History: 25 Years CD Ortofon, Garrard, Decca, Tannoy The Sound of Tubes and Transistors
Your Desert Island Discs Lp Cleaning & DIY Cleaning Formula Elisabeth Lugt Soprano
Turntable & Cartridge Adjustment Marie-Claire Alain, Organist
The Universal Stabilizing Ring SACD: Upsampling & Noiseshaping Decca London Ribbon HF Loudspeaker
DIY: Turntable Weight/Clamp The Joy of Well Positioned Speakers
URSR: Review in HiFi World LP Records for Sale Vintage Equipment
URSR: Picture Gallery The Long Playing Record Guide
The TD124 page The SP10 Page
Joachim Bung: Swiss Precision Stefano Pasini: German Perfection
Mengelberg's St. Matthew Passion Plinth for Technics SP-10 mk2 Record Shops in Amsterdam
Paris Jazz The Sound of The Turntable Mat Acoustic Revive R77 Generator
CLASSIQUE 777 Lp Record Covers The Treasure Trove How to Correct WARPED Records
Klaas A. Posthuma - Remembered Ernst Lumpe: Allegro-Royale Pseudonyms Nostalgia: Violinists on 7" 45 rpm
Steinway-Lyngdorf Model D Infinity KAPPA 7 A Loudspeaker Systems DIY - Draaitafelconstructie - in Dutch
The Turntable Mat - Page in Russian Ajuste de un giradiscos NOTES: The Belt Drive Turntable
Phono Cartridge-Headshell-Plinth Porgy and Bess Active Loudspeaker System
Phono Cartridge Optimizing Gold for Bernard Haitink Rabco SL-8E Tangential Tonearm
Mercury Living Presence Records HiFi Tunes: DAS KLASSIKERBUCH DIY:Tonearm Building
The Bullet Plug Violinist/Violist Paul Godwin The Remington Site
Mercury Recordings on Fontana CINERAMA and Trinaural microphone Placement Concert Hall - Musical Masterpiece Society
Cook-Livingston Binaural Recording System Willem Mengelberg and his orchestra filmed in Epinay in 1931- Contemporary Records - Lester Koenig-
Soundfountain Audio and Music Bulletin logo

hobbyists's views for hobbyists
How to Use an Electronic Crossover Network -
Adding a Subwoofer -
Building a 3 Way Active Loudspeaker -
Luxkit a2003 by Luxman

Thanks to hosting company ARVIXE, my pages are in disarray. Arvixe moved the server from Texas to Provo, Utah without warning and without providing the necessary codes, password and server address, while the yearly payment had been made and acknowledged by Arvixe for the period until October 2023. We certainly hope that the inconvenience to you, the visitor, will gradually end as the pages are now hosted in the Netherlands.




The information is meant to give you more insight
when you consider going active.
Page first published January, 2004.

Characteristic with 12 dB for low and high section and 6 dB slopes for the mid band



Building an active loudspeaker system involves various ingredients and a number of choices:


1. The availability and choice of a suitable electronic crossover unit or module.

2. The determination of how to divide the frequency band (the number of ways per channel).

3. The choice of the appropriate loudspeaker units in relation to efficiency, frequency response (bandwidth) and linearity.

4. The choice and availability of power amplifiers


Units, cabinets, electronic crossover and amplifiers, all these are interrelated. So we have to deal with many variables which make the road to harmony, excellent transient response and efficiency a difficult one.



1. The main advantage of an active system is that the loudspeaker units are directly connected to the amplifiers without the use of filter sections with capacitors, coils and resistors. Especially the omittance of coils used in passive filters (which are placed in series with the voice coil of the loudspeaker unit), make the units do their job in an optimum manner. Coils in series with the voice coil generally deteriorate the behavior of the unit. That is why designers often use coils with a ferrite core or a specific transformer in order to keep the resistance value of the coil low.

There is another advantage. When filtering with a 12 dB slope, there will be no capacitor in parallel needed which has a high value when choosing a low crossover frequency, let's say between 300 and 600 Hz. And for 18 dB the total of uF will be even higher. Capacitors store energy. That is a bad thing in loudspeaker filters. Omitting capacitors as well does mean that the units are working better.

2. The proper loading and filtering of loudspeaker units is translated in a cleaner, clearer and better controlled sound. Only of course if the volume for each unit (acoustic loading) is well chosen. This volume does not necessarily need to be the volume calculated by a computer program for filter and crossover. It can be empirically determined and most of the time it is smaller than the formula calculates because in electronic filtering there are no coils like in passive filter sections.

3. Another important advantage is that one can choose uncommon units which are generally not used in commercial systems: 38 cm bass speakers, horn loaded bass and/or midrange speakers, high efficiency ring radiator tweeters, ribbon tweeters, etc. The choice is yours.

4. Also speaker units can be selected which are built with specific materials for cones. Paper for instance has gone out of fashion. Paper has the advantage that it does not store energy as some modern materials do. Paper gives more detail as it does not smooth out (equalize) the curve. But it needs more attention than, say, polypropylene.

5. Units with heavy magnets are expensive. Units with alnico magnets are expensive. So these are hardly used in commercial systems. When constructing or assembling an active loudspeaker system, such units can of course be incorporated. And if they are too expensive, high quality second hand units may be acquired.



A disadvantage of building an active system is that units have to be well chosen. Units with relatively linear frequency responses are to be prefered.

The objective is to avoid passive components which are generally used to linearize the frequency responses and the impedance curve.

Remember: most electronic crossovers have slopes of 6, 12, 18 and only in some cases 24 dB per octave. That's it. No 9 dB or 16 dB. The correction of the impedance of the units in an active system is of course possible by inserting a well calculated network in parallel (a capacitor, a resistance and sometimes a coil). The effect of such a correction is that the high end of the curve is less free sounding. That is why it is better to choose a unit which has an impedance curve which comes close to a straight line, far into the high end of the curve.

If you want to know what I mean just check the impedance curves of a few units and compare the impedance curves with the frequency curves. A good source for learning is the Visaton website.

Whether it is advisable to apply these corrections depends on the configuration and also can be a matter of personal judgment. And of course what you hear.


Best System: Two-Way

The best loudspeaker system is a 2-way system, whether it is a passive or an active system.

In a 2-way system a wide frequency band is reproduced by a low-mid unit, starting at the fundamental frequency up to about 3000 or 4000 Hz. (or sometimes even higher).
The advantage is that a harmonious buildup of the sounds of instruments is not disturbed by a filter section early in the audio band, nor is it disturbed by phase shifts caused by the mechanical and electrical differences in woofers and midrange units.

So keep the lower end integer as much as possible. The ear is most sensitive to irregularities in the low and mid frequency band up to at least 2000 Hz.

That is why the best system is a 2-way system. That is also why in olden days the best 3-way systems left lows and mid frequencies in tact and filtered at around 600 (and sometimes at 900, 1500 or even at 2000 Hz.). You could encounter a 3-way system with crossover frequencies of 2000 and 6000 for example. Even 7000 Hz. was chosen if the crossover point between woofer and mid band unit was relatively high. Then specific tweeters could be used like the Decca Super Tweeter and Fostex Ring Radiators. Such high crossover frequencies guarantee a harmonious build up in the low and lower mid sections and keep the midband open and whole.


Many Ways
Lead To Rome

So your active system could well be a 2-way system with one or two 7, 8, or 10 inch woofer/midrange unit(s) plus a tweeter. The determination of the crossover frequency depends on the capabilities of the chosen tweeter (linearity and power handling) as well as the technical specifications of the low-mid unit and especially the linearity of the frequency curve. In this way you will avoid all the pitfalls which present themselves when constructing a three-way system with three amplifiers and three different units for low, mid and high frequencies.

The choice of units for an active 2-way system can be inspired by existing DIY sets or projects that do not use complicated filters but serial filters or simple sections with 12 or 6 dB slopes or combinations.

If an existing DIY design from a shop is chosen as a basis of your active system, see to it that the volume for the woofer/low-mid range unit is recalculated as you leave out the passive filter section with the coil in series. The internal volume should be less.

position of sub woofer
Many high fidelity enthusiasts stick to the classic configuration of two satellites and one (or if the budget and the room allows) two sub woofers. Especially analog fanatics go by this configuration to which extra speakers can be added and used when viewing movies.
position of sub woofer in surround sound setup
In all cases the main speakers can consist of various transducers in appropriate cabinets.
3-way and 4-way active with sub woofer

Simple Solution

If you want to go active by just adding a subwoofer to a pair of satellite speakers, you can start by using existing 2-way systems and combine these with just one subwoofer, home built or manufactured commercially. This is the most common configuration. There are many options.
closed box and ported enclusre comparedNot every sub woofer sounds the same. The best options are the closed box systems, and the front loaded (band pass) systems. The Janis subwoofer is in principle such a band pass system.

The advantage of the band pass system is that it sounds like a closed box system with a continuously decreasing excursion. These systems do need woofers with strong magnets and these are of course expensive. That is why so many designers use the cheaper units in bass reflex systems.

The disadvantage of bass reflex (ported) enclosures is that they have an irregular excursion curve with a dip in the 60 to 80 Hz. region. There the excursion is minimal and in fact not logical.

Back loaded horns do not go deep and if you want to reproduce low frequencies through a horn, the mouth of the horn should be extremely large and the way the sound travels has to be rather long. Even if your listening room in your house permits, these horns can occupy the entire back wall of the listening room.

Transmission line systems suffer from the same phase shift as bass reflex systems. And again the cone excursion is restricted in that area. There the sound is slender if compared to a big unit in a closed box. In many transmission lines the lowest frequencies are emphasized and are slow.

There are also designs for open baffles. But the side panels have to be very large (as in the case of the horns). And there are various dipole designs too.

Each and every box, baffle, column or screen has to reproduce sounds which in real life are produced by large instruments (organ, longest and thick strings of the grand piano, big drum, large orchestra), or by instruments with a entirely different shape (violin, cello, flute, etc.), it certainly is no mean task to get a loudspeaker system "right". In this context the question that is often repeated is: Can a large electostatic loudspeaker reproduce a violin sound?



To Boost Or Not To Boost

Many hobbyists and audiophiles use 24 dB Linkwitz filters to separate the woofer completely from the transducers for the mid and high frequencies. It is up to you if you want to opt for such steep filter slopes. It may be that when gaining more experience while assembling an active system you will find out that less steep slopes may result in better phase and a more harmonic and lively sound.

Also you are free to use electronics that help you correct the frequency curve of the unit in the subwoofer enclosure, and can boost the lowest frequencies as some designers do propagate. This is however a method which I personally do not like at all. Better is to give the unit its appropriate volume and curve, as the parameters enable you to calculate.

Also Motional Feedback circuits do hamper transient response. The correction generated by the feed back takes the tops of the pulses of the woofer and other reproducers that are corrected this way. And as the lows are the basis on which the harmonics of the entire system are build, naturally the feedback correction can be heard in the rest of the frequency band, can be heard in the total sound picture. The reproduction may sound clean, but has a less realistic character.


Bass Enclosures

Various subwoofer designs

The Boffle

An ingenious solution which gives a strong and well controlled bass reproduction is that of The Boffle which was designed by H.A. Hartley. I read the discription of The Boffle in one of the old 1950s issues of The Gramophone. The author of the article promised high quality sound.

Hartley said about it: "I suppose I am an old-fogey purist, but it does seem to me that as a speaker in its enclosure has to make an attempt to reproduce every kind of musical noise, it must have no "personality" of its own. It must be absolutely unbiased; it must add nothing and take nothing away."

And that is exactly what The Boffle can do. Of course the music making depends on what unit you choose.

The Boffle is in fact a cube with an edge of say 50 or 60 cm. My version had a Philips dual cone loudspeaker unit in it. It was a full range unit with a stiff suspension (corrugated surround). There was no separate high frequency unit connected to this speaker. So no extra reproducer for the highs, connected with a capacitor in series. In olden days full range units were often used and were a pleasure to listen to as there was no phase shift (which means no time delay) as was often experienced in 2 or 3-way systems if the design of the filter and the position of the units on the baffle were not correct. Being used as a teen ager to my boffle, I was not pleased when I heard a 2-way or 3-way system even from a renown manufacturer.

Any speaker system which uses a full range driver will give you insight of what coherent sound is. Open baffles are more appropriate for mid range reproduction and for a woofer/subwoofer The Boffle is a very good choice.

The Boffle as seen below was drawn from memory by me. The compartments are separated by sheets or panels of 5 mm felt. The original plan stressed that the holes in the felt should vary the way the drawing shows. The holes directly behind the unit should be large in order not to obstruct the flow of air. Also a too small hole at a is not advised. The second half of the boffle (between a and back panel b) also has variations to work correctly. This how I built The Boffle and never adjusted the holes of one or more of the panels.

Principle of the BOFFLE

The Hartley Boffle uses felt panels. Felt was not readily available at the time and had to be ordered. It was also expensive. So in my system I used sheets of soft board. That was practical because the holes could easily be cut out and the panels could be replaced if necessary. On top of that they were easily fixed to the wooden bars. Knowing from years of experience what damping materials do and how to apply them, I suspect that soft board which is a "harder", less absorbent material, gave a very good low mid register, especially the sound of the grand piano had tangeability.
I saw a drawing on the internet of a boffle without the hole in the back panel. And there was also a drawing of a cabinet which was called Boffle by the owner of the web page but was in fact an "ordinary" loudspeaker enclosure.

Richard Schmetterer, President of Hartley Loudspeakers, Inc. in Wilmington, North Carolina writes on April 17, 2013:
"The Boffle was patented by Hartley Turner Radio LTD. of England. This company became the Hartley Co. Ltd of England and now Hartley Loudspeakers, Inc. in the US. The Boffle design was done by Philip Keston Turner (P.K. Turner) and the patent was granted in 1935."

See: A History of Hartley Loudspeakers


A high level of quality can be achieved when using commercially manufactured 3-way systems connected to an active crossover filter. The better quality is specifically attained with satellite three-way systems which are equipped with rather large woofers, preferably 12" or 10" woofers in parallel, and placed vertically in the enclosue or mounted in individual enclosures that are stacked and if you have a large listening room you can place them horizontally.

I have discovered that large woofers for the lower mid section do follow up the frequency band of the subwoofer in a much better way than small sized units do. Large woofers provide a far more realistic sound.

The choice of crossover frequency will generally be around 100 Hz. Such a system can later be updated by adding a second subwoofer.
Although it is said that frequencies under 150 Hz. are non-directional, it does matter whether one or two subwoofers are being used.

Another option is to go active with the small satellite 2-way systems and omit the specific subwoofer, but combine 2-way satellites with large woofers as in a full blooded 3 way system. These woofers have to have their appropriate (well calculated) cabinets, one for the left and one for the right channel. The satellite 2-way systems can be home built and optimized in relation to the woofer. The appropriate turnover-frequency should be determined in practice but could be situated in the region from 350 to 600 Hz.. Such a system has the advantage that it leaves at least three octaves of the fundamentals intact.

Classic 3-Way

The most original and most difficult option for assembling/constructing an active loudspeaker system is not to use existing 2- or 3-way satellites with subwoofers, but to calculate and build the cabinets for the low, the mid and the high frequency bands oneself.

Whether the choice is for incorporating a subwoofer or not, the midrange units must have the ability to perform well up to round and about 3000 or 4000 Hz. As said earlier, a very musical system can consist of a 12" woofer, one 5" or 7" midrange unit and a quality unit for the high frequencies. Crossover frequencies could be 600 and 3000 Hz. in case of a dome-tweeter or horn loaded ribbon tweeter. In the event a ring radiator like one manufactured by Fostex is chosen, the crossover frequency for the treble should be 7000 Hz. With such a system a very harmonious and lifelike reproduction can be achieved especially if you choose high efficiency units. That is the classic way.

There are many options and many possibilities and also many pitfalls. But you can begin by adding a subwoofer and crossover to an existing system. And that can give you already increased listening pleasure, a lot of fun and it is a learning school for gaining knowledge. After some time you can continue from there and expand your active system. Or not. It is up to you.


It takes a lot of calculation and experimentation to find the best units for a three way system with crossover frequencies of 100 Hz. and, say, 4000 Hz. as the mid band speaker should well be able to reproduce the most important section of 100 to 200 Hz. as well as the higher frequencies.

You see: the possibilities are many. The question remains: What is a good option?
Whatever configuration is opted for: the quality of the final result depends on your own experience and expertise which you will gain in practice.

Do not think that you will be ready in a couple of months. An active system is not easily built. To give more insight in the matter the following topics will be considered:

1. Subwoofer and the Electronic Crossover Module
2. Amplifiers, Loudspeaker Cable and Interconnects
3. Woofer, Low-Mid Loudspeaker Unit and High Frequency Unit.


Should You Build
the X-Over?

If you already have experience in constructing tube amplifiers, tube pre-amplifiers, etc. you should have enough knowledge and have had practice to construct the Luxkit a2003 Crossover Unit.
The Luxkit can be constructed if you know about soldering, about etching circuit boards or having them etched, if you know about constructing a cabinet, and if you can find the necessary parts, typically the variable resistors (potentiometers) and the multiple selector switches.

Most people who ask for the manual do already have the Luxkt crossover. They just want to repair it or improve the performance.

If you do not have experience in building tube gear, you better do not start getting involved, or if you want to construct the X-over nevertheless, then it will take much more time as on the way you will get the necessary experience. It is not a job for a month or so. It may take at least a year or even several years.

It is more practical if you buy a unit at an auction site even if the unit is not working properly. You can overhaul it and make it work.


Working with the Luxkit asks for patience and the will to learn and to get experience in the long run. At the same time while making the construction and working with it, you will develop your perception of what sound and reproduction is all about: dynamics, frequency band, about phase (time coherence), and with it width, depth, and even height of the sound image. Much is determined by the choice of amplifiers, of loudspeaker units, of size and shape of cabinets, and by finding the best crossover points, plus that you will be striving for linearity and good dynamics and avoid coloration.

Never focus on the Luxkit and what you are going to make without having a proper stereo set with small or large loudspeakers that you will enjoy and may remind you about what music can and should sound like. That combination should give you peace of mind and make it possible to work on "your active loudspeaker systems". You most certainly will want to make a difference from the commercially available systems. But you will realize that it is not easy to surpass the quality of most commercial systems.

The original manual includes the schematic of just one section that you will have to repeat 6 times (3 sections and these for left and right channel to obtain stereo). Included are the real size images of circuit boards for low, mid and high filter sections, and of course the schematic and image of the circuit board for the Power Supply. They are included also as real size black and white images of these boards on a CDROM.

Each filter section has its own print board on which the tubes are located in their individual sockets. There are three tubes (sockets) for the low section, three tubes (sockets) for the mid section and 2 tubes (sockets) for the high section.

In addition there is the circuit board for the power supply with the values for the transformer and the capicitors. The transformer should be manufactured. You should look for a special factury where the transformer can be wound to the given parameters.

All printboards indicate the positions of the components (and their values) which are to be soldered on the print boards.

In addition there are the small print boards (circuit boards) for the individual x-over frequencies of your choice. These small boards with capacitors can be made for practically every frequency. The manual gives the values for the small film capacitors (if not available you will chose another quality). Overlapping and combining 6 dB slopes with 12 dB slopes for the various sections is possible.
In addition the physical positions of components have to be found. The distances can be determined empirically.

The interior of the Luxman Active Crossover LUXKIT a2003. As you can see that some of the small vertically placed frequency boards in this image are home-made.

Below the underside of the four printboards.


Subwoofer and Electronic X-Over


Subwoofers have existed since the early days of high fidelity. Even before World War II (according to Jean Hiraga in his excellent book 'Les hautparleurs' - Editions Fréquences. Paris, 1980) a man named Olney proposed a loudspeaker specifically designed for the reproduction of the low frequencies. Mr. Olney used a labyrinth (which was later tapered by Dr. Baily and became known as the 'transmission line' loudspeaker).

In the very first issue of High Fidelity Magazine, published in the Summer of 1951 (Vol.I, No. 1), is a description of another low frequency system: the 'air coupler', which is a long slim cabinet with a loudspeaker mounted on the back panel, radiating into it and just above the speaker unit is a wide port.

 Air coupler drawing

Click on the drawing to view a larger image.

The cabinet is 6 feet tall (long). The Air-coupler is meant to reproduce the frequencies from 20 Hz. up to around 350 Hz. The cabinet was complemented by a two way speaker or a dual concentric or a loudspeaker unit like a Goodmans Twinax or some similar construction.

Both air-coupler and full range cabinet had their individual low pass, respectively high pass filters. I have scanned the drawing from my (rare) copy of the first issue of High Fidelity magazine (summer 1951) and have put the drawing on another page. Just click on the drawing above. Maybe it could inspire you to build an air coupler and find the best crossover frequency.

Air coupler in High Fidelity magazine from 1951

The 'air coupler' (in two versions with a length of 6 feet and 8 feet respectively) as pictured in High Fidelity Magazine Vol. 1 No. 1 (1951)

In the nineteen fifties Kelton designed a system which was further developed by Hitachi. This idea became the basis for the bandpass subwoofer of today.


Already in 1970 John Marovskis from the Bronx (New York) designed his extraordinary Janis W-1 subwoofer. The story goesJanis subwoofer W-1 designed by John Marovskis that he made some 50 cabinets before he arrived at the excellent sonic performance of his system which has a flat frequency response from 30 to 100 Hz. (-3dB).

He designed an electronic crossover module with a 60 Watt amplifier to go with the handsomely finished cabinet. He called this elelctronic crossover unit with amplifier, Janis Interphase with a fixed crossover frequency of 100 Hz. Both low pass and high pass sections were filtered with 18 dB slopes.

Later he designed a smaller Janis W-2 to add to his catalog, and also another version of the Interphase. The crossover frequency was fixed at 100 Hz. This Janis Interphase, type 3A, filtered the low pass section (Janis W1 or smaller W2) and the high pass section (satellites) also with slopes of 18 dB per octave.

Another, most important feature of the Interphase modules was the continuous variable phase adjustment of the subwoofer (0-180º) in order to achieve a seamless and harmonious integration of subwoofer and satellites, independent of the subwoofer's position in the room.

Dutch importer Jan Endenburg told in an interview in 1986 that he had read about the Janis W-1 system in 1970 in an article written by Bert Whyte (published in AUDIO) and when he was in New York he called John Marovskis asking where he could hear the Janis subwoofer. Marovskis asked Jan Endenburg where he was calling from. "I am in New York", Endenburg said. John Marovskis invited him to his home and demonstrated the systems set up with Quad ESL loudspeakers.

That is how Endenburg's firm Engasound became the importer for the Netherlands already in the early 1970s. It was Jan Endenburg who explained to me the importance of phase and he pointed out to me that optimum phase coherence could also be obtained by using a 12dB low pass for the Janis W1 and a 6dB high pass filter for the satellite speaker as was achieved by the than cherished and famous Symmetry Electronic Crossover. But I found that the 18 dB slopes were much cleaner and also the Symmetry did not have this variable phase.


Once in a while you read a rather negative review about the Janis subs. But these are always written by people who do not know how to use the Janis sub woofers. They make the mistake of using high current amplifiers (NAD, Rotel, Bryston, etc.) that get their power out of the large toroidal transformer instead of large capacitors. Toroidal transformers cannot deliver enough Principle of Janis subwoofercurrent and are soon exhausted because they need to deliver current directly out of the AC mains.

In the past I have tried a Tandberg TPA 3003 power amplifier for the satellites. They did not match at all. The sound of the Janis Woofers became realistically impressive only when using amplifiers that really go with the nature of the Janis Interphase. For instance when an Amcron power amplifier was used to furnish the power to the satellites. The Janis worked also perfectly well with mono blocks from French manufacturer Elipson which were actually designed by Setton. Just to mention some historical units. If you combine the Janis Interphase with the right amplification for the satellites, the music will get its impact and slam. Not only when playing Tchaikovsky's 1812, but any record from your collection.

In recording studios in Hilversum (The Netherlands) Janis Subs were used in the nineteen seventies en eighties.

Below the frequency response curve of the Janis W-1 apparently non filtered. (From the literature provided by John Marovskis.)

Janis W-1 Subwoofer Response Curve

Another favorite bass system from the past was the Hartley sub woofer which was combined with de Decca Ribbon high frequency unit and the Quad Electrostatic Loudspeakers (ESL). Impressive as well. (See: The Decca London Ribbon Horn Tweeter.)


In the last 15 years many well known manufacturers of expensive subwoofers have forgotten about the importance of the continuous variable phase feature or just do not realize its importance and only adorn their modules with a phase-switch with the two positions of 0º and 180º.
A few years ago I had a severe discussion with the manager of a well known English manufacturer of big subwoofers. Even their most expensive subwoofers can only be integrated with great difficulty into a given high fidelity stereo-set of speakers. At times it is even impossible to find a seamless transition of these subwoofer with the satellites.

Most people think that a boom is what it is all about. This British manufacturer certainly has not done his homework. I have witnessed the horrible Janis Interphase crossover with amplifier and variable phasesound produced by expensive Avalon loudspeakers (and other high-end loudspeaker systems) in various living rooms because the electronics of the expensive subwoofer(s) did not have the continuous variable phase control. Disconnecting the subwoofer brought back the harmonious sound and the slam of the Avalons. Even today, many years after my conversation, these subwoofers still do not have the variable phase control but just a phase switch.

Yet there are electronic crossover modules on the market for filtering the subwoofer and correcting the phase of the signal via the continuous variable control from 0º to 180º. Best known are those of Detonation and Mivoc (search with Google). Thommessen, also from Germany, used to make the Subway Active Crossover. But they have stopped the manufacture.

Dutch readers click here to download the Dutch PDF about John Marowskis Janis Subwoofer.


Use of the Electronic X-Over



If you decide not to add a subwoofer to satellites but you want to assemble a full blooded 3-way system you will have to make your choices with care. There are several x-over units on the market. There are units made especially for studios and for PA purposes.
Try to find a high quality module for real high fidelity application. Three-way modules do not have the adjustable phase control for the subwoofer. In that case you have to find/calculate the best position of the units in relation to each other as you would do when designing a 3-way passive system with stepped/recessed baffles. You may even look for a vintage unit on eBay or Audiogon.

I myself used a vintage crossover unit: the LuxKit a2003 from Luxman which was a present from one of my mentors, Mr. W. de Haan, well-known icon of the hifi-history of the Netherlands, not only because he imported for a long time Tandberg and later the Mission and Cyrus products, and Cabasse loudspeakersystems, but also because he was above all a first rate technician.

The Luxman Crossover (LuxKit a2003) was available in the nineteen seventies. It has 8 tubes of the type 12AU7 (ECC82). This unit has slopes of 6 and 12 dB.

Different turnover (crossover) frequencies can be chosen. It suffice to insert a small circuit board with the appropriate capacitors with the values indicated in the manual.

There are 5 positions to the left and 5 positions to the right of the zero point (0 = the chosen frequency). Each step to the left represents a difference of 1/8 of an octave.

An example. Say you choose for the low pass for the woofer or subwoofer (the section at far left on the module) a cross-over frequency of 200 Hz. This means that 1 octave is 100 Hz. That means that every step down if 1/8th of 100 Hz. is 12.5 Hz. At position -1 the frequency is 200 minus 12.5 Hz make 187.5 Hz. at position minus 2 the frequency is 175 Hz. And so on.

Turning the knob to position +1 gives a different frequency. The octave of 200 Hz. is 400 Hz. That octave has to be devided in 8 parts. Only 5 are usable. Each section adds 25 Hz to the frequency chosen for 0.

Note: Initially the Marchand electronic crossovers were designed with 24 dB or even steeper slopes and they use 12AX7 tubes. The steep slopes are not to everybody's taste. However, now the website says that 6, 12, and 18 dB slopes can also be used by inserting appropriate circuit boards.

So the Lux Kit works with circuit boards, but the advantage of the Lux Kit is that switching from one crossover frequency to another, and from 6 dB to 12 dB, is possible without changing a circuit board. This means that the effect of a different crossover frequency can be heard and measured instantly and helps to determine the best crossover points for the loudspeaker units and the final values of capacitors on the circuit boards. It is even possible to overlap crossover points. This versatility is a great asset and is tempting. However building this crossover from scratch is no minor task. Only experienced hobbyists can do that.


Various configurations are possible. As said the simplest is in fact a 2-way active system consisting of a subwoofer and two satellite speakers. Satellites can be full range speakers or 2-way, 3-way and even more complicated designs.

basic active system 2-way

More experience and a good ear is necessary to establish an active 3-way system where all three transducers have to be adjusted: frequency, power level, and phase.

Principle of 3-way active system including a subwoofer

One visitor of my pages wondered if a combination of a 3-way active system could work together with a separate 2-way active crossover and subwoofer. That is of course possible. But the success depends on the quality of the electronic crossover units, and again the knowledge and perception of the hobbyist. The more adjustments will have to be made, the more difficult it is to optimize such a system. No active system will be assembled overnight, not even the simple 2-way configuration, let alone such a complicated system using two active crossover units.

4-ways active system


As with all cabinets and chassis, also the Luxman shows some microphony when tapping on the metal housing.
The level depends on the way the chassis is built (weight of the material, strength and sturdiness of the construction), on how tight the tubes fit in the sockets, and were the crossover is located and how it is supported.
I have heard tube amplifiers constructed by hobbyists and manufacturers alike which were too heavy and placed the mechanical resonance (resonant frequency of the unit) too low and made the sound muddy to such an extent that they began to sound like solid state amplifiers.

The schematics of the LuxKit do not prescribe the way the cabinet should be built. So it is up to the hobbyist to decide what housing he will build and how he will go about it.
I myself am against too heavy housings for tube amplifiers. I myself am against the silicone tube rings or whatever rings are often used to suppress resonances in tubes. They not only dampen the resonance, but do make the sound un-tube-like and less harmonious. But I must emphasize that these opinions/findings are strictly my personal preferences.

Many tube amplifiers of today sound neat but are too heavy in the lower region and specifically in the bottom.
One of the features of good constructions is the airiness of the sound, specifically of tube amplifiers and valve crossovers. It is wise to seek the best construction and weight of the housing of any amplifier and especially of the Luxman Electronic Crossover filter (Lux Kit).
The total weight of the original Luxkit a2003 is about 6 kg. which is some 13 pounds. Weighing the parts (transformer, circuit boards, components, capacitors, potentiometers, etc.) and subtracting this from the 6 kg. can give some indication about the weight to be chosen for the cabinet of the X-over.


Tube tester


Since the Luxkit has 8 tubes, and brands and quality of tubes are not the same, it is important to know what tubes to get and if the quality is right. Testing tubes is necessary to find out if your Chinese and Russian tubes deliver what they promise and if old Philips, RCA, Telefunken, etc. specimen are still working well.

Above the EMC 213 tube tester. It is a simple to handle, very practical and effective unit to measure the quality of tubes. And you will be surprised how often newly manufactured tubes perform badly. Throw the bad tubes away so you will not accidently use a less performing one. But the quality measurement does not say all. You will have to experimently find the best sound by rearranging the tubes in the crossover. Always of course after listening tests and evaluating each section of the audio band. Before you rearrange make sure the tube crossover is switched off and cooled down for several minutes.


So, you should decide which configuration you are going to make: just add a subwoofer to existing satellites or design a real active loudspeaker system.

Let us start with the subwoofer. There are many subwoofers on the market. When surfing on the world wide web you will encounter many different propositions and plans to build a subwoofer yourself. Read these plans carefully and find out about efficiency, power handling, frequency band, linearity and size of the cabinet, and then decide what way to go. If you are into classical music and into jazz and the refined reproduction of analog material, I advise not to use any electronic bass boost. You just should strive to select a bass unit and construct the appropriate cabinet so that low frequencies will be reproduced without any artificial compensation. That is only if you have rather sensitive ears. If so, you may have heard that boosting and correcting the low frequencies (as in motional feedback systems for example) yields to a completely different (unnatural) low frequency reproduction.

When planning to build a subwoofer you should look for 12" or even 15" speaker units that have a low resonance frequency and the appropriate Qts for functioning in a closed compartment. Try to find units that do not have a foam suspension which will deteriorate in time. The unit should have a rubber roll surround which can stand the test of time. Yes, they are expensive. I was lucky to buy a pair of second hand 1976 Goodmans 3-way systems which had 12" woofers in a volume of 53 liters. The suspension of the woofers is rather stiff, but as I could hear the speakers before I bought them, the resonance frequency seemed relatively low. The units weigh some 10 lb. each. Since the basket is of steel most of the weight goes into the magnet.


Qts = 0.30
Vas = 199 l
Fs = 31 Hz.
Re = 7.19
Le = 3.93
Qms = 1.83
Qes = 0.36
Spl = 90 dB
Bl = 13
Diameter = 26.4 cm
Sd = 0.550 cm2
Pe = 75 Watt
Z = 8 Ohm

There are speaker units which are especially suitable to be used in band pass enclosures I wanted to see if these Goodmans woofers could do the job as well. A friend measured both units and came up with interesting parameters and to my surprise the resonance frequency was 31 Hz. and the QTS had a nice value of 0.30. I would have welcomed a slightly higher QTS and a lower Fs though. In order to calculate a bandpass system I used Juha Hartikainen's valuable WinISD beta-program which you can download from the web. It is a very effective program and rather easy to be used. It can calculate closed boxes and vented enclosures.

Hartikainen's program makes it possible to determine the desired frequency response by altering and adjusting the volumes of the front and rear chambers just by moving the mouse over the drawing. Since a computer program is infatigable one can go on calculating and by trial and error come up with many possible solutions. Also by changing the frequency of the system. The aim is threefold.

First: it is important to change volumes in order to come up with a low -3dB frequency which is situated around 30 Hz. or possibly lower.

Second: it is important to find an upper -3dB frequency situated around 100 Hz. to go with it.

Third: in all cases the curve should be as linear as possible and at the same time the impulse should be the best you can get. If the configuration you calculate shows a bad impulse and a frequency characteristic with the shape of a saddle, than calculate again and again.


Most subwoofer designers choose a frequency curve which is linear within 3 dB. This is all right if you have a large listening room. In small rooms the long wavelengths of low frequencies cannot establish themselves. So what is the use of reproducing 20 or 30 Hz. when you have to sit in the kitchen or in the hall in order to hear them properly? There is a second problem.

When adding a subwoofer to a satellite system there is a possible side effect and that is that the sound in the vicinity of the crossover frequency gets a slender character, especially in cases a satellite with a relatively small bass-midrange unit is used. To avoid this one has to doctor around and find the best integration without changing the strength around the crossover frequency. In Hartikainen's program you can alter the system's resonance frequency and come up with a different port length and characteristic. This all helps to find out what the best rear and front volumes are. The final adjustment of the port length should be done when setting up the system in practice. I calculated a volume of 29 liter for the front and 58 liters for the rear volume. The volumes were actually 30 and 60 liter because the bracing and the woofer's baffle take up some of the volume,

I decided to position the units on oblique baffles. This would help minimizing standing waves. It would also position the coils somewhat farther towards the front of the box for better alignment with the other units. The Goodmans woofers only have four holes for fixing. Four clamps were added for rigid fixing of the units. The box is well braced using metal strips glued and screwed on wooden sticks to minimize panel resonances. The unit is braced as well by means of a beam supporting the magnet. The front of the box is detachable so the unit can be screwed on the baffle and damping material can be adjusted. Each subwoofer has 2 vents which can be adjusted as well in order to achieve a seamless integration.


A Suggestion for A Subwoofer

Goodmans woofer
I was lucky to buy a pair of second hand 1976 Goodmans 3-way systems which had 12" woofers in a volume of 53 liters. The suspension of the woofers is rather stiff but the resonance frequency is relatively low from what I could hear. The unit weighs some 10 lb. Since the basket is of steel most of the weight is of the magnet.
A friend measured the parameters and to my surprise the resonance frequency was 30 Hz. The speakers could be used in bandpass subwoofers. For calculation I used Juha Hartikainen's valuable WinISD beta-program which you can download.
Juha Hartikainen's program makes it possible to determine the desired frequency response by altering and adjusting the volumes of the front and rear chambers. I choose a volume of 29 liter for the front and 58 liters for the rear volume.

The volumes were actually 30 and 60 liter but because bracings and the woofer's baffle, the volumes were calculated slightly less: 29 and 58 liter respectively.

I decided to position the units on oblique baffles. This would help in minimizing standing waves. It would also position the coils farther towards the front of the box for better alignment with the other units. The drawing shows the small difference according to the inclination.
These Goodmans units have only four holes for fixing. Four clamps were used for added rigidity.
The box is well braced using metal strips glued on wooden beams to minimize panel resonances. The unit is braced as well.

The front of the box is detachable so the unit can be fixed to the baffle and damping material can be readjusted if necessary.

The subwoofer has 2 vents and is mounted in an iron frame together with the units for mid and high frequencies.
Qts = 0.30
Vas = 199 l
Fs = 31 Hz.
Re = 7.19
Le = 3.93
Qms = 1.83
Qes = 0.36
Spl = 90 dB
Bl = 13
Diameter = 26.4 cm
Sd = 0.550 cm2
Pe = 75 Watt
Z = 8 Ohm

graph calculated without before filtered


Above is is the edited graph of the calculation done by Win ISD Beta for the Goodmans woofer. For the volume of the closed chamber I choose 59 liter. The corresponding volume for the front chamber is 29 liter. This is the graph for the system resonance of 65 Hz.

The program makes it possible to change the volumes and view the corresponding graphs and thus determine a preferred system. The actual response at -3 dB in this case is 35 to 125 Hz. The equalization lowers the efficiency (sound pressure) of the chassis which was 90 dB originally and now is 86 dB.

Naturally different volumes can be calculated with different system frequencies and the corresponding graphs make the changes visible.

Should these graphs (response curves) always show a straight line? That is for you to determine. And always in relation to the mid range and high frequency sections (the satellites). At left the curve I choose several years ago. The curve is far from a straight line. In most listening rooms the lower frequencies are being amplified. The level depends on size, volume, measurements, and shape of the room.

calculated alternative for useful filtering

The above graph shows the response of the same sub woofer (Goodmans in cabinet) but with a system resonance of 76 Hz. and vents of 5.5 cm in length and of 60 mm in diameter. Now 100 Hz. is just under 90 dB, the original efficiency level of the Goodmans loudspeaker unit. And in this case 50 Hz. and 125 Hz. are at -3 dB, whereas the sound pressure at 30 Hz. is -11 dB.

graph before and after filtering

This curve is of the sub woofer after filtering with the Luxman Electronic Valve Crossover (Luxkit a2003) or any other crossover unit, x-over frequency 100 Hz. The curve has changed. The result is that 40 Hz. is -3 dB. and 30 Hz. is - 8 dB. Increasing the length of the port(s) will straighten the curve from 30 to 50 Hz. and dynamics in that region will be improved.

Of course it is up to you to define the best system frequency in relation to the rest of the active loudspeaker system. These are just examples to show you what you can do by altering the frequency of the system in relation to the filter.

Amplifiers, Units and Power



The choice of amplifiers depends on several criteria.
1. Efficiency of the loudspeaker units.
2. Size of the room.
3. Crossover filter.
4. Available budget.



If you use loudspeaker units with an efficiency of 85 to 88 dB and they can handle enough power, it is advised to use power amplifiers with a high power capability. That means not just Watts, but how these are generated and eventually stored is important. Only amplifiers with large capacitors for storage of lots of energy will do the job.

Amplifiers with large toroidal transformers and just two capacitors with low values do not pump enough power in 85 to 88 dB units. Yes, they generate power. But the power is mainly drawn from the big transformer, i.c. from the net. The effect of these amplifiers (most of which are in the non-clipping category) can not only be measured but can be heard as well. In heavy and loud, complex passages in classical symphonic music the attack of drums, brass and cymbals should be instantaneous and direct. In other words: if the conductor's baton says "now!", it means "now!" for all musicians involved.

When using low efficiency speakers with amplifiers which draw their energy from the transformer, such complex passages do not sound instantaneously but there is a small time delay which in commercial loudspeakers get larger and is really noticeable when complex passive, energy-absorbing filter sections are being used. Many times the sound of the kettle drum which needs more energy comes at a different time than the sound of a cymbal. So it is advised to use these amplifiers with big toroidal transformers in conjunction with relatively high efficiency speakers of 90 to 93 dB.

If you use units with an efficiency of 95 dB and higher, you also should choose your amplifiers with care. Especially high efficiency units make it possible to use small power transistor amplifiers or valve (tube) amplifiers which are fast. Power amplifiers with many hundreds of Watts and large capacitors for energy storage, generally mean overkill for these units.

The most important criterion for choosing amplifiers for an active loudspeaker system is that they should be of the same electronic design. A manufacturer like Cabasse for instance, who have a long history of designing active loudspeaker systems, may choose power amplifiers with different output power for the low-, mid- and high frequencies. But what all these amplifiers have in common is the same design philosophy. This results in phase coherence of the amplifier throughout the entire system.

Physical phase is important

So the best option is to choose the same type of amplifier for all loudspeaker units. Most of the time you will choose amplifiers from the same manufacturer and even with the same reference number.
When choosing amplifiers in relation to the efficiency of the units you can find the optimum balance of speed and power capability. I had very good results with three Hitachi HMA 6500 amplifiers. They are not too powerful and they mix very well with the Luxkit a2003.


It goes without saying that small listening rooms ask for small speakers with an efficiency around 88 dB and relatively small amplifiers. The sound generated by high efficiency units and big amplifiers does not fit in a small room. In very large rooms big units and amplifiers with a lot of energy are a necessity. In average rooms units with an efficiency of 90 to 93 dB will work wonders. These numbers are just an indication. Experience will tell you what the best system is for a given listening room.


Even if you have selected loudspeaker units with linear frequency curves, it is of the Utmost importance what X-Over unit you chose. If you just add subwoofers to existing loudspeaker systems, than the earlier mentioned Thommessen subwoofer crossover or the Detonation will do the job. But if you go for a 3-way system, than it is important to chose a quality unit which enables you to vary the crossover frequencies and the slopes. There are units with slopes of 6 and 12 dB like the Luxkit a2003 tube cross-over.

There are others which also sport 18 dB and sometimes even 24 dB (Linkwitz) filtering. The advantage of steep 24 dB slopes is that the unit just gets that part of the audio band which it is meant for and causes the least distortion. The sound becomes clean. But in my view too clean. I myself have always found that 24 dB is too steep a slope and makes it more difficult to build a harmonious sound pattern without leaving details out. Slopes of 18 and 12 dB, or 12 and 6 dB are better.

Characteristic with 12 dB for low and high section and 6 dB slopes for the mid band

Using 6 dB and/or 12 dB slopes as these are provided by the Luxkit make attaining phase coherence much easier. Using just 24 dB slopes means that you have too choose the units with as much care but an electronic filter with only 24 dB slopes is not versatile at all and soon takes away the hobbyist's enthousiasm. Slopes of 6 dB/oct. mean that the phase is correct (coherent). It is still is necessary to find the correct vertical and horizontal distances between the cabinets and the best acoustic load for the mid range units. The X-over should be well designed and if possible have discrete circuitry or just have not too many operational amplifiers (opamps) and integrated circuits (ICs).

If you have a X-over with settings of 6 and 12 dB (like the Lux kit) you can make a perfect setup with very good phase, filtering the subwoofer (or woofer) with 12 dB, the midband with 6 dB high pass and low pass slopes, and the tweeter or midhigh/high section again with 12 dB.



An important ingredient is the available budget. If the budget is low, the best thing is to start with relatively cheap power amplifiers with the same reference number of a given brand. Maybe you will find them in the second hand marketplace. If after some time when you become more experienced and more knowledgeable and your active system reaches a high level of quality, than more expensive power amplifiers may be chosen. Remember that cheap amplifiers will hardly be able to drive a subwoofer in the lowest register due to little capacitance in the power supply. Also I do not advise the use of high current amplifiers (large toroidal transformer and just two low value capacitors). When high dynamic levels are asked for the transients are smeared out and in fact can not be called transients!


Important is also the choice of cables: interconnects and loudspeaker cables. How many times did you read in threads or mailing lists that someone combined cables of different make and bi-wired a loudspeaker system? And how many times that person described how the sound benefited from this? If different cables and/or cables from different brands can be combined and give a correct sound, it is a chance of one in a hundred. In an active loudspeaker system the same type of loudspeaker cable has to be used for all sections. There may be some difference in internal wiring, connecting the binding posts with the unit inside the cabinets, but even then one has to choose wires which are from the same manufacturer. Using Cardas wire - as a suggestion - of different diameter for internal wiring of low, mid and high frequency unit, will lay the basis for smooth transition from one section to another.

The various amplifiers should also be connected with the same type of interconnect. Optimum slopes and crossover frequencies can only can be chosen if there is no influence brought about by different cables.

© R.A.Bruil, October 10 2002


How to order the Owner's and Service Manual.


Owner's and Service Manual of Luxkit a2003 Electronic Valve Crossover in English on 24 pages, schematic diagram, edited scans of the circuit boards for the power supply, the low, mid and high sections to be used for etching the boards, plus suggestions and considerations.

The price: $ 29. Plus $17 for shipment outside the European Community and plus $13 for EU countries.

Due to the time-consuming preparation of a shipment as demanded by Dutch postal service Postnl, I do not send documents to the United States.

The Luxkit a2003 Electronic Valve Crossover with 8 x 12AU7 (ECC82) is a fine unit and can certainly be built by the amateur with knowledge of electronics and experience in building valve components. The Luxkit a2003 is a 3-way crossover with many possibilities. The low pass and high pass crossover frequencies are variable and can be adjusted in relation to the qualities and behavior of each driver. This means that sections can be adjusted in such a way that they overlap each other or that a gap between two sections can be made in case of an irregular frequency curve of one or more units, combinations of 12 and 6 dB slopes can be made with minimal phase shift.

This Owner's and Service Manual contains all schematic diagrams and values of resisters and capacitors.
It includes images of the 4 circuit boards (3 for the individual channels and 1 for the power supply).

As an extra I have edited the images so they can be used for etching the circuit boards in case you would like to overhaul your unit or build a unit from scratch. I have also added extra information about the unit and how to make the prints for the selection of the crossover frequencies. But you also can etch prints yourself or have them etched.
The manual also explains the functioning of the x-over:

* selecting frequencies,

* polarity of amplifiers,

* polarity of speaker units in relation to the chosen slopes of 6 and 12 dB

* relation to phase.

>>> Click to order Lux kit Schematics




Home | Audio & Music Bulletin | 25 Years CD - Digital Audio Compact Disc
LP List | Record Cleaning Page | Ortofon/Decca/Tannoy Cartridges |
The Remington Site | Record Stabilizer Ring  | Record Weight | Links 

Audio&Music Bulletin - Rudolf A. Bruil, Editor - Copyright 1998-2013 by Rudolf A. Bruil and co-authors 

Top of Page