John Stone wrote on 10/11/2006:
John, xover frequency is determined primarily by the reactive
components.
Any filter network's turnover frequency is calculated through the values
of
source impedance, the value of reactive components, and the impedance of
the
load. Don't believe me? Then substitute a 100 ohm resistor for the tweeter
and run a sweep of the crossover. I'm sure you'll you'll see a pretty huge
change.
OK, but in this case (16 ohms across 3.2 ohms) the xover frequency is
primarily determined by the reactive components.
I did this by applying a full frequency signal to the mid/tweeter (did
exactly the same for the woofer). So a full frequency and full strength
voltage was applied to the mid/tweeter and I measured the current
actually
going to the drivers. In short, the xovers were performing perfectly.
That is, they were "blocking" (little or no current) was flowing for the
frequencies NOT intended for the respective drivers.
I have read this paragraph over and over, and for the life of me, I can't
figure out what you are talking about. What kind of "full frequency and
full
strength voltage" are you referring to? How do you accurately measure the
crossover response with a "full frequency" signal? Do you have a spectrum
analyzer?
I have a dual trace scope and a 20Hz to 200KHz sine and square wave
generator. With this I can monitor voltage while I'm watching current
flowing in the drivers.
The simplest way to accurately measure the response is to run a
sine sweep and plot the voltage/frequency curve using an rms voltmeter at
the driver terminals. Do it with and without the pot in place an you'll
get
an exact picture of what is happening.
You also keep using the term "current". How, and why are you are measuring
current (amps)?
John, this is really trivial. I put a .4 ohm resistor in series with the
mid/tweeter half of the speaker and watch the voltage drop across it. So as
I'm scanning the frequency spectrum and watching the voltage, I can
simultaneously watch the actual current flowing.
Not my rules. Function of the pots is to balance energy going to the
respective drivers to adjust for room conditions and RECORDINGS. I can
perform exactly the same function with the volume controls.
No you can't. You pulled out 2 pots and substituted them with 1 amplifier
gain control, driving the entire mid/tweeter section. You have removed one
of the pots' key functions, which was to allow shelving of the mid and
tweeter levels individually.
True, accept I always set the tweeter pot to max increase. Actually to
achieve a totally FLAT response both pots should be set to MAX and there
should be a slight increase in treble control.
So for me the only real control was over the mid range (which I set fairly
close to max as well). I still have control over the tweeter using my
treble tone control. If I need more I simply turn it up. If I need less, I
turn it down.
With a single amp, John, you have no where near the control over sound
that
I have. And with harmonics riding on fundamentals, you have no where
near
the headroom that I have.
I have no idea what this means.
As an aside, John, a totally flat response on AR-3a's is achieved with
the
pots at max AND a slight increase in the treble volume control.
Anybody who has actually measured AR3a's as opposed to just reading the AR
literature will tell you this isn't true. Here's an actual real world
measurement of the 3a's:
http://murphyblaster.com/content.php?f=AR3a.html
These measurements are backed up in a post on 8/7 by Dick Pierce to you
regarding the frequency response of the AR3a:
We have asbolutley no idea how the pots were set in this dude's experiments.
Nor do we know how well or what he did when he "restored" his AR-3a's. I
can't believe anyone takes this guy's stuff seriously.
And on top of all that, he mentions a problem in the mid-highs that the pots
would actually help correct.
I much prefer the folks at AR who really knew how to test speakers.
Further, I don't consider this advertising literatu
http://www.classicspeakerpages.net/a...ember_1968.jpg
"No, the reason for this is that Roy Allison and Ed Villchur at AR
decided on a balance that resulted in an overall downward trend
in the speaker's response toward the high end. That was their
choice, for whatever reasons.
Urban legend. Simply not true. Per Ken Kantor:
"Quantifying the high frequency response of a speaker, any speaker, is a
much more complex subject than (almost) anyone wants to think about. Short
sounds can evoke a perceptually different response than longer tones.
Distance from the speaker can have a profound effect, as can placement,
listening position and room acoustics. Even more subtly, the whole idea that
there is one uniquely "correct" frequency response is unjustifiable, as
comforting as it may be. (I believe it to be a residue of the Logical
Positivism that arose in the 1960's as an over-reaction to the magical
thinking of 1950's High End subjectivism. Kind of like now...)
To this day, there is little agreement even between very experienced
professionals about how to properly characterize what "flat" response is. AR
had their methods, (which evolved over the years.) Within the limitations of
those methodologies, AR always tried to make accurate speakers, at least
near the top of their line. "
Further, if you read the other reference, you can see that AR recommended
that folks set the pots at mid point to correct for what the record
companies were doing in 1968.
The midrange efficiency is about
2 db less than the woofer, and the tweeter is another 2-3 dB
less efficient than that. That's what the people of AR designed
it to be." (end quote)
A "slight increase" in the treble control will not achieve a totally flat
response as is clear from the actual response measurements.
So you are saying that the pots should be at flat and we need a more than a
slight increase in treble?? OK, so?
Jerry, your ignorance of even basic loudspeaker design principles is
appalling. If I didn't know better, I'd swear you were joking. The
tweeter
network in the AR3a consists of a single series element; i.e., a 6uf
cap
feeding a load consisting of a tweeter and a 16 ohm tapped resistor.
For
all
intents and purposes, when you run the tweeter pot at full, you simply
have
a tweeter and a 16 ohm resistor in parallel. The crossover frequency is
determined by the value of the series cap AND the impedance of the
load.
Removing that 16 ohm resistor alters the load (increases the
impedance),
which in turn alters the crossover frequency, moving it downward. I'm
quite
sure what you've done IS significant,
My ignorance, huh? John, you spouted a lot of "facts", but not a single
number. Are you lazy, John, or perhaps you don't know how to do the
math.
Let's just see how significant.
DC resistance of tweeter is approx. 3.2 ohms
So with the 16 ohm pot full across - net resistance = 2.67 ohms or a
difference of .53 ohms
Sorry Jerry, but yes, your ignorance, or perhaps just lack of attention to
detail, as again demonstrated above. DC resistance is utterly meaningless
to
this discussion, as the last time I checked, tweeters required AC to
produce
sound. Yet you use resistance interchangeably with impedance as if they
were
one and the same. The "nominal" impedance of a tweeter with 3.2ohm dcr is
4
ohms.
Yep, the reactive component combined with the resistive componet yields an
"effective" Z. Z= 3.2 +jwL
But the resistance components will respond exactly as I stated. There is no
reactive component in the pot.
If you take that nominal impedance in parallel with the 16 ohm pot
(the pot will have an impedance close to 16 ohms given that it is largely
non-inductive) the 2 in parallel yield a resulting impedance of 3.2 ohms.
With the pot totally out, you shift the impedance up by a minimum of 20%.
Real, real significant, John ... huh??
Um, yeah, I'd say by any measure a 20% shift is significant. And in a
previous post, you yourself said so:
John, what are you talking about?? Now whose ignorance is showing?
With the pots out, naturally LESS current flows ... all of the time. So
what?? The end result of that is the speakers have increased in
sensitivity.
Your issue was that we have significantly altered the xover frequency. That
is, taking the pots out causes current to flow MORE at some frequencies than
at others. That is we have a frequency altering impact.
(This is awful trying to create any tables in Usenet, but let's try again
and I'll even use your numbers.) Here we show the current flowing through
the high frequency driver under both conditions:
Z Z I @ 2 v I @ 2 v
Hz Z for Cap w/o pot with pot w/o pot with pot % diff
3,000 8.8 ohms 12.8 12.0 .156 .132 14.7%
4,000 6.6 ohms 10.6 9.8 .188 .163 13.5%
5,000 5.3 ohms 9.3 8.5 .215 .188 12.5%
6,000 4.4 ohms 8.4 7.6 .238 .210 11.6%
John, if we look at the last column, we see that the difference in current
varies by 3% between 3000Hz and 6000Hz
As for the improvement in sensitivity, all I can say it's significant
and very noticable.
Those are YOUR words, Jerry. So which is it? Significant or not? 20% is
way
beyond the impedance tolerance AR specified for the drivers alone, so the
crossover is indeed operating into a significantly different load than
intended, with the resultant shift in xover response as previously stated.
Yes, the speakers (not drivers) are more sensitive and it is significant.
The change in xover frequency is INSIGNIFICANT! See the table above and
show us your proof that this is incorrect (I even used your numbers for
impedance).
More importantly 20% is the MINIMUM shift.
Minimum shift in what? Current doesn't change by 20%. From my table
there is 14.7% more current at 3000 Hz and 11.6% more current flowing at
6,000 for the same applied voltage. Over the frequency range of where the
tweeter operates, my guess is the average increase in current through the
driver will be around 10%.
And, YES, that will make the driver put out more SPL.
In reality, the tweeter (and
midrange) impedance curve is nowhere near flat, and deviates considerably
with frequency. Given that these drivers use no ferrofluid they will
exhibit
a fairly high Q at mechanical resonance ( which I'd estimate around 300Hz
for the mid and 3kHz for the tweeter), yielding an impedance peak at
resonance perhaps double the nominal value. The driver impedance also
rises
gradually with increased frequency, due to the voice coil inductance. So
the
pots were there for 2 reasons: provide level adjustment, and to provide a
measure of impedance stabilization for the crossover. By removing them,
you
have changed things a whole bunch.
John, I don't agree with your conclusion at all. I even calculated the
current flow through the tweeter at 3000Hz both ways. At 3000 Hz there is a
14.7% difference .... @ 4000 hz the difference is 13.5% .... @ 5000 hz the
difference is 12.5% ... @ 6000 hz 11.6%
What the pots did is "pad" the output of these drivers to bring them closer
in line with the woofer. With the pots gone the output is higher.
Tell us, John, do you
think there is a human anywhere ... that could actually hear this
difference?
You said you could hear it yourself, so I'd assume others could as well.
Perhaps your arguments would have some weight, John, if you backed them
up
with the math.
There's plenty of "math" above. And it least its based on correct
parameters
(impedance rather than resistance), which makes your "math" less than
useful.
I re-did the math using your numbers (even though I don't agree with them)
and the result is still the same. No significant change in frequency
dependent current.
Pushing down the crossover frequency puts more strain on the tweeter. Like
it or not, those tweeters were notorious for marginal power handling
ability. A.75in voice coil on a paper former with no ferrofluid cannot
take
much power without burning out.
John, your argument bears weight if I removed the pots and single amped. I
don't! I adjust the voltage delivered to the mid/tweeter section with my
voilume controls. This half of the AR-3a's are NOW more sensitive, so I
send LESS voltage. The speakers simply require less voltage and that's part
of where I get more headroom.
Double check my math if you like, John, but there is simply is NO
significant change in frequency dependent current flowing. There is a
change that varies by 3% over the range of 3000 to 6000 hz. I don't
consider that significant and I don't believe there is a human on this earth
that can hear that difference.
Jerry, as others here have told you countless times, nothing is to be
gained
in your configuration compared to a single amp, as long as the driving
amplifier has adequately low output impedance and sufficient undistorted
power to drive the speakers to the desired level.
It's possible that no such amplifier exists. Some could better than others,
but the AR-3a has a very complex impedance map. I don't know if any
amplifier can perform over the entire range ... as well as an amp over a
much narrower, easier and better behave impedance range.
Further, the amp driving the mid/tweeter is operating at a completely
different level in terms of both voltage and current. In addition, it isn't
burdened at all with fundamentals. Whenever we mix fundamentals with
harmonics we create the opportunity for IM distortion. When these
frequencies are NOT mixed that opportunity is greatly reduced.
Virtually all competent
amplifiers act as voltage sources, and within the limit of their power
output capability, perform largely independent of the load they see.
I don't believe that for a minute. One requires enormous "faith" to assume
the simply because an amp can produce a nice wave form open loop, it will
perform the same when driving a dynamic speaker system. I believe that much
is going on inside an amp under continually varying loads and very complex
music signals. I believe they are anything, but pure voltage sources.
The AR3a designers knew this and designed the speakers accordingly. You
have
never provided a shred of proof that your configuration solves any problem
that couldn't be solved with a single more powerful amplifier. To now
extend this argument to ALL amplifiers, regardless of power, is absurd and
based totally on unsubstantiated generalizations about power amplifier
behavior. Yes, by using 2 amps, you can now adjust the mid and tweeter
outputs to levels above the woofer, but that was never the intent of the
original design.
I've demonstrated that the impedance seen by two amps is radically different
than the impedance seen by a single amp. The impedance seen by two amps is
much, much more stable. In an single map with widely varying impedance by
frequency, we'll see widely fluctuating current flows over the frequency
range ... at the same voltage. This "instability" just has to have
ramifications.
Each amp in a bi-amp of the AR-3a's just has a much, much easier job to do
and in my opinion, will always out perform a single amp. Now at the same
time that I say this, I want to go on the record that I don't believe this
is the case with every speaker system. I have bi-amped a fairly efficient
JBL system and while I can hear some minor difference, they are minor. I
believe the AR-3a benefits greatly from bi-amping because of the very
difficult frequency dependent load it presents to any single amp.
I've demonstrated that enormous headroom can be gained in the mid-tweeter
amp. Then others have argued that even if clipping occurs in the woofer
amp, the "harmful" harmonics are totally isolated from the mid-range and
tweeter.
As for adjusting the mid/tweeter above the woofer, I never claimed that as a
benefit. It's true I can, but I don't.
As for my amps, after looking over the facts ... I believe I have far,
far
more voltage headroom than anyone with a single amp.
I know what you believe. Now, please explain the basis of this belief.
Let's
see your experiments or charts showing your "biamp" configuration with
ancient low powered amplifiers of dubious quality having far more headroom
than a single 250 watt/ch amplifier with total stability down to 2 ohms.
Your claim, your proof.
First of all, I don't believe any claims of stable down to 2 ohms. I have
some pretty large 2 ohm resistors. Wanna burn out your amp?
Next, my typical voltage on the mid/tweeter amp is around 2 volts for fairly
loud music. My rails are at 40 volts. Now look at the the voltages of your
amp with fundamentals and see what your real headroom is.
Which brings up another question: If using separate amps gave such a
huge
improvement, why didn't AR suggest it in the first place?
I don't believe at the time that the AR's were developed, people were
even
experimenting with bi-amping for home use.
I don't believe back in 1968 ANYBODY was thinking about bi-amping for
home
use, but please share a reference showing that I am wrong. Quote any
article from 1968 or earlier recommending bi-amping in the home.
Having lived during that era, I can assure you that biamping was indeed
being done in home systems well before 1968. Ever hear of the Marantz 3?
It's an electronic crossover from the '50s. In fact, even Heathkit offered
one called the XO-1 which also dates back well before 1968. There were
others.
I never heard of any of this and was a Heathkit customer, so I can't imagine
it was very popular.
Regards,
Jerry