Friends,
Well...the verdict is in - and it ain't perty. Seems like I have a personal issue of having to do everything in a "big way" - especially when I screw up.
The story goes like this:
During the recent build of this last batch of Timepieces that went out, I had a little time to do some experimentation. I've always wanted to look a little more deeply into the damping material we've been using in a certain tweeter sub-chamber. You know...ongoing product improvement - it's always a good thing - right?
Well...we found a new material that definitely made an improvement with regards to unwanted resonances. OK - lets use it. It was a relatively simple change to implement before the new batch went out. There was a side effect though - it caused a small 1dB drop-off in the tweeter's bottom-end response, right in the crossover region.
No big deal. Just tweak a couple part values to flatten the response again. Yeah, right. In my grand assessment of the matter I assumed that it shouldn't make much difference in the electrical impedance response, so I didn't bother to take the time to measure it. It's a bit of a hassle to change the set-up and re-calibrate, and I was in a hurry (as always). Well, you know what they say about the word "assume." Stupid man.
And I'm the first to bitch at the other guys in our shop about doing the same thing in their areas!
The upshot is that making that change had a major effect on changing the electrical phase angle of the impedance. We're talking a huge negative value. That means a highly capacitive load at about 500Hz. To be specific for the "techies" out there, you can well imagine what a -56 degree phase angle will do to an amp - regardless of actual impedance magnitude.
For the non-technical, I'll try to explain. In a purely capacitive circuit being driven by a sine wave, the current delivered by the amp leads the voltage output by 90 degrees. That means that when the current drawn by the capacitor is at its greatest peak value, the amp is not putting out any voltage to drive the current into the capacitor. That means a major conflict and sort of a "tug of war" going on between the load and the amp driving it. This is not good for any amp - even those of the most robust current capability - although that type will handle the situation better than any other.
Nevertheless, a highly capacitive circuit is not good for any amp. As the electrical phase angle of the load gets closer to approaching -90 degrees, the amp has to progressively work harder and under high power conditions, can actually cause the amp to get hot and either thermal out and shut down...or blow up.
In practice, the negative phase angle should never exceed -45 degrees and in all actuality shouldn't even approach that. A good rule of thumb is that -30 degrees should be the maximum.
Now days most modern amps have "current limit" circuits to protect them from such conditions, which would also include an accidental short circuit on the amp's output. Every time such current limiting "kicks in"...all bets are off with regards to the resulting waveform resembling anything of "fidelity." In fact, we're talking absolute "butt ugly" when that happens. Most current limit circuits don't shut the amp down - they just "clip" the driving signal from the driver stage to the output stage for the duration of the waveform that represents an excess of current draw by the load (speaker). This can happen for every half-cycle of the waveform at whatever frequency(ies) the condition exists.
In the case of these most recent Timepieces, that condition could exist in a narrow band, either side of 500Hz. Because that band is actually quite narrow and the power density of music is not very high in that region, the amp driving them is unlikely to experience much excess heating or be at risk of any failure. But...if any such current limit circuits were to be "tickled" into turning on, the result would definitely be a "hard" or "forward" sound. Not only would the certain frequency be slightly "clipped," but doing so would generate higher frequency harmonic distortion products that could be audible for at least out to 3 KHz or so.
Even if any such current limit circuits were not engaged at any time, the amp's damping factor would be greatly reduced allowing the LC resonant components of the crossover to "ring" right in the 500Hz range. That would generate a very forward "honky" sound - sort of like the "cupped hand/megaphone effect.
So...if that's the sound you guys were hearing, you can thank ol' dumb-arse - yours truly.
In summary, the auditions by
Carlman, Kenyyuan et.al. were nigh unto a waste of their time. In the last three days I have completely corrected the issue and now the worst-case negative phase angle is -26 degrees...less than half that of what this last batch had. That is a major improvement and will make the Timepieces a relatively easy load for most any amp. Not only that, but we achieved an even flatter response than before. Believe me...with this correction installed the Timepieces will sound almost like completely different speakers.
So, all of the last batch of Timepieces need to come back to the factory for this correction. If the above-mentioned guys don't want to bother with a second go-around audition, I completely understand. Please accept my apologies though, as I'm sure you can imagine this situation was not intentional. For those that want to send their units back, we'll work out the details on an individual basis.
I will be posting graphs of the "before" and "after" impedance plots soon so those interested can see for themselves.
-Bob
