The "Elsinore Project" Thread

[johnmark]

Hey Rob, any news on the Hamlet boxes front?

Cheers, Joe



Joe,

Do you think you will try a 2.5 way crossover for the hamlet as well?
Been wanting to build the Elsinore in a more compact form to make a matching lcr set with subs and the Hamlets seems like the ticket. I am guessing that a 2.5 way might have wider off axis response? And Higher efficiency?

Thanks for all the hard work

John

[critofur]

Originally Posted by Francois G

Hey guys, does someone have an answer or opinion for me? Am I missing something important in the requirement for having two separate boards?

Francois

Yes. If you move the tweeter forward/backward relative to the woofers it changes the time alignment and phase alignment (two separate but related issues).

[Francois G]

Critofur, I appreciate for your reply. But I'm asking a diffrent question than the question you (and apparently everyone else) thought I asked. I will try to ask and explain better. (Sorry to be long winded - if I could draw this it would be much more understandable)

Joe's design has the tweeter mounted on a "sub-front" which has openings for the woofer magnets. Then there are two front pieces for each speaker, an upper and lower front. The upper front has one whole for one woofer and the lower has three.

I'I would like to use a single front in nice 19mm oak with a rectangular opening the same height as the original space between upper and lower fronts, still mounting the tweeter on the sub-front 19 mm behind the woofers, but keep the upper and lower fronts together by making the opening 50 mm narrower on each side. My proposed single front has the woofer wholes as before plus a rectangular opening that fits over the tweeter.

The question is really is: "Does it matter if the recessed area does not go all the way across the front (everything is the same, except framing the opening by 50mm wood on each side. It also means the felt is as before, except it stops 50mm short against the "frame" compared to the standard design.

[Joe Rasmussen]

Originally Posted by Francois G

Hey guys, does someone have an answer or opinion for me? Am I missing something important in the requirement for having two separate boards?

Francois

The idea was to make the boxes as easy to construct as possible. But if you made it one piece it would require good skills to cut out the Tweeter cavitity that many would not have or not have the equipment to do it neatly as it is highly visible.

But if you were to make the width smaller and the felt less wide, it should work provided you don't reduce the width of the cavity too much, maybe a max of 15mm either side (which will reduce cavity width by 30mm). We don't want to create diffraction effects that has not been compensated for, espescially in the crossover.

Cheers, Joe

[Francois G]

Joe,

Thanks, I appreciate your reply. It is a bit of a hassle to do this way but I'm trying to avoid the exposed edges of the felt.

Francois

[Joe Rasmussen]

Originally Posted by johnmark

Joe,

Do you think you will try a 2.5 way crossover for the hamlet as well?
Been wanting to build the Elsinore in a more compact form to make a matching lcr set with subs and the Hamlets seems like the ticket. I am guessing that a 2.5 way might have wider off axis response? And Higher efficiency?

Thanks for all the hard work

John

You are welcome.

Doing 2.5 way was not the idea and there is no reason to expect higher efficiency. What would be explored is parallel versus series connection of Nomex drivers. Parallel will certainly increase the voltage sensitivity but also become nominal 4 Ohm and hence double the current relative to 8 Ohm. So the actual efficiency is the same, only voltage sensitivity goes up. The same goes for 2.5 way as below a certain frequency (diffraction loss determines that frequency) the overall the system impedance will be 4 Ohm and there is no gain or loss relatively speaking.

But doing it MTM will mean the Hamlets will be symmetrical both laterally and vertically a la John Dunlavy preached. This makes it virtually a true point source speaker if you get all the other things right like the phase etc.

I better explain this idea of Dunlavy. The Tweeter will be in the dead centre of the front panel or baffle. The baffle above and below the Tweeter to be mirror image pairs (vertical symmetry) and the baffle either side of the Tweeter also mirror image pairs (lateral symmetry). Needless to say, vertical symmetry includes drivers both above and below the Tweeter. Of course all drivers need to be vertically aligned. Now such a speaker, and look at the Hamlet boxes, with the drivers "time" aligned correctly on the front panel (baffle) can have both point source and also symmetrical response both laterally and vertically.

While the Elsinores are only point source from roughly 400 Hertz up (but not truly symmetrical), the Hamlets will be proper point source in true Dunlavy terms. What I am trying to say is that the Hamlets have to be a compliment to the Elsinores and share as many characteristics as possible as multi-channel systems will need both Elsinores and Hamlets. This will also mean the sound fields they generate will be similar and sonically compatible. That was always what we wanted to do.

Cheers, Joe

[Joe Rasmussen]

Originally Posted by Dan_E10

I noticed the number quoted at Madisound for the HDS tweeter is 93 db. However in Scan Speak's pdf, they quote 91.26 db at 2.8V, almost two dB lower. Has the tweeter changed since it was a Peerless design or maybe they just rated it more conservatively?
Dan

I suppose that question is understandable. But what I suspect (and hope) is that the driver has not changed and they've only adjusted the figure. The real question is how do you measure dBSPL of a driver relative to 2.83V when the frequency is not flat? You cannot pick a discrete frequency but have to pick a band of frequencies where it averages such and such dBSPL.

But that is only the start of it. It is really difficult to set up "calibrated" dBSPL measurement to start with. What I do is use a reference Tweeter - the Vifa XT25 that has the flattest frequency response I have ever seen - and set the equipment up as the flat part of its response of the XT25 is my reference for 94dB @ 2.83V @ 0.71 Metre. That sets up my calibration so I can set if at any point calculated from my reference.

No two will have the exact same reference. Even my reference measurement I only use for that particular session. I have to go through the same setup procedure every next session/project. It will be close but never the exact same - but relative measurements during that session will match.

But back to the Tweeter, I expect only the measurement has changed and the Tweeter is the same.

Cheers, Joe

[Joe Rasmussen]

ATTENTION ELSINORE DIY CONSTRUCTORS

Potential Crossover Upgrade Needs Road Testing:

The details are yet to come, but it requires six 33 Ohm resistors rated a minimum of 5 Watt - please use good quality grade, by that I suppose I mean by audiophile standards. They don't need to be ultra-low inductance type, but good ones usually are.

These resistors are added to the existing crossover - no existing values are removed or changed.

I will explain further how exactly to fit those resistors and why.

Do I have any takers?


Cheers, Joe

[mantraxl]

I'm buying xover components for the concrete elsinores right now, so why not play with the new resistor idea, right?

[Joe Rasmussen]

Originally Posted by mantraxl

I'm buying xover components for the concrete elsinores right now, so why not play with the new resistor idea, right?

Yes, that would be good. Whether we end up with six times 33R will depend as six 47R could also be tried - feedback from different users would be welcome. The value of all the resistor need to be the same, three places in the crossover and the max value 100R as above that the difference is small.

I am listening to 33R at the moment and more than just liking it - with the amp I am using it really works well. Probably more gains if tubes are used and less, but maybe still important, with solid state amps.

Cheers, Joe