Domestic sub: experiences please
 

Jim Lesurf wrote:

Seems to me that the amp doesn't 'produce' them as such, but that
they're a natural effect of clipping the outputs.

Not sure what destinction you are trying to make. Power energes from
the amp - and drives the speaker - at frequencies that were not
present in the input. Not sure what definitions of 'produce' and
'natural' you would regard as relevant here.

Reproduce would have been a better word - I was trying to say that the
square wave isn't something that the amp is trying to reproduce from input
to ouput, but that it's an artefact of the output devices (your comment
elsewhere that it might happen earlier in the chain is noted). Consequently,
the amp can have high frequency components in the output that its rated
frequency response would suggest that it cannot.


IOW, it's not a frequency response thing, unless the amp has some
sort of low pass filter after the output devices. For instance, if
you fed a squared-off 30Khz signal of low amplitude in at the
inputs, would the amp faithfully reproduce the wave at the output,
or would the corners be smoothed off? I suspect the latter.

Erm what was your "30khz signal" before you "squared off" the *input*?

A sine wave. If I put a 30KHz sine wave into an amp that is designed to have
that upper frequency response, I get a 30KHz sine wave out. If I put a 30KHz
square wave into the same amp, do I get a 30KHz square wave out?


--
Wally
www.artbywally.com/FiatPandaRally/index.htm
www.wally.myby.co.uk

Domestic sub: experiences please
 

On Thu, 31 Mar 2005 19:09:48 GMT, "Wally" wrote:


A sine wave. If I put a 30KHz sine wave into an amp that is designed to have
that upper frequency response, I get a 30KHz sine wave out. If I put a 30KHz
square wave into the same amp, do I get a 30KHz square wave out?

Assuming that the amp falls over pretty sharply, what you get out is a
sine wave 1.41 times as big as the square wave would have been without
the cutoff.

d

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

Domestic sub: experiences please
 

In message , Wally
writes
Jim Lesurf wrote:

Seems to me that the amp doesn't 'produce' them as such, but that
they're a natural effect of clipping the outputs.

Not sure what destinction you are trying to make. Power energes from
the amp - and drives the speaker - at frequencies that were not
present in the input. Not sure what definitions of 'produce' and
'natural' you would regard as relevant here.

Reproduce would have been a better word - I was trying to say that the
square wave isn't something that the amp is trying to reproduce from input
to ouput, but that it's an artefact of the output devices (your comment
elsewhere that it might happen earlier in the chain is noted). Consequently,
the amp can have high frequency components in the output that its rated
frequency response would suggest that it cannot.


IOW, it's not a frequency response thing, unless the amp has some
sort of low pass filter after the output devices. For instance, if
you fed a squared-off 30Khz signal of low amplitude in at the
inputs, would the amp faithfully reproduce the wave at the output,
or would the corners be smoothed off? I suspect the latter.

Erm what was your "30khz signal" before you "squared off" the *input*?

A sine wave. If I put a 30KHz sine wave into an amp that is designed to have
that upper frequency response, I get a 30KHz sine wave out. If I put a 30KHz
square wave into the same amp, do I get a 30KHz square wave out?



A very simple rule of thumb regarding the reproduction of square waves
is that the channel that the square wave is going through needs an HF
-3dB point that is 10 times the frequency of the square wave itself.
The output will still not be a square wave, but it'll be close enough
for audio purposes!

--
Chris Morriss

Domestic sub: experiences please
 

In article , Chris Morriss
wrote:

I SPICE simulated a 20V pk-pk sine wave of 1kHz, being clipped to 18V
pk-pk by a couple of Schottky diodes and a pair of voltage sources. I
added a series resistor to give the effect of an amp 'soft-clipping'.

FWIW A while ago I did some webpages on clipping (in the context of the
alleged effects on tweeters). This includes a plot of how the relative
harmonic levels vary with the amount of clipping. Details on the pages on
clipping and tweeters in the "analog and audio" section of the Scots Guide.

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

Domestic sub: experiences please
 

In article , Chris Morriss
wrote:
[snip]

I did use a perfect sine wave as the source. I think that in reality
many power-amps don't clip cleanly. The delays due to the internal
saturations that can occur when the NFB loop is inoperative due to
output stage clipping can give a nasty fast glitch on the edge as the
o/p stage comes out of clipping. I bet Jim Lesurf has got lots of info
about this sort of nasty.

Well I did about 20 years ago. :-)

As you say, there can be problems with the amp 'latching' onto its clipped
state and producing a nasty 'recovery' phase as the amp has to sort itself
out and clear any saturations, etc. In some cases this can also prompt a
short burst of instability - bit like some sort of 'snap recovery' harmonic
generator/oscillator.

The resulting rubbish can also be injected into the other channel of a
stereo pair.

All depends on the details of the design.

FWIW I used to try and avoid such problems by a mix of things like:

1) Define the max currents, etc, in the early stages of the amp, and the
bandwidth, etc, from stage to stage. Thus avoiding excessive open loop gain
saturating everything by vast amounts when in clipping/saturation, and
avoiding recovery taking ages.

2) Give the amp such a high output voltage and current capability that the
user would be less likely to ever reach clipping. 8-]

3) Fit a sensor that lit up when the amp clipped, so warning the more
cloth-eared users who couldn't hear it. :-)

4) Hope that listeners actually *do* listen and realise that the volume
control works. ;-

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

Domestic sub: experiences please
 

In article , Wally
wrote:
Jim Lesurf wrote:

Seems to me that the amp doesn't 'produce' them as such, but that
they're a natural effect of clipping the outputs.

Not sure what destinction you are trying to make. Power energes from
the amp - and drives the speaker - at frequencies that were not
present in the input. Not sure what definitions of 'produce' and
'natural' you would regard as relevant here.

Reproduce would have been a better word - I was trying to say that the
square wave isn't something that the amp is trying to reproduce from
input to ouput, but that it's an artefact of the output devices (your
comment elsewhere that it might happen earlier in the chain is noted).

Yes. I'd agree with that. Now makes sense to me. :-)

Consequently, the amp can have high frequency components in the output
that its rated frequency response would suggest that it cannot.

This depends upon how you interpret 'rated frequency response'.

This may be given in terms of given limits - e.g. -3dB at X kHz, or -0.2dB
at Y kHz. However this does not prevent the amp from amplifying higher
frequencies, and it may do so without excessive distortion. Just that the
gain is lower at frequencies well above X or Y. Hence the rated response
should not be taken to indicate that the amp cannot amplify or reproduce
these frequencies, just that they will be amplified by a lower gain factor.

The amp *may* produce more distortion at high frequencies, but that isn't
the same issue, and isn't necessarily indicated by the response. Many
designers choose to passively limit the NH response for various reasons,
but the amp may be able to amplify them when they are present despite that.


IOW, it's not a frequency response thing, unless the amp has some
sort of low pass filter after the output devices. For instance, if
you fed a squared-off 30Khz signal of low amplitude in at the inputs,
would the amp faithfully reproduce the wave at the output, or would
the corners be smoothed off? I suspect the latter.

Erm what was your "30khz signal" before you "squared off" the *input*?

A sine wave. If I put a 30KHz sine wave into an amp that is designed to
have that upper frequency response, I get a 30KHz sine wave out. If I
put a 30KHz square wave into the same amp, do I get a 30KHz square wave
out?

The problem here is that both a 30kHz sinewave and a 30kHz squarewave are
'mathematical' descriptions of an idealised waveform. In reality when you
try to produce either you'll get an approximation to them as you *input*.

For a 'sinewave' the real waveform will always be slightly imperfect and
hence will contain some harmonics. These may be small and at low harmonic
numbers, hence don't really matter in such discussions as these.

However for a 'squarewave' the problem is worse. A 'perfect' squarewave has
non-zero harmonics reaching up to an arbitrarily high frequency. Hence you
can't make an actual squarewave unless your source has an infinite bandwith
and its devices have a zero response time.

This means you can't actually put a 30kHz into an amp, as you can't make
one in the first place.

However you can have an approximation that contains harmonics up to some
frequency which we can assume is well above the normal audio range. e.g. up
to 100 MHz, say.

I don't know how common the practice may be. But I always regarded it as
prudent to have an RC low-pass passive filter at the input of a power amp.
This filters the HF response before the gain stages. This means that even
if the input has components up to 100MHz, their levels getting through the
filter become very small.

The problems in this area with gain devices are essential with slew rates
and rates of change, not frequency as such. Hence small signals, which at a
given frequency have a lower slew rate than bigger ones, are less prone to
causing problems. Thus with appropriate input filtering the 'squarewave'
should not cause a problem. The output will have its 'squarewave' shape
altered as the high harmonics don't get the as much gain as lower ones. But
this isn't 'distortion' in the normal sense as can be understood by the
following example. Thus in the absence of any nonlinear distortion, the
frequency response of the amp means you'd get a slightly different
approximation to a 30kHz squarewave out to that you put in.

Consider using *two* squarewaves of different frequencies which are
band-limited and filtered as described above. If the amp is distorting the
waveform the output will have components at the sum and difference
frequencies of the two input frequencies (and harmonics). Thus it would
show components at frequencies not present in the input. However if the amp
isn't distorting, then it just shows the same components as at the output
as at the imput, but with the higher components reduced in level according
to the frequency response. Thus we can distinguish between alterations
produced by the frequency response from those produced by nonlinearity. The
nonlinearity matters more in this context as it can produce unintended
anharmonic components at audio frequencies that were absent from the input.
The frequency response simply alters slightly the levels at ultrasonic
frequencies.

The above assumes the *input* was an approximation to a squarewave. If,
however, the input was a sinewave and the 'squarewave' was produced by
clipping, then the result is, indeed, distorted. But this isn't quite the
same thing as using a squarewave as the input. Nor does a partly clipped
sinewave look like a squarewave - although it will tend to it as the degree
of clipping grows towards being infinite. If you look at the webpage on the
Scots Guide on clipping I mentioned in an earlier posting you can see how
the harmonics of 'sharp' clipping change with the degree of clipping and
how then eventually tend to an approximation to a squarewave.

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

Domestic sub: experiences please
 

In article , Chris Morriss
wrote:
In message , Wally
writes
Jim Lesurf wrote:
[snip]

Erm what was your "30khz signal" before you "squared off" the *input*?

A sine wave. If I put a 30KHz sine wave into an amp that is designed to
have that upper frequency response, I get a 30KHz sine wave out. If I
put a 30KHz square wave into the same amp, do I get a 30KHz square wave
out?



A very simple rule of thumb regarding the reproduction of square waves
is that the channel that the square wave is going through needs an HF
-3dB point that is 10 times the frequency of the square wave itself.
The output will still not be a square wave, but it'll be close enough
for audio purposes!

I'd distinguish between 'audio purposes' here and the wish to see a
waveform on a scope that looks like a squarewave. :-)

I'd agree that having a flat response to something like x5 - x10 the
fundamental will give a neat squarewave shape. However I don't think that
preserving a neat shape for 30kHz squarewaves in itself matters much for
audio listening purposes - although alterations and the shape of this
waveform might be a useful diagnostic.

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

Domestic sub: car ones are odd
 

"Wally" wrote in message news:EDZ1e.7308
At present its a dual coil and bi-amped from the two channels.

Do you mean there are two drivers? In separate enclosures, one presumes?

Actually now I'm finding out more about sub design I realise how
strange are the ones that are used in cars. I'm using Weems 'Designing,
Building ...' and some info from Eminence. Knocked up a quick
spreadsheet based on Keele's formulae. Upshot is to get cutoff of
less than 40Hz I need one hell of an enclosure and at least a 15"
driver. Anyone know where I can buy BassBox in the UK? I
am awaiting a reply about buying it direct from Harries in the
US but its easier if there's a UK dealer. Google didn't help.

Now cars have small enclosures, typically around one cubic
foot (28 litres). Very low notes *are* produced, but require
very high power. So I guess you just forget resonance, impedance
and all of the rest of it and just drag the cone back and forth. No
wonder my 500W amp was into clipping when trying to fill
my large room using a 12" car sub.

Peter Scott

Domestic sub: car ones are odd
 

On 2005-04-01, Peter Scott wrote:

"Wally" wrote in message news:EDZ1e.7308
At present its a dual coil and bi-amped from the two channels.

Do you mean there are two drivers? In separate enclosures, one presumes?

Actually now I'm finding out more about sub design I realise how
strange are the ones that are used in cars. ...

Now cars have small enclosures, typically around one cubic
foot (28 litres). Very low notes *are* produced, but require
very high power. So I guess you just forget resonance, impedance
and all of the rest of it and just drag the cone back and forth. ...

I discovered that my car's subwoofer works quite well.

I remember the salesman saying something about a subwoofer but I didn't
actually appreciate this until some months later.

There was I having listened to a concert on Radio 3 at quite an enjoyable
level when, unaccountably, I tuned to some new-fangled music station
without returning the volume control to normal and I nearly had my socks
blown off ...

--
John Phillips

Domestic sub: car ones are odd
 

Peter Scott wrote:

Actually now I'm finding out more about sub design I realise how
strange are the ones that are used in cars. I'm using Weems 'Designing,
Building ...' and some info from Eminence. Knocked up a quick
spreadsheet based on Keele's formulae. Upshot is to get cutoff of
less than 40Hz I need one hell of an enclosure and at least a 15"
driver.

Car sub drivers are just like any other, they have T-S parameters.

I use a Sony XSL-121P5, which was very cheap in the Halfords sale,
as a sub for my satellite system.
According to WinISD, in a 65l reflex box it is flat to 30Hz,
-3dB at 24Hz and -12dB at 18Hz, so with room gain it will be flatter.
It's a reasonable compromise between bass-extension and WAF.

Now cars have small enclosures, typically around one cubic
foot (28 litres). Very low notes *are* produced, but require
very high power. So I guess you just forget resonance, impedance
and all of the rest of it and just drag the cone back and forth. No
wonder my 500W amp was into clipping when trying to fill
my large room using a 12" car sub.

Small enclosures give a gentle rolloff from a high frequency which
roughly matches the room gain of the car (with the windows shut,
of course.) You don't need such a lot of power if you correctly
match the levels of sub and main speakers.

--
Eiron.