> ... Phil- Hide quoted text -
>
> - Show quoted text -
Who could ever have hoped that grown men would not argue about speaker
Z measurements? Nobody?
Anyway, keeping speakers away from sound reflecting surfaces does have
a significance. Only a fool would test a driver with its front face
sitting on a bench, or with a wall a foot away.
But what about enclosure effects on drivers?
To make a speaker, ie, enclosure with driver mounted in it, and with
baffles and sound absorbant materials, one needs to know the raw data
about the driver when it is well away from an enclosure or reflecting
wall. Testing in an average room with driver with magnet down and
sitting on a an upright housebrick on a bench, or dangled on a wire
from cieling will negate 95% of effects on Z by nearby objects. The Z
plot will reveal the Fo, the resonance in free air. There are other
tests one can do to determine all of the Thiele and Small parameters,
and then these may be entered into a program like WinISP which will
give you a calculated box size for a predictable bass response. Then
the speaker box + driver fixed in can be completed, and another
impedance test then made with speaker set up in a workshop in similar
situation to how it is used in a typical listening room. It is done
without any crossover, and the use of a power amp set at 10Vrms output
feeding through 1k0 gives a virtually constant 10mAac, and so tweeters
can't be damaged with bass F. The Z at the lowest part of a given
driver's F band will be quite difference from the free air test out of
any box. This Z is the real and relevant Z which must be matched to
work with a crossover filter to give the wanted acoustic response.
Once the Z is known, the acoustic response of each driver in turn can
be tested with pink noise without any X-over with the same level of
applied LOW voltage at all F, ie, with say 0.1Vrms straight from the
amp which has Rout < 1 ohm.
If one builds a 3 way speaker, the acoustic responses of LF, MF and HF
should tell you where the lowest level of sound is made between 40Hz
and 20kHz. You might find these F are where LF and tweeter are -3dB.
In a typical speaker, you may find there is a level response ( +/-
2dB ) between say 80Hz and 200Hz and that this level is lower than
anywhere else above 200Hz and below 20kHz, so this acoustic level sets
the sensitivity of the completed speakers. Its not unusual for LF, MF
and HF drivers to have peaks in their response +12dB above the
REFERENCE level of say 80Hz to 200Hz. So some means of attenuating the
amp voltage reaching all drivers has to be designed to ensure a flat
input voltage gives a flat acoustic response, and that the L&C used
are "terminated" by sufficient R to avoid dips in Z due to resonances,
and that the driver always "see" a lowest possible R as it source. So
the work of making a good crossover is NEVER easy, and NEVER is a
simple task done by an online calculator, with no understanding or
experience with LCR theory.
Most ppl just shove a resistor in series with MF and HF to reduce the
higher acoustic output of MF and HF over the low bass output from most
bass drivers. Let us assume each and every driver has an R+C Zobel
strapped across it so that its mid band Z value is maintained at all F
above the midband. The Zobel adding is called "impedance equalizing,
and makes the driver look more like a pure resistance above the mid
band Z, and equal in ohms to the midband Z.
Let us assume the equalised Z of a driver in a box = 7 ohms. Just
adding say 7 ohms in series makes the total Z roughly 14 ohms, giving
6dB attenuation. The driver is now driven by 7 ohms plus amp Rout so
perhaps by 8 ohms source resistance. People could say this is no good,
because the damping factor has become hopelessly low, only about 1.0
in this case, when theory says we should have DF < 10. The LC filter
designed for 7 ohms will have very different attenuation slope rates
and X-over points if loaded with 14 ohms. One might then decide to
make the Z "seen by" the preceeding LC filter to be 7 ohms, and have
the driver "see a lower source R". This means that after the LC, one
has 3.5 ohms in series with driver and its parallel R&C zobel, but
with an additional 7 ohms across the driver, to make the driver Z look
like 3.5 ohms. The driver sees 3.5 ohms plus 7 ohms in parallel and
the F below the midband Z, ie, 2.67 ohms, and the driver is better
damped, and amp sees 7 ohms, and LC sees 7 ohms, and everyone is
happy, except that the sensitivity has halved for the driver. Well, if
you want a truly flat response with 3 drivers, you usually have to
sacrifice some power lost heating resistors if you are determined to
have a flat response. No free breakfast.
There are other ways to avoid losses in speakers due to driver Z
variatons and R losses. One might use tri-amping, very expensive with
tubes, and you'll waste fat more heat, or you have an audio speaker
voltage transformer for MF and HF. It usually is never necessary to
attenuate bass levels, considering so many speakers are bass deficent,
because they have been designed to sell well and be marketted easily
with high quoted sensitivity figures of at least 87dB/W@1M. The tranny
to attenuate MF and treble can have Fsat at 50Hz, at say 50Vrms input,
and have a series C between its primary and amp. It will need to be
very carefully designed so there is not a horrible low Z formed at the
Fo between the series C and primary inductance. Taps down a winding at
1.5dB steps are wise, giving 0dB to -12dB. By this means, the
crossovers and the following drivers are driven by a source resistance
lower than the amp Rout, so DF improves, ie, gets higher. The amp then
sees a higher load, welcomed, because most tube amps function best if
load R is higher. So if one has 2:1 tranny voltage ratio, with sec
load = 7 ohms, the amp will see 28 ohms, and no losses occur in series
& shunting R.
I don't know ONE manufacturer who uses transformer attenuators for
multi drivered dynamic speakers.
The exceptions are for ribbons and ESL which form a tiny % of speakers
sold.
So sure, speaker impedance is hugely important, if we wish to save our
amplifiers from overheating, and if we wish to hear the best sound,
from using the available electric energy most efficiently. Efficiency
becomes important when the Watts become expensive, or when we have
made the decision to have so few Watts at the outset, ie, chosen a
lone 300B to fill a large room with divorce causing sound levels at
low distortion.
Patrick Turner.