https://purifi-audio.com/ptt6-5x04-naa-08/
I was thinking to myself all along that an aluminum cone would make the Purifi woofers even better. Problem is, it's saddled with the same issue that most other aluminum cones have - a relatively low breakup frequency. I'm sure HiFiCompass is going to do a test on these at some point, but I expect to see the high-order sub-harmonic HD amplification from that breakup manifest in a significant way. You already see it in the manufacturer's distortion curves for HD2 and HD3.
When I was reading the audioXpress article, I was saying out loud "please don't let it be a vanilla cone"...feels like the Purifi team just mailed it in, here. Oh well.
EDIT: Taking back some of the comments above. Comparing to other modern aluminum cone drivers, really the performance is not lacking in any way... still wish they had come out with some brand new technology though.
I was thinking to myself all along that an aluminum cone would make the Purifi woofers even better. Problem is, it's saddled with the same issue that most other aluminum cones have - a relatively low breakup frequency. I'm sure HiFiCompass is going to do a test on these at some point, but I expect to see the high-order sub-harmonic HD amplification from that breakup manifest in a significant way. You already see it in the manufacturer's distortion curves for HD2 and HD3.
When I was reading the audioXpress article, I was saying out loud "please don't let it be a vanilla cone"...feels like the Purifi team just mailed it in, here. Oh well.
EDIT: Taking back some of the comments above. Comparing to other modern aluminum cone drivers, really the performance is not lacking in any way... still wish they had come out with some brand new technology though.
Last edited by a moderator:
I seriously doubt the Purifi team just mails in anything. Even at break up, you're still down 50dB. Besides, you should not be anywhere near that break up crossing over from an 8" driver.
hi, we are preparing an app note showing how to do xover networks for the alu version that take the peaks away. measured through such network, the distortion peaks a gone including the hint of sub harmonic peak. cheers, Lars/PURIFI
Thanks Lars. Could you share any information regarding the decision to go with a standard (I assume) cone profile vs something with higher geometric rigidity to push the resonance up even further, so that the breakup doesn't need to be mitigated with additional passive components? SB Acoustics and Alpine have/had pressed ridges, KEF has the molded ribs on the back (poly cone), Fostex has their wacky HP shape (paper cone), YG Acoustics has their BilletCore technology, etc.
Figure 6 looks super tight even at 5kHz, but I'm guessing the 2.83V is nominal, hence the slightly elevated 30Hz range? (As compared to 30Hz in figure 5).
But almost everything else on the original driver was pretty ground-breaking. Granted, the paper cone didn't seem to be anything special.
The sub-harmonic HD amplification will hit at just over 1 kHz. HD6 and up usually aren't measured (and should be much lower in level), but if they exist in any meaningful way they will then show up at ~870 Hz, ~750 Hz, etc.
In many applications, you'd want to run this driver up to 1.5 kHz or higher, so I think you'd normally have to work around this effect.
Figure 6 is "current" harmonic distortion, meaning it's the distortion isolated to the motor itself. Figure 5 is for the whole driver.
IIRC, stamping out the breakup peaks can "fix" the HD profile.
The method I used was to create a "bottomless" notch filter on the primary breakup peak - LC (series) in parallel with the driver. By presenting a short-circuit to the driver at the desired frequency, strong electrical damping is applied, which reduces the acoustic output, but I think it also becomes more difficult to excite those resonances in the form of harmonic distortion.
Chris
The method I used was to create a "bottomless" notch filter on the primary breakup peak - LC (series) in parallel with the driver. By presenting a short-circuit to the driver at the desired frequency, strong electrical damping is applied, which reduces the acoustic output, but I think it also becomes more difficult to excite those resonances in the form of harmonic distortion.
Chris
there is nothing standard about the cone profile. it is the result of an automated optimisation process in Comsol/Matlab where 1000’s of profiles have been simulated. Experiments with ribs did not show much gain and this requires very slow 3D simulations whereas a reviled cone profile simulates much faster in 2D. This does of course not rule out that there could be fancy 3D profiles that work better. The main 5k break up involves me not the inner neck part and former so ribs would probably not help but admittedly look fancy
the subharmonic distortion peaks are due to motor distortion (hysteresis /current distortion) being amplified by the cone peak. However, our motor has extremely little distortion to begin with. Moreover, by adding a notch as an event in series with the driver then the current distortion is proportionally suppressed at the peak frequency thus killing the subharmonic peak. the THD peak at the peak frequency it self is also killed by the EQ. we come out with an app note about this
yes exactly. a shunt tank provides electrical damping but this increases the current distortion (increasing impedance reduces it). so better to use an LC parallel tank in series with the driver
Lars,
How can EQing one frequency affect the harmonic distortion at a different frequency? I don't see how this is possible.
Also, wouldn't the addition of the low-pass inductor provide a large impedance rise anyway?
I have some drivers with hard aluminium cones here, and will try some measurements of low-pass inductor + notch, with the notch being in series or parallel with the driver.
Chris
I've shown this before, but this is with a parallel notch in series with the driver (a modified Seas tweeter). Accuton used to provide before / after HD plots in some (some) of their old data sheets; they were being coy, but that's what they were doing. Series & parallel notches work, but parallel (in series) is generally my preference as it works better, unless you're incorporating, say, an LC shunt into the transfer function & it either just works out that way or you need to save money on components. Whether you'd call it EQ is a bit open to question as in both cases we're usually talking bottomless notches being used purely as that, rather than being used to shape the driver to a given response.
Attachments
Last edited:
Chris,
Let’s say the cone has a peak at 3k. this amplifies the 3rd harmonic of a 1kHz arising in the motor and the distortion plot shows a peak for HD3 at 1kHz. The dominant motor distortion above a few 100 Hz is the hysteresis distortion showing up as distortion of the current. This can be modelled as a dynamic modulation of the coil inductance. an inductor produces an EMf=L•i’(t) and the EMF induces a current in the coil i=EMF/Z where Z is the impedance seen by the coil. A notch or inductor between the amp and coil changes this impedance and can reduce the hysteresis distortion showing up as a peak at 1k. A shunt notch does not work since it lowers the Z .
🙂
cheers
Lars
I actually think that the specs and meausurement of this driver are quiet good, certainly compared to a lot of the other high praised metal cones who have way worse specs on resonance and distortion. But it will of course nevere be like a treated paper cone (that has other disadvantages). It's not that i'm in the market for such drivers, but if i was this one would certainly interest me. It can easely be crosssed at 2 to 2.5kHz, even with a lower order crossover in my opinion.
Based on the published data, I'd say it's more than 'quite good'; it looks a storming example of the type. It's not going to have a whole lot of character of its own, based on the FR, HD, which is generally what I want from a midbass -hopefully I'll get chance to work with it at some point this year. It might just be the resolution of the impedance plot, but I think I can see a tiny glitch at about 1.3KHz; I doubt in this case it's indicating energy storage though, so it seems more or less to be a pure piston out to ~5KHz. Great job again, Lars. 🙂
thank you!
The tiny glitch at 1.3kHz can be from our baffle or a resonance in the baffle
Usually, EQ is applied at line-level. ie, the impedance seen by the driver won't change - it's only the output voltage of the amplifier that's varying by frequency.
Lars, thanks. It's been a while since my degree was put to use, so I think it's time to brush up again.
Chris
What about a passive eq to kill the difraction aluminum cone with a solid lense in front of the driver as did Phil Jones with Boston Acojstic Lynnfield serie on its mid unit ?
I suspected it was something like that or a reflection, since the magnitude is so small & there wasn't anything notable in the HD. Thanks for confirming!