Tri-amping, driver time alignment, and carbon fibre cones
 

On Mon, 16 May 2005 19:58:04 GMT, "Tim Martin"
wrote:


"Stewart Pinkerton" wrote in message
.. .

That's true, but unless you have access to good SPL measuring gear,
you may have some difficulty knowing where 'flat' is unless everything
is matched.

You don't need to be able to measure the SPL, just the signal generated by a
microphone.

Only if you know that the mic response is flat!

Take some frequency within the range of the bass and mid-range
speaker, feed a signal to the bass speaker, and measure the microphone
output. Now feed the frequency to the mid-range, and adjust its level until
the microphone is generating the same ouput.

Repeat for the mid-range and HF speakers.

You don't even need a calibrated microphone.

You do, if you want a flat response from your speakers. Also, note
that you need to be *very* careful, if you're measurements are not to
be swamped by room effects. You might even need to move into the
garden for this, with the mic suspended above the speaker, itself
several feet in the air and pointing upwards. And then you've got to
set up your angles and positioning properly to get your full family of
responses to calculate when you have a flat *power* response, rather
than just a flat axial response, or do you want a slightly falling
power response, and if so at what slope, etc etc.

Measuring speakers properly ain't rocket science, but it's
close...................
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Tri-amping, driver time alignment, and carbon fibre cones
 

"Stewart Pinkerton" wrote in message
...
On Mon, 16 May 2005 19:58:04 GMT, "Tim Martin"
wrote:


"Stewart Pinkerton" wrote in message
.. .

That's true, but unless you have access to good SPL measuring gear,
you may have some difficulty knowing where 'flat' is unless everything
is matched.

You don't need to be able to measure the SPL, just the signal generated
by a
microphone.

Only if you know that the mic response is flat!

No, you don't need a flat response from the mike, because it's measuring the
same frequency from the LF and mid-range speaker, and the same frequency
from the mid-range and HFdriver.

All you're trying to do is adjust the amplifier volume level so that a
source signal generates the same energy from the LF speaker as from the
mid-tange speaker.

Tim

Tri-amping, driver time alignment, and carbon fibre cones
 

"Stewart Pinkerton" wrote in message
...

You seem to be missing a critical point here. Three drivers in a box
will not in themselves have flat responses across their working
frequency bands, so the fact that you are supplying a flat signal in
three parts will have very little to do with the FR of the final
speaker.

Sure, but that has nothing to do with whether identical or different
amplifiers are used for each frequency band.

Tim

Tri-amping, driver time alignment, and carbon fibre cones
 

Stewart Pinkerton wrote:

You seem to be missing a critical point here. Three drivers in a box
will not in themselves have flat responses across their working
frequency bands,

Are there any that do? Flat to within what spec?


... so the fact that you are supplying a flat signal in
three parts will have very little to do with the FR of the final
speaker.

Check out the crossovers in real high-quality active speakers such as
Meridian or ATC, and you'll find that they are far from flat in their
electrical output.

If this becomes an issue, I have the option of moving the digital EQ to
before the crossover and using it to apply corrections to the full band.
(Or, indeed, adding additional EQ units.)


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

Tri-amping, driver time alignment, and carbon fibre cones
 

On Tue, 17 May 2005 20:38:24 GMT, "Tim Martin"
wrote:

"Stewart Pinkerton" wrote in message
.. .
On Mon, 16 May 2005 19:58:04 GMT, "Tim Martin"
wrote:

"Stewart Pinkerton" wrote in message
.. .

That's true, but unless you have access to good SPL measuring gear,
you may have some difficulty knowing where 'flat' is unless everything
is matched.

You don't need to be able to measure the SPL, just the signal generated by a
microphone.

Only if you know that the mic response is flat!

No, you don't need a flat response from the mike, because it's measuring the
same frequency from the LF and mid-range speaker, and the same frequency
from the mid-range and HFdriver.

True, but are you saying that you don't care what happens in the pass
band? You're not going to do any equalisation? You think that passive
crossovers *only* perform frequency division?

All you're trying to do is adjust the amplifier volume level so that a
source signal generates the same energy from the LF speaker as from the
mid-tange speaker.

And there's a *lot* more to that, than being sure that the levels are
the same at the crossover points!
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Tri-amping, driver time alignment, and carbon fibre cones
 

On Tue, 17 May 2005 20:41:44 GMT, "Tim Martin"
wrote:

"Stewart Pinkerton" wrote in message
.. .

You seem to be missing a critical point here. Three drivers in a box
will not in themselves have flat responses across their working
frequency bands, so the fact that you are supplying a flat signal in
three parts will have very little to do with the FR of the final
speaker.

Sure, but that has nothing to do with whether identical or different
amplifiers are used for each frequency band.

Indeed, but aren't you going to do anything but an utterly simplistic
frequency division? That's certainly not the way to get the best sound
from such a complex and flexible system.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

Tri-amping, driver time alignment, and carbon fibre cones
 

On Tue, 17 May 2005 22:37:43 GMT, "Wally" wrote:

Stewart Pinkerton wrote:

You seem to be missing a critical point here. Three drivers in a box
will not in themselves have flat responses across their working
frequency bands,

Are there any that do? Flat to within what spec?

There are a few that are flat within a couple of dB, but the *vast*
majority require some compensation in the crossover. Especially the
critical midrange unit. And of course the cabinet also has a
significant contribution in most cases.

... so the fact that you are supplying a flat signal in
three parts will have very little to do with the FR of the final
speaker.

Check out the crossovers in real high-quality active speakers such as
Meridian or ATC, and you'll find that they are far from flat in their
electrical output.

If this becomes an issue, I have the option of moving the digital EQ to
before the crossover and using it to apply corrections to the full band.
(Or, indeed, adding additional EQ units.)

That would be a wise move - but then you start to need really good
measuring gear to optimise results. Bottom line, this is why even the
most exotic home-builds invariably get trashed in blind comparisons
with midprice commercial speakers from the likes of B&W, KEF and
Mission. You simply can't support the necessary R&D when you're only
making a couple of speakers.

--

Stewart Pinkerton | Music is Art - Audio is Engineering

Tri-amping, driver time alignment, and carbon fibre cones
 

In article , Wally
wrote:
Jim Lesurf wrote:

2) The power amps may have different gains.

They have, but, as I said to Stewart, they each have volume controls,
none is turned up full, and the crossover provides +/-6dB on each
band.

But do you now have the same relative effective gains in the bands as
before? If not, then some of the differences can be due to 'tone
control' effects...

I don't wish to labour the point, but 'before' is history. I don't have
the 'before' numbers, so any attempt to draw comparisons is academic.

But that may not be the case for others who are considering taking a
similar path in the future. :-)

Not necessarily. :-) The point is that as you go through the
crossover region of the spectrum changing the slopes will alter the
resulting response. For example, using steeper slopes may produce a
'dip'.

Surely the crossover will have been designed to not have a dip?

With the slopes and roll-overs initially applied, yes. However if you
change the order or slope of the crossover then you may get a dip or a
peak. You alter the filtering, you alter the response.

Alternatively steeper slopes also means that the phases change in a
different way, so you may get other changes for that reason.

I thought this active crossover thing was supposed to be all
phase-compensated or something...

Depends what you mean. Some designs might ensure that the vector sum of the
outputs remains 'in phase'. However many designs do not, and the speakers
also affect this, etc. You change the steepness of the filters, and in
general you alter their phase responses as well as their amplitude
responses.

4) The effective gains for the active crossover may now be different
in the various bands.

As I say above, they're settable - even if there are discrepancies
between bands when the gains are set to zero, there's scope to
compensate in any case.

Yes. But if you are doing this 'by ear' to get the result you like,
the result may be that you settle on a different response to before.

'Before' is history. I'm not interested in 'before', I'm interested in
taking what i've got and seeing what can be done to make it more
enjoyable.

See above. :-)

This then means that you are using the xover as a tone control and
getting 'improvements' from that.

Fine by me. I don't care what happens within the system, so long as I
like what comes out.


Hence having multiple amps may not be giving you the changes you
assume.

They've added power, clarity and dynamics.

Alternatively, the changes you perceive may be due to other factors. Since
you have applied a 'tone control' and may have altered the overall
response, you can't be sure what the reason is for your perception here.
Well, you can be 'sure' in the sense of having a belief, but you don't have
the evidence that would rule out other reasons for the changes.

As I said, the tri-amped set up is the baseline. The way I see it,
it's easier for me to make changes and assess their effects.

That is fair enough. However as I have indicated, people who have
*not* yet spent the money on more power amps, etc, might find it
useful to be cautious and assess what I am saying. There may be a
cheaper and easier route for them. Difficult to say without suitable
experiments, etc, in each individual case.

I'm doing what I'm doing for myself. What other people choose to do is
entirely up to them.

I agree. That is why I am seeking to ensure they are informed on the points
we are discussing and do not assume that the use of multiple amps is
certain to be the reason for the changes your perceive. The cause may be
something else, which they might then obtain more easily and cheaply.

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

Tri-amping, driver time alignment, and carbon fibre cones
 

In article , Tim Martin
wrote:

[snip]

You don't even need a calibrated microphone.

Agreed. In this context you only need a measurement system whose behaviour
is the same 'before' and 'after' to indicate what changes have occurred.
Hence the mic has to be 'stable' in its performance, but not calibrated or
otherwise of 'lab measurement' quality.

if you want the *actual* in room response rather than a 'before'-'after'
comparison the situation would bequite difficult and require better
measurement kit.

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

Tri-amping, driver time alignment, and carbon fibre cones
 

In article , Tim Martin
wrote:

"Wally" wrote in message
...

I thought this active crossover thing was supposed to be all
phase-compensated or something...

The Behringer crossover slope is fixed at 24dB/octave, and the output is
summed to flat.

That is interesting. It indicates that the vector sum of the outputs is
(with gains equal) such that the vector summed output would look like the
input.

The problem is that the original speaker networks may not have been like
this. However that would not matter if the results were adjudged to be
preferrable.

These things simply work. I should think that, as one accumulates a
collection of amplifiers, it becomes cheaper to use an active crossover
with "free" amplifiers than to build a high-slope passive crossover.

Well, if the cost of the amps isn't a concern, then yes, that makes sense.
I also agree that well-designed active systems can work very well, and it
is potentially 'good' to avoid networks between power amp and speaker.
However I am wary of the idea that multiple amps, etc, are invariably worth
the change. A single amp and speaker (perhaps of 'better' quality) might
prove a more effective improvement in some cases.

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