By all means send me email, but please let them be
about the existing design and not about request to alter the design
presented as is.
Reserve the right to
retain editorial authority for the sake of clarity. Full names or personal
links will not be given.
The following questions from
Resistors - 15 6.8 2.7 22
Ohms, Can you suggest a power value. Would 10W metal oxide 5% be suitable.
Minimum 5W is OK. I am not sure of the sound of Metal
Oxide. I use Dale CW-5 5% series. I get them from Mouser Electronics, but it
requires quantities to get reasonable value due to shipping cost. Use
quality resistors known to have been accepted for good sound. A little
searching online should find something suitable.
Is a laminated steel core or
iron core type of coil suitable for the bass low-pass filter ?
I use laminated steel core from Jaycar Electronics -
part LF1336, A$14.50 each. they do the job nicely and cost is
reasonable. Since the load seen by the choke is 16 Ohm nominal, 1R DC
resistance would be equal to 0.25R when compared to 4 Ohm speaker/load. In
the Elsinore it is less sensitive to DC resistance than almost any other
design, neat eh?
Also, this also makes it less sensitive to core saturation, so while air
core is good, a decent metal or powder works surprisingly well. They still
need to be rated high power. Other choke coils should be air cored or audio
grade foil types.
Capacitor C3 is 68uF and you
specify "non-polar". I am intending to use SOLEN Fast Caps and a 70uF is
available. Is this suitable ? Did you have in mind a non-polarized
The value 68uF is a standard value, 70uF is not. It is
just close enough, but I would be wanting to be out more than that. Yes, I
use a standard 100V Bi-Polar, which is the same as Non-Polar. This part does
not need to be all that expensive as it works near 600Hz to max effect and
is sort of buffered by the series choke, way below the Tweeter's passband.
But it is your money.
In the Mk3 cabinet design,
the front-most panel is 18mm thick. The tweeter sits flush in the 25mm panel
just behind it. Thus the cavity in front of the tweeter is 18mm. All the
drawings pertain to a single box; do the felt drawings relate to a single
If so, the front of the four
layers of felt sits 6mm out of the cavity. OR are there only two layers of
felt for each box; thus the felt is 6mm below the cabinet front?
Correct, it is 18mm. The Tweeter must be flush mounted. It is best
that 830875 MidBass NOT to be flush mounted. I used a router for the Tweeter
set to 4-5mm depth.
The felt I sourced was rated 6.4mm and I had in mind a three layer
arrangement. But felt is a compressive material and in real life it was more
like 8.5mm. That made it almost flush with the front panel proper with two
layers - and I went with that. So check out the felt you will be using, and
get something that is more than 6.4mm for real.
These were reasonable questions,
not question requesting design changes, but about the existing design.
Thanks for your response - i will have a go - have located the Peerless
drivers in the US (cheaper than UK). Will go for MK3 without separate
tweeter panel and let you know how I get on.
Please note, use the drawings here:
Then you will NOT need the Tweeter Panel as shown
The construction will be much simpler and the end result the same.
Let me know if you had understood this as it may not have been clear on the
http://customanalogue.com/elsinore/elsinore_18.htm is for those who have
already made earlier Mk1 or Mk2 Elsinores.
The stuff on
http://customanalogue.com/elsinore/elsinore_4.htm has been updated to
reflect Mk3 approach.
Let me know.
Great website was looking to upgrade my speakers and had been looking for a
nice pair of Yamaha NS1000. Now having second thoughts and very tempted to
build a set of Elsinores.
Are you expecting to include cut out sheets for Mk3 on the website - rather
than the inserted tweeter panel would a one piece panel work as well?
I wonder how the Mk3 speakers compare to the NS1000 - would the bass be as
strong with peerless HDS 830875
Best regards, Goran
Sorry, hadn't gotten around to doing the Cutting Sheets
yet. (it is there now). I would do the MK3 without the inserted tweeter
panel as the box construction is simpler by a fair margin, but both work
quite well. Use the new Mk3 drawings (they have been updated) and Cutting Sheets.
The Elsinores are not bass shy and yes, IMO, the bass
will better than NS1000, and that is no criticism of NS1000. The Elsinores
have very low dynamic compression that will show up in the bass. In my room
they are flat down to 30Hz and useful to 25Hz - I have subs here of more
than just good quality, I just don't need to use them. It's not just the
bass extension but also the sheer QUALITY of the bass. They are world class.
Not my words, but others.
Please go ahead and build them. The DIY constructors who have built them
have only been full of praise. No disappointments and only better than
Let me know how you go.
I'm not sure which version to
build. I already expect much work for building, the finish of veneer and
slightly angled edges. Is the reason for MK3 only simplicity or better
Definitely build Mk3. It is done for BOTH reasons. First it was done to get
a better stitch at 7KHz to 8KHz - and then the whole speaker system stays
minimum phase all the way up to 16KHz. The fact that it also made box
construction easier was an extra benefit. Careful with those angled sides,
it so make it no more than 6-7mm. Otherwise the diffraction effects will be
different from those I compensated for.
Tubes with outside diameter of 90 mm are hard to
get in Germany. Do you suggest any alternatives ? If not, what is the inner
diameter of the tube.
The tube I use is 90mm outside and about 86-87mm inside. IF you use a
slightly smaller diameter, the tube needs to be a shorter than 80mm.
IF the internal diameter is 80mm the L = 65mm
IF the internal diameter is 75mm the L = 55mm
IF the internal diameter is 70mm the L = 45mm
IF the diameter is in between, the above should give you a good idea. For
example if between 75mm and 80mm then make L 60mm etc (this is called
interpolation). I think 70mm is too small, but only use if you have to.
To be sure, if you know how to measure the box tuning
(using a decent AC multimeter, a frequency meter, test amp and signal
generator), then make sure the box tuning is near 33 Hertz. The danger of
using too small diameter tube is that the peak velocity in the port is
amplified. Larger port are better, but you are limited by something less
than 100mm because internal bracing gets in the way. Get as reasonably close
to that as possible.
And I figure the Mk3
crossover + box design assumes the HDS 810921/830875 combination. But if I
have/decide to go with the PPB drivers in stead of the nomex ones, is it
enough to return C2 to 3.9uF?
Correct on the first one. If using 830874 PPB plus new HDS tweeter, then
increase both the resistor (to 3R9 or 4R) AND use 4.7uF - then you have the
PPB version of the Mk3.
Mk3 MUST use HDS tweeter!!!
Clearly much work and experience are embodied in your Elsinore LS design
which I have just come across when looking up the Peerless 830875 on the
I wondered at seeing the square-wave spikes and transient pulse reverse
spike until I saw the crossover diagram with its reverse connected tweeter.
I am aware of crossover phase changes upsetting an overall response at the
crossover itself, but the transient and steady state responses are clearly
'at odds', and this puts me off.
Is there no way this can be corrected ?
Don't be put off by this - everything is not always
what it seems and there is something interesting going on here.
Much research has been done on this subject. What is not clearly understood
is that a tweeter connected in phase, most of its output is actually
NEGATIVE. The OPPOSITE is also true and that is why it is used. The tweeters
rise time if 0.6 millisecond - it only shows up well beyond 10KHz. When this
rise time is negative, the rest of the output is positive!
Hard to believe but most of the tweeter's output is POSITIVE! The proof has
been supplied, read on.
There is an extremely exhaustive explanation given here:
I know some of it is rather technical, but an attempt has been made to make
it least so. Even if you don't understand all of it, read it right through
to the end and I think some of it will be clear, if not all the details.
The reason a reversed tweeter can be electrically out of phase is because
the 90° phase shift which has incomplete summing of the acoustic
output of the drivers, but by allowing phase shift by both drivers
(complimentary phase) the tweeter is actually locked in phase acoustically
when it is connected out of phase electrically. The alignment of the drivers
must then compliment that arrangement.
The Elsinores - and this is measured - is minimum phase up to 16KHz.
That is why it produces up to >1Khz square waves and this amazes so much
that it augers disbelief. Too bad for the non-believers.
Yes, the fact that the tweeter is electrically
out of phase does not show up until 16KHz.
The research of Linkwitz & Riley have made these facts well laid down - as
you read between the lines the rule is get the phase right. But it is
not widely known or understood, that can only happen when you invert
electrically - them are the physics. The fact that L&R proposes 2nd
and 4th order does not invalidate the facts even if they don't recommend 1st
order. Fact are facts.
Read the page as it extols the virtue of
acoustic phase as opposed to electrical phase (you cannot have both). The
benefits are multiple.
Re Mk 3, is the rear tweeter
stiffening panel still required, it makes it three inches?
I somehow knew somebody was going to ask this question. The answer is YES!
By making the 25mm Sub Front Panel ONE piece, I have actually weakened it. I
know it doesn't look obvious.
Much of the construction and bracing is based on thinking by R&D at KEF
during the Fincham/Cook era. In fact Raymond Cook used to show that
longitudinal bracing was superior to lateral and/or edge mass load/bracing.
He did this during talks and showing how different panels audibly sounded
with different types/methods of bracing. The rules are simple once
understood. It is understanding mass versus stiffness. Longitudinal bracing
where the brace is off centre (asymmetrical) is better than lateral bracing.
Bracing the edges of a panel is totally ineffective. That means the joins of
the box have little bracing value.
The original box design meant that the Tweeter was mounted where the box has
its greatest strength and rigidity as well as concentrated mass. The Tweeter
Panel was set behind two 43mm thick panel, the small dimension of the panel
and the lateral (front to back stability) was rated good under those
conditions. It is 230mm by 200mm and 25mm thick. Try to make that bend or
bow. The more 'cubic' the shape the harder it will be to excite. Whereas the
same amount of the same material made thin and larger in other dimensions,
will vibrate like crazy. They will both have the same mass though. It is
mass AND stiffness that counts.
The two 43mm thicknesses in front of it means lateral mass AND stiffness is
high. Simple really.
Making the 25mm Sub Front Panel a single piece and mounting the tweeter on
it meant greater potential longitudinal bowing (like a violin) of the Front
Panel. This causes the lateral stability of the tweeter is weakened, it is
more likely to move to and fro. The tweeter operates at frequencies where
the wavelengths are small - that tweeter must be as stable as possible and
not move backward or forward. Or you lose information and 'focus'. Time
smear is the result.
Also, it means the internal volume will stay the same, although this is of
Hope that is clearer.
In your experience can I
drive these with 25w per channel valve amps?
Absolutely. In fact a 25 Watt amp with these speakers sound like 100 Watt
with other typical speakers. The Elsinore goes LOUD with not so many Watts.
I use mine with 30 Watt tube amps - and I have power to spare.
Hans Hammerbauer in Switzerland tried the 18 Watt Vacuum State DPA-300B amps
($16,000) rated 18 Watts and found that was plenty in an average room. That
just makes the Elsinores an even better bargain.
I am interesting in building the Elsinore loudspeaker. Please provide some
clarity on the wave guide wedge. This the wave guide used on the final
My understanding was it assisted in using a low crossover point and provided
the right acoustic load. Mk3 appears not to use the waveguide? Yes/No..
Aligning the impulse/step response to line up properly, so that the point at
which the Tweeter's impulse goes positive and lining that up in time when
the impulse of the MidBass driver also goes positive - necessitated that the
tweeter mounted about 29mm behind the MidBass' mounting flange to mounting
flange. The correction amounts to about 60microSeconds - which means the
Elsinores end up being minimum phase to about 16KHz.
Hence the waveguide is now shallower and it was decided to use felt. It also
made Mk3 box construction simpler. I like the idea of waveguides, but the
time alignment was more important. It is still a shallow waveguide.
Interestingly, there is evidence that waveguides actually helps build up off
axis response - the tweeter beams less. This shallower waveguide still does
If you haven't read this more recent page, then this will explain further
the rationale behind the time alignment, especially towards the end of the
BTW, if you are in Sydney, please come around for a closer look and listen.
You are most welcome. The number of people building the Elsinores continues
to increase, but the proof is in the listening. Most speakers sounds like
toys in comparison as speakers that produce real scale (whether small or
large) are either rare and/or expensive.
Just ran across your design
on diyaudio.com. Looks really great; I am thinking about giving it a try. I
would choose the Nomex woofers and the HDS tweeter or maybe a seas excel or
Fountek ribbon. My main question - would I use the same crossover if I
wanted to turn the drivers so that the cut sides are vertical, and therefore
could go with a narrower baffle? I would keep the box volume the same by
making it deeper. (name with-held)
(Take the following in the right
spirit and do not be discouraged building the Elsinores as is.)
Please avoid asking these kinds of questions. I am a
designer for hire and simply can't afford the time re-designing the Elsinores all
over again with different drivers and other changed parameters. It is a gift - and a
valuable one at that - and you have permission only to build a personal pair and
NOT FOR PROFIT. The design is provided as is -
the complexities are enormous and the variables uncountable. I cannot
predict many changes you may have in mind and have not the time or
inclination to sort out even a small number of them, out of necessity and
Ribbons? They won't work in this design as they rarely
have enough bandwidth to suit here.
Narrower baffle? Diffraction and step response
compensation now won't work with existing crossover. Your tonal balance
could be all wrong. Any ideas how critical the width of the cabinet is? More
than a few millimetres could create a different balance. Would
you suggest a boat with a narrower keel to a boat designer? He might suggest
it would sink, right?
Different tweeter? All acoustic measurements needs to
be remade of all drivers and remodelling the crossover, and only if the
tweeter has the right characteristics. You are asking me to buy the driver
and crossover components? And then spend dozens of hours in integrating the
new tweeter? I cannot afford to respond to every whim and the results may
not even be up to scratch.
Any other questions? Please think first what you are
asking, OK? If somebody gave you a present, would you have thought it good
manners to make suggestions after the fact. I suspect you wouldn't.
So, bottom line. Either build a design or build
a disaster. But if you stick to the plan, chances are
you will be overjoyed by the results and no headaches for
Hi Joe. No, you got me wrong
- I would never think of asking for more ! I think that your providing the
info that you already have is awesome for the amateur looking to build a
pair of speakers..
If anybody wants to make changes, it is no longer my design, but theirs (go
right ahead - but don't put my name to it). If I recommend design changes it
is my reputation that is on the line and I have a lot more to lose. I have
always retained a right to veto a design when dealing with people in this
business. IF I recommend something I have to be right (at least satisfied in
my own mind). This is not vanity but hard-headed reality. I did not
make the world the way it operates.
Although I am a audio nut, I know almost nothing about speaker design (as
you can probably tell by my question).. I was just asking a question about
how the design would work if I narrowed the front baffle. Thanks very much.
An Example of a Single Development Cycle
this actually changes the frequency response MAJORLY. Even bevelling the
edges will change the response as will any
change to the front baffle.
The reason I did NOT specify bevelled edges is
because not all home constructors are able to do a good job of that. If this
was a commercial design I would have liked bevelled edges. But ANY changes
to the front baffle (or anything else) means pulling out and setting up a
bunch of equipment, disconnecting the crossover, measure the acoustic
responses of ALL drivers, both far-field and nearfield, ON axis and OFF axis
sets. Put away equipment (space is always a premium), merge farfield and
nearfield files and reconstructions with the electrical measurements of
drivers (which also must be done IN-box) into complex transfer
characteristics files of each in-situ driver - complete Hilbert-Bode
transforms (a strict mathematical relationship between amplitude and phase)
from accurate calibrated data. And at this point we have only started.
ONLY NOW CAN WE reconstruct the overall
project file and start the analysis.
Want to count the hours so far... ???
Now we make the required changes to the crossover (and if the required parts
are not to hand that means spending money). Then we pull out all the
equipment again to check and verify that our modelling is correct. You must
always start with the HARDWARE (calibrated measurements of the 'bits'), then
move into SOFTWARE (reconstructions and modelling) and then finally back to
HARDWARE to complete/verify the development LOOP.
Often some fine tuning is still required (not just listening, yet to
come), such as
in-room 1/3rd and 1/6th octave pink noise. As with the last change of
tweeter - to the current HDS tweeter - this showed up that there was still
possibly too much energy in the presence region. Going back to the modelling
showed that the component that had the single largest effect on this area of
the response, was the main series capacitor to the tweeter. Other components
in combination would also do the trick, but there were other complications.
It was decided that the cap had to be tweaked VERY carefully.
The speakers were hauled to a different locations and with other trusted ears
and we did fine tune that - it resulted in a cap value reduction. It was
decided that we should go for no artificial boost in the presence region,
and YES, at lot of commercial speaker DO exactly that and even horrendously
priced ones. But the Elsinores now utterly avoid it and is better because of
it. In OUR opinion (yes, there is a BUNCH of people involved, but I have the
chief design responsibility).
I just want to give an idea of the process involved - and I will be posting
this, or the substance of it, in the FAQ page (that is why I have spent a
bit more time on this than I would otherwise have)...
Please don't feel discouraged in building the
Elsinores. You will benefit from a huge amount of work, not only upon my
shoulders but also the help I have received kindly from others.
And you can always say to your fellow listeners that
you had a sparring match with the designer.