Noise Weighting Curves
 

On Sun, 7 Oct 2007 20:45:41 +0300, "Iain Churches"
wrote:


"Dave Plowman (News)" wrote in message
...
In article i,
Iain Churches wrote:

My 50W PPP tube amp has a noise floor of 80µV
that's -106dB

That's not typical.

Bear in mind that traditionally valve amps had AC heaters.
These days, by simply using DC, and biasing the heaters
above the cathode can make 10dB improvement in the
LF noise floor.

In the hey-day of valve amps, iron was cheap and large
electrolytics very very expensive. Now the reverse is the
case. When valve retifiers were used, the reservoir cap was
usually limited to 47µF. Now 470µF is often seen as the first
cap in a chain with one or more chokes of 10 or 20H.
This results in a PSU of low impedance with very
low ripple.

I can understand that it might suit the agenda of some
to try to maintain the idea that valve amps are plagued
with hum and noise:-)

Fortunately, this need not be the case.


Regards to all
Iain


We aren't talking about what need be, but what is typical. The typical
valve amp lives in an open backed box with a 12 inch Celestion speaker
and quarter inch jacks for inputs. It also hums like it has forgotten
the words.

d

--
Pearce Consulting
http://www.pearce.uk.com

Noise Weighting Curves
 

"John Phillips" wrote in message
...

In article i,
Iain Churches wrote:

I have a Lowther LL26 (EL 34's PP) 26W at 0.1% THD.
It is half the power of my PPP amp, but still the noise
floor is only 120µV a very presentable -98dB.

Very presentable indeed. I guess I look on amplifier noise floors
as follows:

- specified WRT full power they indicate the absolute maximum dynamic
range available from a system. I'm afraid I'm keen on good dynamic
range.

- specified WRT the nominal 1 W into 8 ohms (2.83 V RMS) you can add the
speaker sensitivity and approximately check if the hiss will be
audible (at 1 m or at the litening position by correcting at 6 dB for
each doubling).

I have heard systems (SS systems) in the past, even at dealers, that
had quite audible noise from the 'speakers and wondered why. It seems
perfectly possible to do an engineering check to see if a system will
exhibit a number of avoidable limitations like this.

That is the reason I think it more sensible to quote noise
floor levels in µV rather than in dB, as a power amp is seldom if ever
running at full power.

I fixed a system some years ago which had a Hafler DH-100 pre-amp with
20 dB of gain from the AUX input, connected to a Quad 405 with its high
gain - somewhat untypical of the US power amps with which the Hafler
might have been designed to work, and some high-ish sensitivity 'speakers.

The combination was noisy. I guess no-one designed it. It just got
assembled. I had to reduce the gain of the preamp (checking it for
stability) to 10 dB, when it just became silent at the listening position.

I had a similar experience with a mu-follower valve/tube preamp
(gain 25dB) running into a 50W PP tube amp (input sensitivity 600mV)
and full range speakers which had a sensitivity approaching that of
Lowthers (SPL 100) The source was a Quad 66 CD player.

Taking the preamp out altogether, and fitting a DACT stepped
attenuator to the power amp, resulted in a very good sounding
system, with no standing hiss or hum.

Iain

Noise Weighting Curves
 

"Don Pearce" wrote in message
...


We aren't talking about what need be, but what is typical. The typical
valve amp lives in an open backed box with a 12 inch Celestion speaker
and quarter inch jacks for inputs. It also hums like it has forgotten
the words.


That is what *was* typical.
Things have changed:-)

Iain

Noise Weighting Curves
 

On Sun, 7 Oct 2007 21:02:56 +0300, "Iain Churches"
wrote:


"Don Pearce" wrote in message
...


We aren't talking about what need be, but what is typical. The typical
valve amp lives in an open backed box with a 12 inch Celestion speaker
and quarter inch jacks for inputs. It also hums like it has forgotten
the words.


That is what *was* typical.
Things have changed:-)

Iain



How many hundreds of thousands of AC30s, Fender Tweeds, Marshall 50s
etc are still out there and going strong in pubs every Friday night.
They are still very much typical.

d

--
Pearce Consulting
http://www.pearce.uk.com

Noise Weighting Curves
 

"Don Pearce" wrote in message
...
On Sun, 7 Oct 2007 21:02:56 +0300, "Iain Churches"
wrote:


"Don Pearce" wrote in message
...


We aren't talking about what need be, but what is typical. The typical
valve amp lives in an open backed box with a 12 inch Celestion speaker
and quarter inch jacks for inputs. It also hums like it has forgotten
the words.


That is what *was* typical.
Things have changed:-)

Iain



How many hundreds of thousands of AC30s, Fender Tweeds, Marshall 50s
etc are still out there and going strong in pubs every Friday night.
They are still very much typical.

I thought we were talking hi-fi here?
Iain

Noise Weighting Curves
 

On Sun, 7 Oct 2007 21:25:51 +0300, "Iain Churches"
wrote:


"Don Pearce" wrote in message
...
On Sun, 7 Oct 2007 21:02:56 +0300, "Iain Churches"
wrote:


"Don Pearce" wrote in message
...


We aren't talking about what need be, but what is typical. The typical
valve amp lives in an open backed box with a 12 inch Celestion speaker
and quarter inch jacks for inputs. It also hums like it has forgotten
the words.


That is what *was* typical.
Things have changed:-)

Iain



How many hundreds of thousands of AC30s, Fender Tweeds, Marshall 50s
etc are still out there and going strong in pubs every Friday night.
They are still very much typical.

I thought we were talking hi-fi here?
Iain


No - typical valve amps. Hi Fi is a begged question in that it will
necessarily have low hum and noise, which would make the whole
discussion sort of pointless.

d

--
Pearce Consulting
http://www.pearce.uk.com

Noise Weighting Curves
 

"Don Pearce" wrote in message
...
On Sun, 7 Oct 2007 21:25:51 +0300, "Iain Churches"
wrote:


"Don Pearce" wrote in message
...
On Sun, 7 Oct 2007 21:02:56 +0300, "Iain Churches"
wrote:


"Don Pearce" wrote in message
...


We aren't talking about what need be, but what is typical. The typical
valve amp lives in an open backed box with a 12 inch Celestion speaker
and quarter inch jacks for inputs. It also hums like it has forgotten
the words.


That is what *was* typical.
Things have changed:-)

Iain



How many hundreds of thousands of AC30s, Fender Tweeds, Marshall 50s
etc are still out there and going strong in pubs every Friday night.
They are still very much typical.

I thought we were talking hi-fi here?
Iain


No - typical valve amps. Hi Fi is a begged question in that it will
necessarily have low hum and noise, which would make the whole
discussion sort of pointless.

How can we play football here when you keep moving the
goal posts? :-)

I have some more info about ITU-468ARM. Are you still
interested?

Iain

Noise Weighting Curves
 

On Sun, 7 Oct 2007 23:21:27 +0300, "Iain Churches"
wrote:


"Don Pearce" wrote in message
...
On Sun, 7 Oct 2007 21:25:51 +0300, "Iain Churches"
wrote:


"Don Pearce" wrote in message
...
On Sun, 7 Oct 2007 21:02:56 +0300, "Iain Churches"
wrote:


"Don Pearce" wrote in message
...


We aren't talking about what need be, but what is typical. The typical
valve amp lives in an open backed box with a 12 inch Celestion speaker
and quarter inch jacks for inputs. It also hums like it has forgotten
the words.


That is what *was* typical.
Things have changed:-)

Iain



How many hundreds of thousands of AC30s, Fender Tweeds, Marshall 50s
etc are still out there and going strong in pubs every Friday night.
They are still very much typical.

I thought we were talking hi-fi here?
Iain


No - typical valve amps. Hi Fi is a begged question in that it will
necessarily have low hum and noise, which would make the whole
discussion sort of pointless.

How can we play football here when you keep moving the
goal posts? :-)

Just insisting on a full-sized goal. :-)

I have some more info about ITU-468ARM. Are you still
interested?

Of course. What do you have? I've been playing with making noise files
with various weightings lately, And the ARM version is interesting
when applied in third octaves to noise, and seeing how loud they
sound.

d

--
Pearce Consulting
http://www.pearce.uk.com

Noise Weighting Curves
 

In article ,
Iain Churches wrote:
How many hundreds of thousands of AC30s, Fender Tweeds, Marshall 50s
etc are still out there and going strong in pubs every Friday night.
They are still very much typical.

I thought we were talking hi-fi here?

Thought you were talking about valves?

--
*Tell me to 'stuff it' - I'm a taxidermist.

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

Noise Weighting Curves
 

Iain Churches wrote:
"Rob" wrote in message
...
Iain Churches wrote:
"Rob" wrote in message
...

FWIW the three bits of valve kit I have (Beard power, AI integrated and
EAR phono), and the one I had (Carmenta pre), are pretty quiet. The
Beard has a slight hum from the unit (transformers?), but less than some
SS amps I've had.
The Beard valve amps that I have seen perform well. They have fairly
small
mains transformers, which are bolted straight to the chassis. Setting
them
on rubber grommets (and also tightening the bolts that secure the
laminations)
gets rid of the transformer noise.

Thanks Iain - the transformers are tucked away and last time I looked they
weren't readily accessible, and I suspect they're a fair size given the
amp weighs about 35kg. And if my fettle causes a problem, they're a
problem to replace - £500 each springs to mind.

Rob. The Beard that I worked on (SP35 IIRC) had the
mains transformer mounted centre chassis under the rear cage.
Before removing it, check the tightness of the bolts through
the laminations. This might solve the problem completely.

If not, you will need to drill out the transformer mounting
holes to say M8 and then fit the grommets. Put a tab
washer top and bottom, and then ensure that the transformer
bell is grounded to chassis (a separate black wire may
be required with a solder tag at each end from one of the
lamination securing screws to the transformer fixing bolt -
clean off the varnish if required) Don't bolt the transformer
down too tight.

Best regards
Iain

Cheers Iain - it's a P100, a dual mono affair. The transformers are
under covers that flank the valves, and at my last under the lid peak I
couldn't work out how to get the covers off. But next time I'll set at
it with a new determination.

Rob

Noise Weighting Curves
 

"Rob" wrote in message
...

Cheers Iain - it's a P100, a dual mono affair.


That's a good 'un:-)

Iain

Noise Weighting Curves
 

"Iain Churches" wrote in message
ti.fi...
My thanks to all those, both on and off list, who
provided useful info on the noise weighting curves.

The two ITU curves are similar but
ITU-R ARM is a later Dolby Labs
proposal which moves the whole curve
1kHz to the right.

Thanks also to my pal Richard in the UK,
I now have a chart in Excel showing all three
IEC curves, A,B and C, plus the two ITU curves.

http://www.kolumbus.fi/iain.churches...se/ABC+ITU.jpg

Comparison is interesting. One can also see why
the old IEC "A" weighted curve is still popular:-)

Do bear in mind what these curves should be approximating. They should be
approximating the appropriate Fletcher-Munson curve for the SPL being
listened to.

Noise Weighting Curves
 

On Mon, 15 Oct 2007 08:11:09 -0400, "Arny Krueger"
wrote:


"Iain Churches" wrote in message
hti.fi...
My thanks to all those, both on and off list, who
provided useful info on the noise weighting curves.

The two ITU curves are similar but
ITU-R ARM is a later Dolby Labs
proposal which moves the whole curve
1kHz to the right.

Thanks also to my pal Richard in the UK,
I now have a chart in Excel showing all three
IEC curves, A,B and C, plus the two ITU curves.

http://www.kolumbus.fi/iain.churches...se/ABC+ITU.jpg

Comparison is interesting. One can also see why
the old IEC "A" weighted curve is still popular:-)

Do bear in mind what these curves should be approximating. They should be
approximating the appropriate Fletcher-Munson curve for the SPL being
listened to.


Is that true? The F-M curve is for the threshold of audibility and
equivalent loudness for tones on their own at various frequencies. A
noise weighting curve is designed to weight the equivalent
contribution of each frequency to a broad agglomeration. I am not
convinced that these two amount to the same thing.

Do you know of any work that has examined the similarities or
differences between these two?

d

--
Pearce Consulting
http://www.pearce.uk.com

Noise Weighting Curves
 

"Don Pearce" wrote in message
...
On Mon, 15 Oct 2007 08:11:09 -0400, "Arny Krueger"
wrote:


"Iain Churches" wrote in message
ahti.fi...
My thanks to all those, both on and off list, who
provided useful info on the noise weighting curves.

The two ITU curves are similar but
ITU-R ARM is a later Dolby Labs
proposal which moves the whole curve
1kHz to the right.

Thanks also to my pal Richard in the UK,
I now have a chart in Excel showing all three
IEC curves, A,B and C, plus the two ITU curves.

http://www.kolumbus.fi/iain.churches...se/ABC+ITU.jpg

Comparison is interesting. One can also see why
the old IEC "A" weighted curve is still popular:-)

Do bear in mind what these curves should be approximating. They should be
approximating the appropriate Fletcher-Munson curve for the SPL being
listened to.

Is that true? The F-M curve is for the threshold of audibility and
equivalent loudness for tones on their own at various frequencies.

Good point - F-M doesn't include masking.

Thing is that masking is usually applied to sounds whose frequencies are
similar - within the same critical band or in nearby critical bands. F-M
indicates few differences in the ear's sensitivity to frequencies that are
close to each other.

A noise weighting curve is designed to weight the equivalent
contribution of each frequency to a broad agglomeration.

Agreed.

I am not convinced that these two amount to the same thing.

It is my understanding that when calculating the effect of things like
masking, the F-M curve is (at least conceptually) first applied before the
masking effect is estimated. Usually, the frequencies involved are so
similar that it is presumed that the relevant F-M curve would have no
significant differential effect, and it is thus ignored.





Do you know of any work that has examined the similarities or
differences between these two?

d

--
Pearce Consulting
http://www.pearce.uk.com