For me it would be an exercise in amp tuning to get the most out of the speaker.
Something like: a current amplifier with a DC servo that brings in a little voltage feedback in the bass but then rolls off.
Active filters. The way I understand it, an active notch at 5kHz wouldn't normally work because of the short-circuiting action of the amp, but in this case the output impedance would be high at that frequency.
Absolutely at line level, especially active. As Lars explained, at speaker level a different set of conditions comes into force.
My thoughts exactly. A driver of this size (Sd ~= 130 cm^2) is best utilized below 2k due to narrowing dispersion. This new driver is ideal for use up to 2khz.
Getting a pure piston to 4 or would be great. I mean really great as that is far better for most tweeters, just that darn breakup. It looks like I might try the PTT6.5X08-NAA-08 to 2500 or 3K, but would need a really steep crossover. I do not like steep crossovers. Maybe active DSP would be a better choice.
Any plans to apply this to the 4 inch? Push the breakup further up and maybe, finally, we can have that killer midrange. 500 to 5K. Just like the old monkey coffin days with "decade" drivers. It really was a good formula.
I am sure there is more that can be done to distribute the breakup to different bands, different Qs, without upsetting the mass distribution. Modern simulations can do wonders we never dreamed of in the old days. I am glad Purifi is taking this on, but not convinced me yet. Just ordered a set of CSS to do a "cheap build" as they look on paper ( pun intended) to be able to reach an XT-25. I like the VIfa again, for very well behaved above the pass band.
Any plans to apply this to the 4 inch? Push the breakup further up and maybe, finally, we can have that killer midrange. 500 to 5K. Just like the old monkey coffin days with "decade" drivers. It really was a good formula.
I am sure there is more that can be done to distribute the breakup to different bands, different Qs, without upsetting the mass distribution. Modern simulations can do wonders we never dreamed of in the old days. I am glad Purifi is taking this on, but not convinced me yet. Just ordered a set of CSS to do a "cheap build" as they look on paper ( pun intended) to be able to reach an XT-25. I like the VIfa again, for very well behaved above the pass band.
I'd call that a bit of a waste, since it would be sacrificing some of the good qualities of the driver purely for the sake of not using a (not very large or costly) inductor and capacitor.
Wouldn't that depend on what the tweeter is (in addition to spacing on the X & Y axis)? Not all 1in domes have their own coils in the same location -especially those with waveguides. And that's before you introduce different tweeter sizes. So unless Lars can default to a single example, I'm not sure it'd be especially practical.
What is AC offset? If an AC voltage signal has an offset, wouldn't that be a DC offset? I am confused by what you mean
AC (presumably) was referring to 'Acoustic Centre'.
OK, if we're on the 'I want breakup distributed' concept, let's look at this logically for a minute. This unit, like some others, obviously has the design objective of concentrating the main modes to a single well-defined peak (the secondaries don't start until almost an octave higher, which gives a good idea of how effective this cone profile must be), and act more or less as a pure piston below that. A single well-defined peak, especially at a reasonably high frequency, is a hell of a lot easier to deal with in a crossover than multiples, as well as resulting in lower overall distortion levels. Multiple distributed modes are certainly appropriate for a wideband driver like a Jordan, MA etc. -that's how they work. But that is not what this is for: it's a midbass with very different design method & priorities. Lars & Purifi are obviously aiming for a slightly less ambitious, but still quite broad BW over which the driver is more or less unresonant (i.e. 'pistonic') with very low HD and a stopband mode that is easily dealt with in the filter. True, that does partly preclude a low-order filter at a relatively high frequency, as was suggested above, but those have their own issues with a 6 1/2in driver in terms of centre-to-centre spacing, polar response etc., so presumably compromises like that weren't a priority when designing this unit.
OK, if we're on the 'I want breakup distributed' concept, let's look at this logically for a minute. This unit, like some others, obviously has the design objective of concentrating the main modes to a single well-defined peak (the secondaries don't start until almost an octave higher, which gives a good idea of how effective this cone profile must be), and act more or less as a pure piston below that. A single well-defined peak, especially at a reasonably high frequency, is a hell of a lot easier to deal with in a crossover than multiples, as well as resulting in lower overall distortion levels. Multiple distributed modes are certainly appropriate for a wideband driver like a Jordan, MA etc. -that's how they work. But that is not what this is for: it's a midbass with very different design method & priorities. Lars & Purifi are obviously aiming for a slightly less ambitious, but still quite broad BW over which the driver is more or less unresonant (i.e. 'pistonic') with very low HD and a stopband mode that is easily dealt with in the filter. True, that does partly preclude a low-order filter at a relatively high frequency, as was suggested above, but those have their own issues with a 6 1/2in driver in terms of centre-to-centre spacing, polar response etc., so presumably compromises like that weren't a priority when designing this unit.
Last edited:
What's wrong with a 1st order slope up to say 4kHz, together with a notch at 5k and notch at 10k? If the amplifier is not screwing things up then those resonances should just be a kind of 'falsetto' with no reason to suspect a loss of purity as such except for the radiation pattern.
If you're saying 1st order acoustic at 4KHz, then you'll only be about 4.5dB down at 5KHz, with a wide transition band, and with a bottomless notch placing a high Q hole in the response within this. If you simply EQ the response to the target curve, you won't stop the resonant amplified HD -you need a bottomless notch to do that. Granted, it's low as-is, but higher than it would & could be if you use the driver as Lars & co. appear to have in mind. On top of that, the polars would be lousy with the midbass narrowing significantly relative to the (presumably) tweeter with its much broader dispersion pattern through that BW and below. Usual message applies: on axis response only gives part of the story.
Right, a shunt would be a way to work around the peak, of course.
The product info contains the BL(x), Kms(x), and Le(x) performance, is there a chart somewhere that also shows the Le(i)?
Sorry, I didn't mean to say it wasn't an optimized profile, but it's still basically a conical, or mildly curvilinear cone, no? Its performance is excellent but I feel like there was a missed opportunity to do something revolutionary. It's the downside to coming out with a killer first album, the expectations for the second are too high... 🙂
Last edited:
I'd wager that most people who would want to "time align" their drivers will have the ability to measure the actual offset.
I played a bit with traced fdr and imp from the specs to see, and it's quiet hard from that data to cross higher than 2kHz with a second order passive crossover to avoid that resonances. I'm not using dsp myself for the moment, but i guess it will be similar. You will need high order crossovers to go higher than 2kHz, and once past 3kHz it will be impossible to tame that peak. But at that point dispertion is getting so poor it's not feasable anymore for such an expensive high end driver made for neutral sound to cross. Even from dispertion point of view 2kHz is more or less the limit. Pure as illustration of this I added my result (tweeter is SB26ADC, no calculations were made for alignment and so, it's a pure theoretical try out).
we do not measure Le(i) but you can get an idea from looking at the current distortion. The motor topology has inherently very little Le(i) modulation
applying notches: one has basically to tune the center frequency, Q and depth of the notch and any peak can be ironed flat. Of course, unit to unit variation and component tolerances sets a limit at the end of the day
YES! But if everyone listed their acoustic center ( AC) offset from the front of the flange, it would be most handy. All drivers. I have indeed found a couple mm differences in domes and guessing on a cone is almost useless.
I use an old GoldSound "Pulser" to measure mine, but my scope died and as it is a fixed and well defined parameter, it could be included, where things like the frequency response needs to be measured in situ.
My desire is for a 6 or so inch driver, that is reasonable to 60 or so, and be able to reach where most tweeters are happy, 2500 minimum, 3K better and get away with no more than second order electrical. I am not so greedy, or unaware of phase, to expect a first order electrical filter. My own testing concurs about with the published charts from Purify, that on the low end, distortion skyrockets at about 1/2 X-Max. ( Maybe the Purify does better 🙂 ) That applies to tweeters as well. So many look tempting to cross low as they get a pretty frequency response, but not good for distortion. These harmonics may be one of the causes for the breakup whose additional sub harmonics are causing the "glare" issue I am working on. That is why a LP on the tweeter only helps a little bit. Well behaved tweeter helps a little, and some yet to be determined behavior of DAC and amps effect it a little but less if a well behaved tweeter. Bad mixing effects it a lot, but nothing I can do about that.
Second it that magical 4 inch or so mid for an old fashioned 3-way. It looks like Purifi has a heck of a start on it.
What do you mean by "waste" ? In a treble unit, noone is noticing an acoustical lense made for difraction purpose is bad ? Id the passive "lense", "rod", whatever you call it is only taming what is around the break-up, why it could be worse than a serie parrallel notch as shown by Lribo which is very widthband large ?
While I like the Lribo, already imputed by HifiCompass, serie // notch. I use it myself in a treble unit but it has a large area of taming when the break ups is very focused most of the time (understand better DSP with firewall notches), no ?
I think an aluminum cone version of the 4" mid would be an extremely versatile unit.