl replace all their electronics about every 7
years on average.

Well, I don't know about your factor "nearly all". Nor do I know about how
reliable the basis is for your assumption that the average for everyone
else is about 1.5x on your own number of repairs/year.

Nor do we know from your figures how many valve amps you and others
actually repair per year. The problem here is that you are 'extrapolating'
from your own results at the same time as saying your results are not
representitive. Given some of the sweeping comments you have made in other
postings I am not sure how well you judge this.

Nor do we know what fraction of the amps in service in your area are not
SS.

So, if were to assume that you are nearer the average that you say, then
the above value would drop a bit. And it would drop a bit more if we take
the "nearly all" into account. Lacking reliable figures it is hard to
decide just how reliable your estimate may be.

FWIW If you'd asked me to guess, then based purely on recollections from
many years ago, I'd have said that fair-quality audio amps (as distinct
from the cheaper items many people buy in general electronics shops) would
have rates of the order of 1% per year averaged over the order of the first
five years. But this would be a guess. Despite that, it isn't a long
distance away from the kind of values your comments imply.

The problems a

1) We still have no reasonable 'like for like' (in terms of cost and market
level) comparison of reliability of SS versus valve from what you say.
Hence if valve amps had the same level of reliability and service
requirements, but were only 1% of the market, you would have around a 50:50
chance of getting one valve amp to repair in a typical year. Hard to do
statistics on such a small sample, but how many of the amps you repair are
not in the "nearly all" category you defined above? i.e. How many are
valve? Without reliable info on this we can't say if you should be
regarding valve amps as being just as 'unreliable' - or worse - than SS.

2) Nor do we know if what you say really sweeping claims you made some
time ago that prompted this discussion.

If we make the blanket assumption that amps are just as likely to require
service in the first year as the second, or in any subsequent year, then a
time-averaged result of 1 % implies that around 95% would survive 5 years
with no service requirement. If we assume 2 % then this drops to 90%
surviving the first five years with no service requirement. These effects
unequally distributed between models depending on their cost and target
audience.

Assuming we are comsidering good quality units, I am not sure that knowing
that the order of 95% or more do not require any service for 5 years or
more would cause most people to worry that the units were 'unreliable'.
However I also don't know how representitive your estimates or mine are of
this class of amp. Maybe they do better than your figures imply. The
problem her is that reliable units will be under-represented in what you
see for the obvious reason that they won't require service so often. :-)

However my own recollection is also that the likelyhood of a service
requirement tends to follow the standard 'bathtub curve'. This does not
distribute the rate in the way assumed by the above. The main effect is a
'child mortality' one and then after some time a wearing out. My experience
of this, though, is also that service after the child mortality with SS
tends to be due to things like switches, etc. Not as a result of SS devices
per se.

For the above reasons I still do not think that the experience you quote
justifies singling out SS amps and claiming they are unreliable if you are
comparing them with valve amps. You have not provided data that illuminates
that question, or some of the others I mention above.

Your idea of a one in 300,000 failure rate p.a. is just **** and wind.

I've left discussing *your* "idea" of that to a separate posting. Here I've
just concentrated on the estimates you make. :-)

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

Tube amplifiers
 

In article , Iain M Churches
wrote:

"Jim Lesurf" wrote in message
...
In article , Iain M Churches
wrote:

Stewart has misquoted my figure which was Ro=0.4 Ohms, so DF=20 I
have been given to understand by others that this is a good figure,
and that increases above DF 12 do not seem to yield audible
improvement.

The difficulty here is with judgements like "audible improvement" and
with assuming this can be taken to always be so simply on the basis
that you have been "given to understand" this.

Both Olson and Tremaine have documented test results with DFs up to 100.
I think earlier in the thread I gave the page in Tremaine's book.

I'm afraid I do not have the book. Is their research reported in JAES? If
so, I can look it up in the JAES CDROM set I have.

These test were carried out using panels of listeners drawn from varying
age groups. Is there any reason to think that they cannot be relied
upon?

I can't say without reading what they did and the details of their process,
conclusions, etc. However the distinction I am making here is between
"difference" and "improvement". Lowering the o/p impedance may have an
audible effect until any further reduction ceases to be noticable. But this
is not the same as assuming that a lower (or higher!) value is always an
"improvement" as that is a value judgement that will depend on the
circumstances.

So if Tremaine, etc are defining what you call "improvement" as 'becoming
like what you hear with zero o/p impedance then that would be specific. But
if they have some other meaning, I would be wary of "improvement" as the
term to use here. Can't say more without reading the details, I'm afraid.


When I write "given to understand", this is my way of stating something
which I have been taught, and believe to be true, but have no way of
proving for my self. Actually, in the New Year, I have the opportunity
to do some testing in a broadcast lab. This will make an interesting
experiment.

OK. For obvious reasons, the choice of speakers and the room acoustics will
also matter. FWIW I'd prefer to do tests like this with various sorts of
music, but test tones would also be of interest. If you can, try an 'ABX'
approach on people.


The problem is that the change may be modest in many cases. But may be
audible, and I'd suspect that if people simply were given a speaker,
amp, and room acoustic, and asked to judge the effect of switching
in/out a series resistance of this order, that there would be no
reason to think most people would clearly prefer an effective source
resistance of 1 Ohm to one 0.1 Ohms.


Values quoted here seem to vary with the direction of the wind:-) From
where to do we get this 1 Ohm now? I talked originally of Ro=0.4 Ohms,
and DF=20, so a series resistor of 0.33 Ohms would be more appropriate.

I am not sure, but I think it was Patrick who made a comment to the effect
that most people would not notice an output impedance of 1 Ohm. I'm not too
fussed about the specific values chosen for a basis of discussion, but I'd
call 1 Ohm 'high' and 0.1 'low' in general terms when thinking of '8 Ohm'
speakers and their typical impedance variations, accepting that this does
depend upon the circumstances of use.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

Tube amplifiers
 

In article ,
Jim Lesurf wrote:
For the above reasons I still do not think that the experience you quote
justifies singling out SS amps and claiming they are unreliable if you
are comparing them with valve amps. You have not provided data that
illuminates that question, or some of the others I mention above.

Anyone with any real life experience of valve amps when they were common
will remember having to have them repaired. And I'd include, Radford, Quad
and Leak in this as I've owned them all. I've got several decent quality
SS amps here that are probably in use far more than the valve ones ever
were, and only one failure I can think of - an input coupling cap on a
home built Lindsey Hood.

--
*Geeks shall inherit the earth *

Dave Plowman London SW
To e-mail, change noise into sound.

Tube amplifiers
 

"Jim Lesurf" wrote in message
...
As an aside he Can you give me a reference for the AES standard? I
don't
have this to hand and I'd be interested to see what they specifically say.

I can find the reference for you from the AES index. You could probably do
this
yourself immediately by going to the website.

I am also curious as to what value they would accept for 'unusual'
speakers
like the old Quad 57's whose impedance varies all over the shop. :-)

It may well be that the recommendations were not in place at that time.
Another possibility, is that Peter Walker, being the man he wasm, had
special
dispensation from the Pope:-)

I am less sure than yourself that an o/p impedance of 0.4 Ohms would only
be noticable on sweep tones. However the only tests I've done on this
sort
of thing in the past involved music, so I can't comment on the relative
audibility music versus tones.

I too cannot be sure. That is why I have asked the broadcast company to
do some tests with/for me.

Perhaps if Arny is interested we could make up some 'ABX' type tests that
apply the kind of modification a 0.4 Ohm (or 1 Ohm) o/p would impose on
some typical speaker(s) and then invite people to judge for themselves? It
seems to me this would be an interesting experiment.

Indeed it would. That is exactly what I propose to do.
Can we take a typical SS amp Ro to be 0.1 Ohms? What do
you think? If so, then a 0.39 Ohm resistor would do nicely.

How much information should the panel of listeners have before hand?

Iain