Crossover Differences: Refers to how steep the slope of frequency
isolation is and whether the power used is created by the circuit to hide the loss of signal.
Throughput is the efficiency or amount of lost signal strength due to using the properties
of phase shifting that makes it possible to divide a high frequencies from low the ones.
A diplexer is often made from two different capacitors in parallel, on the high pass side,
so that the voltage split sees higher resistance on the lowest cycles, increasing the
skirt steepness. The goal for the best filters is to completely seperate in half, allowing
no overlapping frequencies, or minimizing them because they would interfere.

A Audio Triplexor Audio Triplexor which I invented, allows a smaller cap to recharge instantly for short fast

frequencies transients AND/OR long step waves, by backing up the dominant short signal capacitor with long discharge
lower frequency power cap diplexor and V2 or signal2 to make the ideal fuller range crossover with much reduced distortion of unusual wave sound.

More than 12 dB per octave is required to do this, which makes
it a second order filter or more. The 'order' (differential equations) is just how redundant, or how many stages of capacitors or inductors are used.

The triplexor can be used in the OTL type interface to hi frequency compression drivers.
I wrote in COcalc.htm to have 2.5 watts per diaphragm. Finding the current needed

Ipeak = SQR(2 x Watts x Min Resistance)
     Min Resistance

This for the prefered 16 ohm 2 inch compression driver would mean we need 0.560 amperes per channel.
A simple 450uf adds 0.88 ohms at 400 hertz but the resistance divided by time constant defines the actual
curve slope, which gives some idea of the slew rate achieved. With multiple crossovers, this matters.
Therefore with an 18 watt tube amplifier it may be possible to deliver enough current and voltage to drive
a heavy 16 ohm compression driver, without risking the burnout that is typical of OTL interfaces, even for
heavy duty standard speakers. One function of the output transformer in tube amplifiers is to convert some
of the high voltage (low current) into more current. Note that this interface allows you to use a multi-tap
transformer, with the 450uf going to the 2 ohm tap. If however you include an output transformer at the
input of your speakers to impedance match, you may be able to achieve different effects - like added guitar
distortion - by reversing the connections described. The reason for this is? Guitar amplifiers mostly distort
by overdriving the output transformer, not over-powering inverter tubes (12Ax7 12AT7) or the power tube:
6L6, 6L6G, 6L6GT, KT-88, EL-34, KT-100 It's best done by replaceing with KT-100 although the 6L6 wasn't
designed to take so much input power. 6L6GC glass beam pentode was designed in the 1930s, modified in 1950's; GC 30 watts.
Some contemporary stereophonic users value rare, old or exotic, like 300B or 350B (6L6G replacements).
In any case, great tube systems are capable of unusually bright sparkling midrange. They cost more to operate.
The technology can get very pricey. The catch is, to overdrive an output transformer to sound good, you will have to
be replacing the power transformer with one that supplies twice the voltage, which makes the 'original'
circuit nastalgia factor fade away, and with it the hyper-price advantage of reselling for more than paid.

Now, without re-transforming, the highest tap value, hopefully 500 ohm, could go into the 15uf capacitor.
In my opinion, if you want undistorted audiophile sound, you risk wasting money on 'special' mystic tubes.
On the other hand, if you want a great ELECTRIC guitar sound at vastly reduced price, you risk chasing
after the wrong type of circuits (complex organ circuits, underpowered power transformers, cheapened
down output transformer like the Ultra-Linears type) as well as high shipping charges and condition issues.
The mass produced and simple designs like the DynaKit, used cost reduction tricks like Ultra-Linear power.
They came in 3 ways, Cathode Bias, Fixed bias, and Acrosound. None used standard center taps.
Ultra-linear adds an output transformer tap to reduce cost, by eliminating the Screen-Grid power supply.
However, because the taps are not center taps, and because this configuration defeats and counteracts
distortion in the output transformer, it is NOT what is sought after by vintage guitar and instrument amplifier enthusiasts who deal in the heavy metal and distortion enhanced technical setups.

I do understand it is fun to experiment with exotic tubes, it's just that they tend to design distortion
out of the tube systems that the guitar enthusiasts are trying after that process to reverse reengineer.
This becomes a battle to out-engineer the institutional manufacturers, so that compromise is inevitable.
I have found that mixing signal paths and multiple interfaces produces solid results, with the exception,
for the same reasons, of headphone interfaces and amplification methods. You can get a clear, yet
pleasingly distorted sound this way, like having your cake and eating it too. It eliminates the killer
intermodulation distortion that many simple guitar pedal effects fail in living up to expectations in.
Today, digital post-processed effects, sound bytes, on TV & internet, have made for 'great expectations'.
When you do go out and hear a band, they do not have that 'retail' sound that they used to be proud of.
The tendency is to cave in to high profit, portable, monotoned solid state unharmonic interfaces.

The combination of 15uf and 450uf values may seem at first to be a noise source, careful examination
of the discharge voltages shows that a 450uf cap will discharge half a stored charge Q in 1 time constant.
At 500 hertz t=.002 and we will use 10 ohms considering an 8 ohm voice coil and capacitor resistance ESR.
The TC = RC for discharging it takes 1TC to dump 1/2 its Q instead of 1-1/2 to 3 TC to top-off half charge.
Instantanious voltage is supply voltage times 1-(e^(.002/.0045)) and that is .36. The 15uf capacitor
has long sinced reached full discharge so that it cannot reproduce frequencies in the 2 millisecond range.
Therefore discharges of 450uf diplexors recharges itself as well as the dominant 15uf capacitor.
This effectively cuts off signals below 500hz, although more components could be used in practice.
The voltage does not need to be as high because the large buffer capacitor yields 30 times the Q.
At 15 volts I=CV/T and if the 15uf supplies 1.5 amps, the backup supplies 1.215 times that in that time.
And that is enough to supply the coil temporarily while recharging the small tweeter capacitor's pole.

While either an active or passive crossover design has the disadvantage of introducing distortion in
square wave input and very low throughput or efficiency, it can be made with its own power op-amp
to mask the lost volume, called an 'active crossover' as compared with a passive one.
That has been a successful way of overcoming buyer objections, but such 'engineering' is not what
I am writing all these comments about. There is too many bad commentaries on the internet, and it
is too hard for reputable dealers to play combat with the manufacturer and the pundits. So many just
provide a good product or service and then they shut up, unless asked a direct question.
Well, the buyer cannot beware unless he has some help. And free enterprice DEPENDS on buyers
being 'aware' if not being informed so that they CAN BE 'Beware'.

For symmetrical linear filters, only the first-order passes phase and / or amplitude scrutiny as
being a fairly low distortion. This as I have explained, is insufficient to eliminate phase interference noise.

See diagram Below of single stage first order crossover or simplistic capacitorized signal.
SPECIAL NOTE- The Voltage LEADS the Current and thus the two are not longer in phase...
The VOLTAGE therefore has influence over the crossover frequency. At 15 volts a stereo with 40uf may
have a crossover at min. 500 hertz but because the change in current drives the speaker induction coil,
the increase to 25 volts typical of most stereo power supplies almost doubles the capacitor
used because of the tax effect placed on the power as it tries to supply larger current
at a slew rate which has increased in magnitide, (no pun intended) but not steepness/slope.
If the capacitor is not changed, the crossover point may increase to almost 700hz at 25v,
depending upon the inductance of the voice coil and/or the voltage applied.

Math: Q = C X V. Think of Q as coulombs which is really the number of electrons flux or Current.
The energy in Joules is 1/2QV so that if V doubles then Q must also try to double the 'flux' energy
that is supplied in what is now essentially half the time, because the formula assumes Time constance.
Looking at the chart below, you can confirm that when a voltage peaks, current only goes up
about half as far as the voltage peak, because it is a response in the oppposite direction.
It's not then that capacitive reactance alone or the AC resistivity due to frequency is acting
independently. This is why some DJ's run power amplifiers near full power and control
them with preamplifiers - so that the crossover interface to the speakers does not plateau.
Many guitar amplifiers are known to have a distinct plateau at half volume, especially with
more than one speaker being driven when the power supply is sized for one transducer.

This fact is just why the in-phase or signal aligned calculator shows the Cap value changes vs volts.
Using series capacitors, minimally 2 or three, the signal can match original shifted 360 deg.
For inductors, jumping them with a resistor pulls up the voltage lag and likewise helps align.
The inductors, of course, shift phase or delay in the opposite direction of capacitors.

Technically speaking, the modeling of current as half of the voltage here is because it is a
normal acceleration derived distribution. Assume 10 ohms settling time resistance for example.
A more accurate assumption would be the Wyatt approximation for balanced RC trees.
This would place the current at the 50% propagation time defined by Elmore for step input
properly at 0.693 times the propagation time. With a 15uf capacitor then 15 volts would
yield an output of 1.5 amperes max and 750ma within 1.00 milliseconds. At 25 volts input
the current reaches 2.5 amperes on a 1 milliHenry coil and 1.2 amps within 1.00 msec.

If you work with lower voltage DC capacitor discharge formulas, you will notice that
this full charge is 6.1 percent higher than the usual R*C = 1 time constant charge percent.
The difference is that electronics standards were written for DC lower voltage bipolars, not
the AC line voltages, Pentode tubes or 200 watt PA system speaker crossover capacitors.
In any case we are more concerned with the discharge rate, which delivers output faster than the charge
input rates, effectively causing the shift in phase. When I calculate T for larger ripple current
capacitors, I consider that since they may not discharge as completely in a mid or high frequency applications.
That is the reason that they act as filters - but we are only talking a few ohms, especially on inductors,
in capacitive reactance Xc/Lc, which most folks use exclusively to calculate cap values for filters.
Larger capacitors benefit in, 'effectively' or pseudo-reduced time constant T charge rates or multiples to time to full charge.

For a midrange filter a 40uf capacitor at 15 volts yields also 1.5a but then 1.2 amps in
the same time span, a more flat or damped curve than the 25 volt 15uf. At 25V 40uf
we would expect 2.5 amps (25v/10 ohms) falling to 2 amps, a 20% step distortion.
The step function happens regardless of if the signal is a spike or 'transient'.

Therefore running a capacitor so it is unlikely to fully discharge enhances transients and square waves
Increasing capacitor values modulates the 'damping', however we want it critically damped.
By underdamping we increase risk. Overdamping however reduces throughput. 3 common methods:
1)fly-back diodes 2)Zobel resistor and 3)damping (+ input) pre-coil resistor (1/2 ohm).
For balanced systems, always put the same cap / damping resistors on BOTH poles, not just one.
Most voice coils fall between .3 and 2 mH. When using capacitors in the millifarad range, the
resistor becomes a 10 ohm minimal power buffer, or else a 450uf followed by a 1.5 ohm.
This combination is also a good suggestion for jumping inductors that remove peaks or
act as low pass filters. It decreases the current lag or moves it nearer to original phase.
below example is at 500 hertz. Always Check Any Online Calculators Used for E = I X R
There are some that have order of magnitude errors. The AMPS ALWAYS equals R times input Volts.
'Distorted Distortion'
The above shows response to square wave guitar distortion pedal through a simple crossover.
You'ld say this is distorted distortion. It is why most tube guitar amplifiers do not use filters.
Solid state amplifiers have this quality of turning on and off voltage, as shown in the diagram
rather then following a perfectly curved sine wave, therefore they approximate signals anyway.
However, a second pair of wires can be used to jump over a simple transformer isolation filter
which in general can give very satisfactory results while preventing voice coil burning in smaller
speakers and compression type driver diaphragms. Phasers and Flangers are delay based distortions,
so instruments get a little such effect from a simple capacitor crossover or first order filter.

Polarized capacitors are used on the professional sound equipment to maximize lifespan:
a typical so called audio electrolytic capacitor only has a reliable life of 1000 hours.
It is a good idea to use polarized capacitors with transformers. Stepping down voltage
increases the slew rate, so that is used with tweeter capacitors. Higher wattage speakers
require stepping voltage UP. For very large wattage, 450uf is pretty small. For instance:
at 15 volts we know 1/2 * (15*15) * .000450 is the watt seconds or joules. For a 75 watt
rated compression driver or midrange, this means enough power to power that voice coil
for a time period equal to about 1500 hertz. Personally I use several branches, one with
a transformer and one without, but for each branch on the tree the voltage gets divided.
To keep things more or less in balance with the resistance of the transformer coil, I put
about a 6 ohm resistor before the 450uf diplexer capacitor pair. This further reduces the
watt-seconds for the large capacitance branch, but then I don't want all the power going
into the lower midrange frequencies anyway, as they tend to heat up compression drivers
too fast and burn them out. Often 3uf to 10uf on expensive 2 inch diaphragm, 200 watt drivers
are the norm, but you have to consider that these run at 116dB, a very sensitive response.

You also should consider when doing these calculations, that when I chose 15 volts, I was
actually saying that the supplied voltage to that branch is 1/0.693 times 15 volts.
This means I need an amplifier capabile of putting into that branch 1.565*15 or 23.47 volts.
And THAT is a typical home stereo, even a Krell at 300 watts per channel's top possible voltage.

Sometime back in the history of sound, an engineer named Bose re-designed the paper
cones that had always been made with different formulations of flat paper. He wrote pages
of articles claiming that sound other than longitudinal waves must be attenuated as
the sound of other waves are a source of distortion. He was refering to good bass.
By making the paper thicker near the center or changing the stiffness radially, he
established a standard that endures to this day. Hence to recone a speaker is very
different from repairing a speaker cone. If the original material is lost, the value
to most buyers has gone with it. It became taboo to re-cone speakers with any other type
of surface than originally used, although voice coil types are commonly modified today.
Today we know that standing waves are just the way that midranges come off any surface.
Because of the way speakers are still made, midrange tones are almost never accurate or even
easy to listen to, although the ear canal of different people does cause large changes in
each human ear, quite possibly this was the reason, along with limited microphone technology,
that the midrange sound spectrum was originally engineered out of most transducers. That along
with complaints from guitarists that acoustic/magnetic pickups distorted by means of
the electronic magnetic signals being mixed with the hollow body waves, the only available
types of guitars until the introduction of the Fender Squire.

The distortion you get depends upon several factors - the most important being the pickup type
of the guitar. The Les Paul or Humbucker type is really a dual coil - like spinning an aircraft
propeller in both directions by adding a second propeller on the same shaft and cancelling the
sideways torque. But hard rock is NOT what Les Paul had in mind - rather he wanted a non-linear
response that is resistant to intermodulation distortion (playing chords).

The Stratocaster pickup is a famous single coil, sometimes its modified as a split single coil.
Whatever else you can say, its a sure thing that you can not add two distortion pedals either
in parallel or series with a humbucker pickup type - if that distortion is a square wave type.
What happens even with one distortion and humbucker dual coils is that either the high half of
the guitar register has distortion and there is none on the low E and A strings at all, or else
the low register has good distortion but chords played on the higher notes sound smeared and
like a screwdriver was stuck between the strings with the very higest notes above the 7th fret
having a faint presence, getting even fainter as you strike the upper E string harder.

That is essentially what a humbucker is designed to do - so consider when you are working on
your sound that you won't be able to tweek you guitar or amplifier's multi-speaker system until
you begin there. The single speaker combination amplifier does help, especially the tube
balanced circuit systems, but I would be remiss if I did not point out that switching from
a tube amplifier to a multiple specialized speaker solid state amplifier will require much
more effort, often unanticipated effort, to dial in your prefered sound profile. Watch for
upcoming videos on this subject as we explore and expand our product line in this direction.

The inductive source impedance inherent in humbuckers makes them less linear than
single coil pickups and the voltage out is twice that of a single coil as well.
Of course there would be problems getting a full spectrum even distortion on chords.
This may also be why putting distortion stomp boxes between the first stage output of
a preamplifier and the final stage - often called the effects loop - does not produce
as good a result as putting the stomp box after the preamplifier and turning up the volume.
Check to see if it is possible that the extra voltage from dual coil pickups is making it
more problematic to introduce distotion into the signal, thus making you use more boost,
gain, overdrive, or other compensation labled circuitry or to go to a 24 volt pedal instead
of a nine volt - like a tubeman true tube based distortion pedal.



For background on Guitar TONE Capacitor Selection See This discussion of guitar crossovers, filters and bypass capacitors.

'Sphinxs Distoring Lead Amplifier'

Above is photo of author and amplifier wired with 2 inch compression driver capacitively coupled
to a triplexor and input transformer. The upper cabinet also contains two 12 inch speakers and the
lower cabinet is a JBL 2228. Work is in process to design a summing amplifier that will combine
several stomp box distortion effects into one, which requires careful attention to the details
of analog sound stability as well as voltage and current drive amplification systems.



For ALL Capacitor Crossover Calculator Section See :

to calculate lower frequency high throughput signal aligned multiplexor.