Flippin' Scandalous

Post date: May 19, 2013 4:39:01 AM

This all started a few weeks ago, when a Parts Express forum member, who had built two pairs of Speedsters, posted that there seemed to be something "not right" about the sound of them... too bassy, "thick" sounding. He went so far as to take some Frequency Response measurements of them, and posted them online. What he found was a pretty noticeable dip around 3500 Hz. Another forum member measured his, and found a similar sort of response.

So I took some action.

First, I measured my Speedsters. They looked like this (please note that since I use gated MLS, the measurements below 250 Hz are nil). Looks good to me. However, it should be noted that the Fountek NeoCD1.0 ribbons I used in mine seem to be from some older run, as mine have a plastic faceplate, and the "waveguide" is rectangular.

Mark from Meniscus had sent me a new pair of NeoCD1.0. These ones have a metal faceplate, and the "waveguide" is more of a rounded rectangle. I tested these new tweeters in the same cabinet with the same crossover and wiring as my original, and this is what I got:

Doesn't take a speaker design whiz to think that maybe the polarity ought to be flipped. So I did just that, and this is the measurement I got:

So it seems that what has happened is that Fountek has flipped the polarity on the NeoCD1.0 ribbon, but didn't tell anyone. I don't know exactly when this happened, but it's probably a good guess it happened around the time they re-tooled for the metal faceplate. If you have built the Speedsters, and are unsure whether yours are affected, the simplest way to test is to flip the polarity on the tweeter and listen. The version that sounds "overly hyped" and/or "thick" will be the one with the null. (Unfortunately, the "battery" trick that works great to test the polarity of woofers and tweeters doesn't apply to ribbons)

I am really sorry to all the people who this has affected. I honestly didn't know it was an issue, because my tweeters were fine. This sort of "bug tracking" is not fun at all, and I had to give up my free time over several nights in order to get to the bottom of this. It's stuff like this that make me not want to publish designs anymore. Because something like this is completely different than someone just going online and saying, "Hey, I built design XYZ by Paul Carmody, and I don't like it." That sort of thing wouldn't bother me, because--assuming he built it to spec--I liked the way the design sounded, I'm sorry he didn't. Tastes differ; that's life. But in this case, what he built and what I designed were actually two different things. It's slightly maddening.

Meniscus has been very responsive to this, and is now offering to test drivers the drivers used in kits every couple of months, to ensure that they are still within manufacturer specifications. That makes me feel much better, and hopefully we won't have to have this discussion again in the future.