SL2000 response charts in sda compendium show no spikes

daifanshi wrote: »

Instead of bribing and theorizing to get to the end conclusion. The correct way to determine what is going on with the SL2000 tweeters is to make careful measurements of a physical sample. And they do not have to be anaechoic sine-wave sweeps. Gated burst tone measurements are mathematically equivalent and can be done almost anywhere. And as far as I know, there are no publically available distortion tests available either. We need these too.

Plots of the in-room response of the entire speaker system such as those found in magazines or done with your SPL meter in your living room are of limited utility, since the environment is not typically known or repeatable, and over-averaging can really skew the results. System tests also can't be used to extract accurate imformation about the drivers because of crossover electrical effects.

I have already started the process of measurements of the SL3000 tweeter and MW6510 woofer for my own purposes. I finished my impedance measurements already and will be able to extract the Thiel-Small parameters shortly. These will be as close to reality as one can get and not taken from fresh engineering samples built and measured in the mid 80's, but well-aged and broken-in units.

With respect to the tweeter, there are several things that can cause "sibilance", "peaking", "harshness", etc... From reading this board, the folklore here says we have a "spike", "peak", "bump" in the frequency response around 13kHz.

Instantaneous frequency response is only a superficial indication that a problem might exist. I will make "standard" frequency response measurements but also important are the following:

1. Stored energy in the tweeter response. Waterfall plots are useful in seeing this. e.g. A "peak" can sound bad, but a smaller "peak" will sound worse if it stays with the tweeter for several hundred milliseconds. I have the ability to test this. I've heard this called "linear" distortion.

2. Non-linear distortion. Also known as intermodulation distortion. These are basically mixing products (a non-linear effect) that happen when multiple-tones (as in music) modulate with one another and produce other tones. These "sub-tones" can be large compared to the desired tones and the odd-order tones are close in to the desired tones, which can make them the worst offending. This distortion is different depending on frequency and is always worse when operating a tweeter at too low a frequency. This is due to the non-linearities in the motor when there are large excursions. I can make this measurement as well.

3. And of course all the effects I listed above are power-dependent. I can test this but I won't. Too tedious. But the effect can be very significant as the voice coil heats up.


Cheers.