"Proac 2.5 clone" or "Elsinore project"

[Joe Rasmussen]

A little belatedly I thought I'd add my tuppence worth.

There is always an inherent and potential problem when doing a high-end DIY speaker design: That somebody beyond the control of the original designer may not get the design exactly right. Guess who may have to unfairly wear criticism? I know of at least one instance where the DIY'er dispensed with the passive crossover and went for an active crossover. Other than providing the most basic of advice, it is impossible to predict how something like that will work out. Not surprisingly, the Tweeter was turfed out and replaced with a "better" unit. If only things were that simple.

What follows below will help explain why the Tweeter didn't work out.

Also: Even the slightest chance of somebody getting it wrong in the crossover can have a dramatic effect on the performance. Now I will bare myself, even I have made a mistake, but realising it very quickly. I would of course quickly know, where others might not. But crossover mistakes are easy to make.

The issue re the HDS cum Scan-Speak Discovery Tweeter: Distortion in Tweeters is a major issue, particularly in dome Tweeters.

We do not have the benefit of LR4 crossover here - so we need to come up with a working solution.

The solution in getting low distortion out of the HDS Tweeter in a 1st order crossover, is how you control it's output below 2KHz and the source impedance the Tweeter sees looking backwards. In the Elsinores it sees a sub 1 Ohm at circa 650 Hertz (Fs), keeping the dome output down to less than 10% compared to many other situations. This is done by a carefully tuned null-filter and it dramatically lowers distortion. The Q of the null is critical as it has significant effect on the output in the 1-2KHz octave.

(BTW, the same trick also works well with XT25, but the HDS is better.)

The response is 1st Order for at least one octave below 3Khz and then goes into a 5th-6th Order fall which bottoms out below 1 Ohm as discussed above. See graph.

Need I also mention the phase characteristics? These are also well controlled.

Thanks for Joel - I appreciate him starting this thread.

Cheers, Joe R.