The "Elsinore Project" Thread

Joe,
What is crossover frequency point between woofer, mid-range and tweeter?

It is technically a 2.5 way, the upper Crossover is near 3KHz and the bottom two MidBass drivers are gradually rolled out above 400-500 Hertz. This does a good job of correcting the diffraction loss of the box and also to keep sensitivity high.

Cheers, Joe

.
 
Hi Joe,

Did the new rebate for the speakers remove the old speaker screw holes.

Or do they need to be bogged up before routing.

By calculation its a close call.

Also if 2 x 10mh with rcd of 3.6R was used could the resistor be eliminated.

Cheers

I was expecting you would chip in - what has happened was not supposed to have happened.

I didn't have to bog up the holes beforehand. I found that after 7mm rebating, that the new drivers would seal them up anyway - they have inbuilt gaskets, really nice touch.

Re 2 x 10mH, if the DCR makes up the whole value of the chokes, then in theory the answer is yes. But that could only be achieved by reducing the wire gauge used to wind the choke(s). But I would not necessarily recommend it as now the choke will dissipate a lot more heat and it would be better to share that with a minimum 10 Watt external resistor.

But keep in mind that 2 x 9mH or 18mH is also acceptable, only 20mH is ideal, but would be splitting hairs.

Cheers, Joe

.
 
Thanks Joe. Old Hamlet crossover post below.

Hi Guys

Here we go. It been coming for some time and I thank you for your patience. Not only for the requests made here but also more directly via emails.

PROVISIONAL HAMLET CROSSOVER

Hamlet_xover_provis.gif


The results here are based on modeling results from the Elsinores. I think that this provional crossover should be reasonably close to the final posted result. Compare both, the Elsinore and provisional Hamlet crossovers, and you will see that many of the component values are the same. You would expect the 4KHz LCR trap and the LC for the tweeter to be identical. The parallel LR was changed to cope with the different diffraction loss, the "step" that is not filled in by the addditional pair of bass drivers. This step/loss is based on the Elsinore box and since the width of the box is the same, we should be in a similar ball park. The unknown is the additional baffle area below the two MidBass that will be missing in the Hamlet. But since I see a loss of sensitivity, based on practical experience, I would say that -5dB should be pretty close and certainly within 1dB. Eventually when Rob comes up with the actual box to work with, all this will be revealed.

Let us look at the Frequency Response:

Hamlet_Freq_provis.gif


This would indicate about or close to 87dB sensitivity. Considering that this is achieved with high Z like the BBC LS3/5A which was also 16 Ohm but 81dB sensitivity. This will be a speaker that is not current hungry. Let take a look at...

The Z (Impedance) Curve:

Hamlet_Z_provis.gif


Below 3KHz this comforms to a 16 Ohm speaker, above that it qualifies as 12 Ohm since the Z does not go below 8.5 Ohm (divide the nominal Z by 1.414 is the convention used).

Now that Z curve only tells a partial story, it needs to be looked at in the context of...

The Impedance Phase:

Hamlet_Z-Phase_provis.gif


Mildly negative. The lowest is around 85 Hertz and -50 degrees. But the Z plot shows it is 28 Ohm. This is no problem. The second lowest is 2.5KHz and -20 degrees. Yet the Z is a very high 12 Ohm. Above 6KHz the phase is positive.

One of the major features of the Elsinore is its phase response...

Minimum Phase:

Hamlet_A-Phase_provis.gif


We can see then minimum phase extends from 20 Hertz to almost 20KHz. This is depite breaking the rule and wiring the tweeter in reverse phase based on The Renegade Tweeter Theory that states that a fast rising tweeter inverts the acoustic phase relative to the electrical phase below a specific frequency and input level. Yes, tweeters are Invertors.

I also looked at simulating that time response using a square wave:

Simulated Square Wave @ 750 Hertz:

An externally hosted image should be here but it was not working when we last tested it.


The spikes you see in this example show up beyond 17KHz as shown by the minimum phase plot.

Note the two last graphs are based on 4 Watts input, or 5.65V RMS. With lower levels the phase improves further and above comes down in frequency.

Coming back to those spikes, they largely disappear with actual measurement as opposed to simulation. The following example is the Elsinore:

Sq_700Hz_2M_15-OFF-4_750W.gif


Notice how much the spikes are truncated, and as this is a real measurement rather than a simulated one, we can see that in the real world The Renegade Tweeter Theory works even better than one might think. Reality is better than even the theory indicates.

Conclusion: The Hamlet retains key features use in the Elsinore Project speaker. In addition to above it retains the use of series connection of the main drivers. In all honesty I have not seen this used in any commercial speaker and come to think of it, I can't recall any diy designs either. Normally we would connect the drivers in parallel and get a 3-4 Ohm speaker. Yes, the voltage sensitivity would likely match the Elsinore, but the current required would quadruple, so efficiency is not improved, just a heavier load on the amp. I make no excuses for being in a particular camp (a minority one I would say).

OK, over to you guys.

Cheers.

Joe R.
 
Last edited:
Joe, I am loving this thread of yours, and I think you are onto something with your SB Acoustics low inductance drivers, simple small value filters and a flat impedance. The best threads make you THINK! :cool:

But I want to rain on your parade a bit, and maybe help you take something into consideration. I had a good ding dong with MarkK about his First order plus Fs notch tweeter circuit in the ER18DXT. I said he would do much better to use a third order butterworth. I'm going to justify that here.

There's a problem with notches, whether at tweeter Fs or at the top end of woofers' natural 3-6kHz cone resonance. They spoil the natural rolloff even if they improve phase alignment.

Below are some simulations with a tweeter Fs around 1.3kHz. My conclusion is inescapable. Third Order butterworth lets through far less low frequencies than a first order plus Fs notch, which means lower intermodulation distortion. Having built KEF Acoustic Butterworth notch filters, thinking they would be near perfect, I actually found they sounded horrible. What they are really, is "First Order Improved", especially on phase and suppression of the Fs resonance. Improved to the point where you can actually hear how horrible first order filters generally are! :eek:

What is interesting is that notches generally aren't needed to align phase when you time align the drivers. Which might be something to consider. :)
 

Attachments

  • System7 First Order.png
    System7 First Order.png
    12.5 KB · Views: 319
  • System7 Acoustic Butterworth.png
    System7 Acoustic Butterworth.png
    15.1 KB · Views: 311
  • System7 Third Order.png
    System7 Third Order.png
    12.1 KB · Views: 307
  • Impedance Acoustic Butterworth Circuit.PNG
    Impedance Acoustic Butterworth Circuit.PNG
    15.1 KB · Views: 317
  • Impedance 3rd Order.PNG
    Impedance 3rd Order.PNG
    13.4 KB · Views: 305
  • KEF 3rd Order Acoustic Butterworth.JPG
    KEF 3rd Order Acoustic Butterworth.JPG
    42.9 KB · Views: 243
Can anyone recommend if there will be much of an audible difference with an internal crossover rather than external? Can't seem to find much about this in previous posts

Should be no audible difference but if made internal, tweaking crossover component values or upgrading to higher spec components is much more difficult. I've made several component changes to mine over the last two and a half years. I would recommend external.
 
Member
Joined 2004
Paid Member
Joe & Co.:

I'm new to building speakers but have been approached by a friend with access to a professional woodworking shop that will build enclosures for us at a very attractive price. We are focused on the Elsinore, in part because of the detailed information available and the knowledge of its fan base. One of the unknowns we are currently wrestling with is whether the waveguides for the Elsinore Mark VI version are commercially available or must be assembled (and if so, from what). Frankly, I have no idea what it takes to construct a waveguide or the skills and equipment needed to make them properly. What's the story?

Thanks,
Scott