ScanSpeak
Ellipticor-28
Copyright 2025 © Troels Gravesen
Go to on this page:
DRIVERS CROSSOVER CABINET WORKSHOP PICS MEASUREMENTS
SPEAKER-KIT The Sound CROSSOVER LAYOUT HYPEX program
April-2025:
This speaker construction was started more than 18 months ago and was meant
for myself only. It is built for our 24 sqm living room and I wasn't
planning a kit, but the option of a simpler mid-tweeter cabinet might
make it possible for most DIY'er to follow. The midrange-tweeter cabinet
I made is crazy and I made three versions before I was had something I
can live with due to non-horizontal bends
of the side panels. This makes crazy odd angles all over the place. Some CAD nerd might
possibly make a 3D drawing of it.
The speaker below is in no way a beginner's project and I urge you to go
through the whole page and carefully
consider if this is something you should engage in. If you have
questions to cabinet construction, implementation of crossover or doing
the Hypex programming - please do not engage.
You will preferably need a preamplifier with double balanced output - or you will
have to find out yourself how to balance the bass driver to mid-tweeter
section from a RCA line output - or taking the signal from speaker
terminals. If you cannot do this - do not engage. And there will no fully
passive version either, so forget about that. You need two Hypex FA501s
to make this speaker happen.
The reason for saying so, is that over the last
few years we have reached an audience that in a few instances shouldn't
have engaged
in speaker building. Strangely enough these people have chosen the most
complicated constructions, like The-Loudspeaker-1-2 or Ellipticor-A50
without ever having had a solder iron at hand. One builder suddenly
found out that resistors had different values!! Obviously we're happy
for the sale - but saddened that people were not able to finish their
construction due to lack of basic knowledge in cabinet construction, nor
following the instructions on how to make a crossover that doesn't look
like a bird's nest, with no possibility of checking layout and wiring.
We do urge all customers to consider their skills in woodwork and fairly
simple electronics before engaging in speaker building. It may be more
complicated than you think and our DIY kits are not paint by numbers!
Some of you have to learn stuff to make it right. Please go through
all of the tips page and see if there is anything alien to you. If
so, please reconsider.
I also start with this warning as my request to follow six links (see
below) before starting your project is highly neglected. People just DO NOT
read. Sorry to say so, but countless mails on e.g. differences between
crossover schematics and layout tells the story. There may be a lot to
learn for newbies.
I'm sorry to start out all grumpy as this has been one of the most
joyful constructions I've ever made. I'm so happy to have them in our
living room. Please read below - and hopefully enjoy!
For many years I used the Jazz at the Pawnshop as a reference recording. Then new things came along and I haven't actually played it for years. Many things have changed in my setup, better turntables, better cartridges, better amplifiers, etc., etc. Pulling out the recording again after many years made me realise that what I thought was perfect back then, really had some shortcomings. I recall the clarinet always has some edginess to it and the soundstage some lack of depth. Not any more. With the Ellipticor-28 in my workshop and living room - and even a modded Rega/Candenza Black/EAR MC4/EAR-868PL/Parasound A21+ there were lots of new things to discover. Smooth, smooth clarinet and well, like sitting 2nd row in the Pawnshop with the stage full width. I hadn't seen this coming.
Speaker Diary
Oct. 2023:
Sensitive, dynamic, powerful, deep bass and compact will be the five primary
objectives of this speaker. And of coarse all the usual qualities of
transparent midrange and smooth treble. Can we really have it all?
Indeed so. For bass we will have 500 watts driving a
powerful 11" bass driver, that, given its rigid straight-sided sandwich
cone, 3" titanium voice coil and ideal TS data, will deliver dynamic, thunderous and
deep bass from a relatively small cabinet. And the Hypex will lift it to
the possible sensitivity of midrange and tweeter.
Tweeters in play:
Over the years we may become particularly fond of specific drivers and
wanting them combined in a single construction if possible. Not that
this guarantees success, but we can't know until we have tried. So I
have it with the Illuminator D3004/660000, the Ellipticor D2404/552000
and Revelator D2904/710002 tweeters.
The -02 version of the Revelator is the silver faceplate version of the
D2904/710003, I ordered a pair of both to see if they are truly
identical and what looks the best. They are absolute identical I can
tell.
The midbass:
The Ellipticor 18WE/8542T00 midbass driver is to my mind one of the few
very, very best midbass drivers around, and there was no doubt in my
mind this should be the corner stone of this construction, hence
Ellipticor-28 despite two Revelator companions.
Bass:
In my
MUN-17-3W construction I used the magnificent sandwich coned
28W/4878T00 bass driver - and it has a sibling, the 28W/4878T01.
The -T01 has twice the efficiency of the -T00 due to reduced Mms,
higher BL, etc. Yet, it does F3 = 32 Hz in a 50 litre cabinet. Quite
impressive. And Xmax of +/- 7 mm is enough for me. I don't play that
loud.
The MUN-17 never caught on as a DIY kit, probably a combination of cost
and the use of active bass section. Some builders don't want
semi-active solutions despite all the benefits of room adaption, so much
needed for most.
The MUN-17-3WC was not a cheap project, I admit. And this is not going
to be cheaper using the 18WE midbass driver, although the use of
6600/7100 tweeter in replacement of the Be dome, 7140, may make it about
even. The D2404 is as expensive as the Be dome, but I'm not building
this speaker to be a DIY kit, rather because I just like having it - and
because size wise it suits my living room just perfect. Should someone
fancy to try it out we have decided to set up the kit, but be prepared,
the midrange cabinet is a nightmare in terms of construction, although
an easier way is suggested and will work as well as the one show
here.
The
Hypex FA501 plate-amp will run the bass section from your preamp
fitted with XLR output (analogue XLR). As I have doubled balanced output on both my EAR
pre-amps, I run them from the line stage and prefer this option - for no
particular reason. But if you can't follow this you have to find a way
yourself to match the bass to the mid-tweeter section.
Now, there are several options setting up a classic 3-way system like this and I'll go
through them here. First of all, and as mentioned many times before, the
efficiency of the bass driver is normally what determines the overall possible
sensitivity if a speaker system. We can, and usually do, attenuate midrange and
tweeter, but we cannot make a bass driver produce higher output than
what is set by the TS data of the driver and the baffle on which it
sits. There is an exception from
this in case you have two amps with variable gain, where you can boost
the bass driver, say 3-6 dB, compared to the amp driving the mid-tweeter
section, but this will not be shown here.
In a 3-way system with the 18WE/8542T00 driver as midrange we can
achieve a 91-92 dB system, because the bass and tweeter will lift the
overall response in the midrange, at least from shallow-sloped 2nd
order crossovers, because here we can set the Hypex module to add the
extra decibels to the bass driver. So, the bass driver is no longer the
limiting factor for system sensitivity; here it's actually the 18WE
driver that sets the bar. And well, the D3004 and D2404 tweeters can
also only match some 92 dB and still we have to add very modest
attenuation to the tweeter. The D2904 tweeter can do up to 95 dB, but
not needed here, but it may save us a few expensive uF on tweeter crossover based on
practical experiments. The D3004, D2404 and D2904 tweeters work from a little above 2 kHz
and they do well. Beryllium? Fine, but matter of taste. No, the
D2404/D2904/D3004 all work so well I won't even try. Forget it, it won't
happen. And BTW, Be is on its way out and will be replaced by TPCD,
carbon domes.
So, basically we have no damping components in line with the
bass and midrange drivers and this ensures maximum dynamics, just
what I wanted.
From a fully passive version we have to first look at the 28W driver and
we can manage a system sensitivity of around 89 dB and mid and tweeter
must be attenuated accordingly. This was tried and worked well, but
did not manage the dynamics of the high-efficiency Hypex version.
A fully passive version driven from a single amplifier would also mean a
minimum impedance in the area of 2.5 Ohms. The same driven by bi-amping
would do better, but still, what's the point if we can have some 3 dB
extra sensitivity?
As can be seen, many options. And I'm not trying to please the
"single-amp, non-bi-amped" people here, rather make the most of our
precious drivers.
The point of crossover between bass and midrange is another option to
consider. Basically we can have anything between 200 and ~600 Hz,
although 200 Hz would mean lower system sensitivity due to baffle step
loss of the midrange, so it has to be a bit higher. The discerned reader
may ask why I didn't use the 4 Ohm version of the 18WE driver to
overcome this, but having the bass driver running not only the bass but
also lower midrange adds to the dynamics of the system, because a lot of
energy is located in this area.
Another reason for choosing the 8 Ohm 18WE is that I so much love the
Ellipticor-2F8 and this speaker really didn't work so
well - for some reason - with the 4 Ohm 18WE, so - we get our preferences based on history
and not always on rationales. Also we would need a lot more uF for the
high-pass section. Costs - after all, I am concerned as all of
this is out of my own pocket, as are most speakers these days.
February 2024:
There is no deadline to this speaker. It has been playing in my workshop
now for months with almost no changes to the crossover - and I'm not
getting bored, not at all. It fulfills all my objectives, deep, deep
bass, high efficiency, exceptional transparent, modest size (after all).
I feel I could settle with this speaker and never build another one. I
know I won't, because there is also a business to maintain and our customers expect
new speakers at a regular rate, although a feel I have a portfolio that
may stay for years - or as long as speaker drivers are available. Some
speakers I made some 10 years ago I wouldn't change a bit if I was to
build them again, like the
Illuminator-4. I liked this speaker a lot and it would probably
rival this speaker, albeit not on bass dynamics and efficiency, but for
sure a speaker I could live with for years to come.
One reason people keep building loudspeakers may be boredom. We get used
to a certain sound and the question inevitable arises, "what if?".
Larger speakers always do stuff smaller speakers don't and we tend to
increase size to improve efficiency or get deeper bass, etc. I
deeply respect people who stick to a given system and maybe make small
improvements over the years, maybe to the system, maybe to the acoustics
of their room. All worth while. In the end it's a matter of enjoying the
music and stick to what we've got. All speakers are compromises.
I have just ordered all components for the final crossover, so, after 4
months I feel certain how and what it should be. As said, this is all out of my
own pocket as I only ask for specific
new drivers once from my sponsors. After this, I buy. The drivers + Hypex'
here are around ~4800 EUR. The tweeter copper caps cost a fortune and
replacing them with Alumen-Z saves you around 500 EUR when we use the
D2904 tweeter. Ask Jantzen Audio for Amber-Z if you want this.
Basics:
3-driver speaker. Semi-active.
Dimensions: 34 x 44 x 112 cm, WxDxH.
System sensitivity: 92 dB/2.8V/1 meter.
Impedance: 4-8 Ohms.
Useful links (Please
follow and read info from all links before e-mailing!):
http://www.troelsgravesen.dk/LCR-RC.htm
http://www.troelsgravesen.dk/tips.htm
http://www.troelsgravesen.dk/tips.htm#CONSTRUCTION_OF_CROSSOVERS
http://www.troelsgravesen.dk/crossovers.htm
http://www.troelsgravesen.dk/Inverted-Polarity.htm
http://www.troelsgravesen.dk/choices.htm
From countless mails I know you DO NOT read the
http://www.troelsgravesen.dk/LCR-RC.htm
article on the layout of RC/LR/LCR parallel circuits in crossover
layout. Please do, it will save me a lot of time, and it may also save
you some time not having to explain why you think the layout is wrong.
DRIVERS used
for testing.
BACK TO INDEX
Ellipticor D2404/552000
Revelator D2904/710002/-03
Illuminator D3004/660000
18WE/8542T00
28W/4878T01
Download specs here: D2404/552000 D2904/710002 D3004/660000 18WE/8542T00
Crossover display a standard LR2 topology where R3+C4+L4 flattens in
impedance peak of the 18WE driver, which takes quite a few uF being just
a single peak at ~67 Hz from the aperiodic tuning of the cabinet.
Contrary to previous designs with the 18WE driver, here we have the
driver working in a relatively narrow range as both the bass and tweeter
can manage well above and below their points of crossover. This relieves
the 18WE of some energy and pays off in midrange clarity. Initially I
used Alumen-Z for the tweeter, but trying out Amber-Z proved a small
improvement, thus Amber-Z it is despite a horrific price tag.
The tweeter topology is the same for D3004 and D2404 except for R1. For
D2904, R1 and C1 are different.
Click image to view large. Save and open in MSPaint and 1 pixel = ~1 mm.
All dimensions can be +/-2 mm, internal panel dimensions MUST be
adjusted to your conditions of panel thickness, so my measurements only
as a guidance.
Bass cabinets were made from 20 mm BB and all internal panels were clad
with 12 mm HDF to add stiffness, weight, reduce panel vibration - and to
reduce volume to around 45 litres. Damping materials will add to
virtual volume and we end up with something in the area of 50 litres.
The port's straight tube is 250 mm and with the flanges this make a port
tuning of 33 Hz.
The front panel is retracted 10 mm to make room for front grill. Grill
frame was made from 12 mm plywood and kept in place by 6 pcs 8x8 mm
neodym magnets, 6 in frame and 6 in front panel.
Above some guidelines for the impossible mid-tweeter cabinet. See
further info below. The above was not the first iteration of the MT
cabinet.
And please do
NOT ask for further info. If this is not enough, build the easy straight
forward stumped pyramid, but keep front panel dimensions, tilt and
drivers' placement. Something like this:
Here you have the sides bent horizontally, believe me, MUCH much easier.
Cabinet volume is around 18 litre, not critical due to the
aperiodic tuning.
The bass cabinet
Side panels made. 20 mm BB with maple
fillets.
All parts ready for trimming internal bracing
and rear panels.
Bracings ready for making ventilation.
Rear with cavity for crossover.
All holes in braces done - and cutout for 28W
driver.
Before adding any glue, I want to make sure
all corners will be in place.
Ready for gluing, almost -
Routing for the Hypex module, holes for
terminals and hole for port.
Test assembly of one bass module.
Oak base support ready and frame for front
grill in preparation.
Time for internal cladding with 12 mm HDF:
Left: Seen from bottom. Right: Seen from top.
Use
vinyl glue.
The cabs are going to have a front grill,
kept in place by 12 pcs 8x8 mm neodym magnets.
All internal panels added felt damping.
Acoustilux added, fasten with 10 mm staples.
At bottom add 4 layers, fold.
Due to the 12 mm cladding I had to remove
material for the binding post to reach through the panel.
Base support made from 25 mm solid oak. Adjustable
feet.
First time setup as subwoofers for my
Illuminator-71s. 2nd order low-pass @ 80 Hz.
Not bad at all. Adding my 1st order high-pass filter to the EAR-861 and
reversing polarity of subwoofers, this could play insanely loud.
Yes, front grills to the bass modules held in
place by 2 x 6 pcs 8x8 mm neodym magnets.
The neodym magnets posed an embarrassing problem as I couldn't remove
the grills again because of rounded edges....
As can be seen below I had to add a "handle" to the grill in order to
have a hold.
The mid-tweeter section:
Building a mockup of the mid-tweeter took me a week and I'm not really sure I can replicate this in Baltic birch with maple fillets... These cabs are extremely difficult to make as all angles are totally odd. But I did learn a lot of things, so maybe. The basic frequency responses were much better than anticipated and from simulation and practical tests didn't make me want to make any changes to the basic design. This works and works really, really well.
Except for the bottom plate, every single
piece of material here is cut in weird angles.
I amp not going to make a suitable drawing of this for you to follow.
You have to do with the sketches above - or make a straight truncated
pyramid making the same internal volume. And don't ask questions,
please. If the drawings above is not enough, don't build it.
The port is Ø50 x 100 mm and stuffed with a 10 x 20 cm roll of acoustilux.
Aperiodic tuning.
The MT cab is damped with felt and acoustilux.
First time setup in workshop. The Hypex
program took some time to make a proper transition to the midrange and
don't worry about latency of the Hypex. Reversing polarity (same
polarity of bass and mid) produced a HUGE dip in response at point of
crossover, so they integrate well.
This bass driver is just phenomenal!
Phenomenal attack, phenomenal low-end extension.
This driver gives full credit to the organ on Bach's Toccata from my
very old
Acoustic Research test album:
The deep organ notes shakes my workshop like I haven't experienced since
The Loudspeaker-1 and -2.
One guest thought it was serious feedback to the turntable - it wasn't.
All this with no equalisation to the bass, Hypex module set to linear
response.
The sound?
OMG!
Speaker Diary
Dec. 2023:
I think it has been 7 weeks since I started this project, many many
workshop hours, and I had high hopes for the system.
When you set up a new construction you expect to hear something you
hadn't had before. That be a small speaker like the Illuminator-71 - or
huge speaker like the Ellipticor-84. All speakers are a compromise and
even from a small speaker we expect something, well knowing it can't
move air like a 15" bass driver. That is just a matter a of being
realistic about what we can expect from a speaker of a given size. The
smaller it gets, the more listening to music become an intellectual
exercise, where big speakers also provide some of pure physical
dimension of music.
Long story short: I was blown away! I was worried about the
midrange due to the relative high point of crossover between bass and mid, but
this worry was unsubstantiated, in fact it turned out to be really,
really good. The midrange appeared even
more clear and unconstrained than ever, and obviously is pays of not
having a midrange playing upper bass and at the same time lower treble.
Remember treble starts at 1280 Hz. The cone
really doesn't move much, even at seriously loud levels. Well, the same
lesson as learned from the
Ellipticor-84.
I took record after record to my workshop to hear
them out and every time I was blessed with new discoveries from my
record collection, not least the bass gave new dimensions to my records.
I recently bought the 2023 double album by John Scofield, Uncle
John's Band, and Vicente Archer's upright bass was a delight
through the 28W bass unit. I heard the band a month ago at Klejtrup
Musikefterskole and well - great concert, the sound seems almost better
from the LP. In a studio they can do all sorts of things, but a
magnificent recording.
For new speakers I sometimes take out some recordings I haven't heard in
a long time, this time the Uncharted Land by our late Niels
Henning Ørsted Pedersen - and, must say, a dream team of musicians. BTW,
a magnificent recording and despite some wear, still sounds great.
This is a recording rich in almost everything from deep powerful bass to shimmering treble and not least Jan Garbarek's piercing saxophones. On the mockup I could enjoy this recording at levels my wife would call insane. I know, I like music loud at times. Piano, acoustic guitar, drums...everything came through as I've never heard before.
I hasten to say the last time I heard this record might have been on a
very different setup.
The Kuzma/MSL/modded-EAR-868PL/modded-EAR-861
- and the Ellipticor-28 - brought out things in this recording I didn't know existed.
Suffice to say, I like my workshop setup very much.
Highly recommended.
I rarely buy these kind of records, but going through the list of tracks, I couldn't resist despite its ~100€ price tag. And I wasn't disappointed, these two 45 rpm vinyls are sugar for our ears - on a good system that is. Only sad the drum solo is so short - it is phenomenal. One respondent claimed I shouldn't tune a speaker from vinyl playback, rather high-res streaming. Long story short, I have done high-res Qobuz streaming and it makes no difference - none at all. I guess the respondent has never heard high quality vinyl playback. I pick whatever pleases my ears the most - and I wish all happy in their beliefs.
Making the "final" MT cabs:
Day 1, the easy part.
Fillets cut 10 deg. for the side tilt.
Day 2, the troublesome parts:
Making the lower sides and brace.
Day 3, the no less troublesome parts:
Making the front and rear panels is major trouble, all sides cut at odd
angles.
No guidance available. Have a goniometer to do your own measurements.
Day 4:
Starting on the tweeter section.
Rear panel of MT section is a complete
nightmare!
I made four before I had two suitable.
Bitumen pads in place in MT section.
Felt in MT section.
Making the felt coverings. This is 5 mm wool
felt.
I thought this would be a good idea - read below -->
Making the star shaped felt damping pads.
Download drawing to the right, open in MSPaint, set to cm and you have
the correct proportions. Print out to make Ø85 for the star shape and
make a replica on thin cardboard as a template to cut the felt with a
scalpel.
200 grams sheets wool added to MT section. ~100 grams behind brace, ~100
grams in front of brace.
Right: From bottom with spikes. Rear spike placed on 12 mm oak bar.
The "finished" speakers from front and
rear.
So, am I satisfied with the result? Bass cabinets? Yes, 100% satisfied. MT-section: No, let's say 80% satisfied with the final appearance, in particular with the back side which is a pain to construct. Also the fillets used on the front should have been maple, not oak. Can I live with it? For some time yes, but have already plans for the mk3 MT section.
Now, this all looks very nice and I didn't expect any change in sound
from the prototypes MT cabinets - but was I surprised!!!
The felt arrangement around mid and tweeter proved a disaster and the
wool filling also for the 18WE. This really didn't sound well and
measurements confirmed peaks and dips that weren't there before, so
these MT cabs were ditched - or at least the front panel design and
damping were in for change and the prototype cabs were reinstalled - and
to overcome the immediate depression, time for doing some serious tweeter testing:
D2904, D2404 and D3004 tweeters for testing.
So, what's the difference between these three tweeters?
Well, in evaluating different tweeters like here, we not only have to
focus on the intrinsic quality of the tweeter, but also on how well i
blends with the midbass. None of these tweeters can be tuned to exactly
the same frequency response although I did manage a response +/- 1 dB in
the critical range. But 1 dB in a certain area may have us favor one
tweeter for the other, so in the end also a matter of taste.
The Ellipticor D2404/552000 is undoubtedly the most neutral of the lot,
and it blends well with the 18WE driver- no surprise, so it did in my
Ellipticor-2F8 construction. This is indeed a very, very fine tweeter.
The D3004/660000 does it all so well being the most linear of the lot. I
have listened to this tweeter for months now, so I have a wide range of
recordings well in memory from this tweeter.
The D2904/710002 requires only 2/3 of the capacitance of C1 - and some
further attenuation due to its high sensitivity. The reason for
favouring the D2904 is subtle, but it did seem to reveal some extra
rosin dust from cello strings, it made a very robust treble on the
saxophone. It did everything so well and never seemed to run out of
steam. I guess the D2904 does have some extra dynamic headroom compared
to the D3004, or maybe it's just its extra sensitivity that pays off.
Playing a flamenco track that usually make me seize the remote control
convinced me to stick to the D2904, but ultimately a matter of taste.
There is no best!
Now, the D2904 looks better in black together with the 18WE drivers, so
I ordered a pair - and tests revealed the -02 and -03 are the same,
actually measuring exactly the same despite the rear damping
ring on the dome is placed differently. I know from ScanSpeak the
diameter of this damping ring is tuned for each new batch of fabric
domes. So, what is it good for - this damping ring? Normally, at high
frequencies, the center of the dome - a soft dome that is - starts
wobbling; it can't follow the motion of the voice coil. Adding a damping
ring - adding mass - decouples the center dome, and the dome starts
acting like a ring-radiator without the smear of a center dome more or
less in reverse phase of the majority of the dome. This goes on
somewhere above 15 kHz. In short, cleaner treble.
D2904/710002 and /710003 tweeters.
The damping ring on the inside of the dome has different diameter, but
don't worry, they measure exactly the same.
So, the wool felt around the tweeter in my #2 MT cabs were removed and
replaced by some temporary 6 mm plywood with black paint - and everything was back
to normal:
Summary on mounting Ellipticor drivers:
http://www.troelsgravesen.dk/Ellipticor.htm
March 2024:
As said, I wasn't completely satisfied with the appearance of my #2 cabinets, so this time I took it slow and started building up the parts:
Right hand image: Bottom panel (1) and (2) both cut 5 deg.
Left image: Panel 2 and 3 cut 1 deg.
Time to make the rear cutout for the cables to come in.
Bottom section shaping up. Making the single brace of the construction.
No dimensions given for this. Make something similar to the image.
This time I started a bit more methodically, so let's see the dimension
seen flat from the side, and break it up into rectangles.
What angles the panels have to be cut towards the center is up for you to
find out. When we start tilting the panels towards the center, the edges,
front and back,
will no longer be perpendicular to the side. Some may have to be cut
1,2, to 3 degrees to make a flat surface towards front and back. The top
some 8 deg. to form a flat top surface.
And as said before, if you have to ask questions regarding the cabinets, DON'T START BUILDING THE SPEAKER.
Starting on front panels. Front, top and rear panels were made from 25
mm HDF.
Even HDF/Valchromat is not the best thing for screws, hence adding
lacquer to the pre-drilled holes helps securing a tight fix. Fill the
holes with lacquer and let the HDF suck up the liquid.
BTW: I hate
T-nuts!
While working on the MT section mkIII, I took the speakers to our living room:
I've never had a pair of new speakers in my workshop for four months and
spent countless hours listening to all genres of music so intensively
before taking the speakers to the living room. After four months the
quality of speakers are pretty much engraved in your memory, thus the
transition to the living room was quite a bit more than surprising. I
usually know the changes that will take place, fine-tuning the treble
level and adjusting to a very different bass.
What struck me the most was the increase in overall transparence and
space around performers, the enhanced level of resolution. Treble level
didn't need adjustments as I knew what to expect from moving the
speakers to a new room. Going from a concrete environment to a 100%
wooden structure has quite some impact on bass as well as the doubling
of room size. My living room and the placement of speakers produce
limited room-gain, where the small workshop space delivers lots of
room-gain, thunderous bass, actually too much bass for a speaker this
size.
Surely the small workshop offers lots of early reflections smearing the
midrange and reducing the level of transparency, but never did I hear it
so clearly. A lightweight wooden structure will absorb quite some bass
and this took some mental adjustments as I've kind of gotten used to the
gut shaking quality in the workshop. However, no doubt the bass quality
in my living room is what it should be like, better definition and still punchy and very deep.
MT-mk3 shaping up. Ready for gluing the final top/rear panel in one.
I think I'll be 90% satisfied this time.
After 2nd coat of lacquer. The black HDF will have beeswax.
Maybe I will, after all, be 95% satisfied with the appearance...
Nothing can compete with good-enough!
So, with the final mk3 mid-tweeter cabinet. Very close to satisfied...
Although, maybe I'll use the black /700003 tweeter...
May 2024:
The Ellipticor-28 has been running now for more than a year, and would I change anything? No. It really does what I had hoped for some 12+ months ago. It fits our living room and the initial targets of sensitivity, dynamics, powerfulness and deep bass from a compact design has all been met. Mission accomplished.
One of the C1 capacitors removed to match the 7100 tweeter and new R1s
inserted.
Lid on!
Almost sorry not to view the goodies every time I pass the speakers.
A few comments on
MEASUREMENTS before you start interpreting the readings below.
First of all, if we think measurements will
tell us how a speaker sounds, we're wrong. The perception of sound is
way too subjective to be reflected in any measurements we can perform. A
loudspeaker system is meant to give us a satisfying idea of an acoustic
event and for some people a pair of 5 USD ear-plugs are enough, others
spend 200 kUSD on a truly full-range pair of speakers - and the latter
may not be happier than the former.
Measurements may give us an idea of tonal balance of a system, i.e. too
much or too little energy in certain areas, although dispersion
characteristics play a vital role here. A two-way 7+1 and a three-way
7+4+1 may display similar horizontal dispersion, yet sound very
different. Measurements may tell us about bass extension if far-field
measurements are merged with near-field measurements. In addition to
this, ports may contribute to bass extension. Most of we diy'ers do not
have access to an anechoic room for full-range measurements from
20-20000 Hz.
What cannot be seen is what kind of bass performance we get in a given
room. Bass performance is highly dependent on in-room placement of your
speaker and the same speaker can be boomy in one place and lean in
another. Actual SPL level at 1 meter distance and 2.8V input is useful
for en estimate of system sensitivity and combined with the impedance
profile may give an idea of how powerful an amplifier is needed to drive
the speaker to adequate levels.
What measurements do not tell is the very sound of the speaker unless
displaying serious linear distortion. The level of transparency, the
ability to resolve micro-details, the "speed" of the bass, etc., cannot
be derived from these data. Distortion measurements rarely tell much
unless seriously bad, and most modern drivers display low distortion
within their specified operating range.
Many people put way too much into these graphs and my comments here are
only meant as warning against over-interpretation. There are more to
good sound than what can be extracted from a few graphs. Every graph
needs interpretation in terms of what it means sonically and how it
impacts our choice of mating drivers, cabinet and crossover design.
What measurements certainly do not tell is the sonic signature of the
speaker, because speaker cones made from polypropylene, aluminum,
Kevlar, paper, glass fiber, carbon fiber, magnesium, ceramics or even
diamonds all have their way of adding spices to the stew. Nor do
measurements tell what impact the quality of the crossover components
add to the sound, from state of the art components to the cheapest of
coils and caps, they all measure the same if values are correct, yet
sound very different - in particular capacitors, coils not so much as
long as they have the proper power rating.
Frequency response of MT section @ ½ meter
distance, normalised for 2.8V/1 meter.
I dare say we reach a sensitivity of 92 dB.
And I can't reiterate enough, these graphs tell nothing about the sound
of a speaker! The
Discovery-18W mkII measures pretty much the same, but do they sound
the same? They very much don't.
Tests: A few companies favor felt around the tweeter, obviously to reduce the radiating area around the dome, so I had some 5 mm felt material and made a star shaped damping pad - and indeed, it did sound different and I was eager to do measurements to see if this would have any impact on frequency response. I was surprised to see it didn't, or at least very little. See graphs below. I had a friend sitting in and played him the w/wo scenarios and he immediately said it made better focus and 3-dimensionality. What appears to happen is a clearer treble range with better localisation of voices and instruments. Many records later it decided to keep the felt and also apply it to the midrange panel. This should later turn out to be not such a great idea.
6600 tweeter.
Horizontal dispersion at 0, 10, 20, 30 and 40
deg. off-axis. Red, green, yellow, orange and purple.
All in all, this felt has very little impact on measured performance,
but clearly audible impact.
6600 tweeter.
Above raw tweeter response with (green) and
without (red) felt. Measured 0.25 m on-axis.
This is close to being within experimental error.
Final system impedance of MT section, 6600
tweeter.
D2904 tweeter.
Horizontal dispersion @ 1 meter height @ 0-10-20-30-40 deg. off-axis.
Here's the impact on having felt
around the tweeter (not the star shaped felt, see workshop images above) flush mounted, here 6600 tweeter.
The green curve, no felt, is easy on crossover, the red curve (with
felt) cause major trouble. And the difference was audible.
Standard kit comes with Alumen-Z for the tweeter and D2904/710003 (black
faceplate).
Should you want
Amber-Z for C1, please ask Jantzen Audio.
This will cost you some EUR and the difference is subtle.
To configure
a DIY kit and get a written offer including shipping cost, please use
the DIY kit configurator in the Jantzen Audio website:
https://jantzen-audio.com/diy-kits/
All technical questions to troels.gravesen@hotmail.com
CROSSOVER-LAYOUT
BACK TO INDEX
Check this out before start making crossovers:
http://www.troelsgravesen.dk/tips.htm#CONSTRUCTION_OF_CROSSOVERS
And just in case you forgot to read this:
http://www.troelsgravesen.dk/LCR-RC.htm
The crossover is seen with three C1 (6600
version). For the the D2904 we only need two.
This saves the extra cost of the D2904 tweeter!
Speaker wiring:
My preferred setup for the bass drivers.
Balanced input from pre-amplifier.
With the power amp having ~26 dB gain, set the Hypex potentiometer in
the middle click.
Download zip file for Hypex program here.
