Anyone who cared who had them could do before and after acoustic measurements. No one in his thread has shown any interest.
Has anyone read this?
https://www.amazon.com/Audio-Production-Critical-Listening-Engineering/dp/0240812956
George
https://www.amazon.com/Audio-Production-Critical-Listening-Engineering/dp/0240812956
George
I struggle to see what is new that Joe is trying to explain in obtuse new language. I have given up that there is anything of note to discover as there is a new maths to be used that only Joe understands.
Have browsed the book, haven't listened to the examples. Introductory level ear training for recording engineers, is how I would describe it.
That's his USP, not unknown in the business, designers design new phrases and terms to make themselves and their products appear different and interesting, don't question them or they'll get upset and/or their guard dogs will attack. 😀
For those interested in reduced distortion in loudspeakers in case you missed it Exploring Purifi Woofer Speaker Builds might be of interest. It's a link to an interview with Purifi over some of the technology in their new drivers. Nice part is the interview has been bookmarked in that post if you just want to jump to the meat.
Now I will admit to not understanding significant chunks of what they have done in the motor system, but was intrigued that a significant area of distortion was found to be the surround.
Interesting stuff although of limited interest to those of you who have space for 10" or larger drive units and can get reduced distortion by traditional brute force methods 🙂
Now I will admit to not understanding significant chunks of what they have done in the motor system, but was intrigued that a significant area of distortion was found to be the surround.
Interesting stuff although of limited interest to those of you who have space for 10" or larger drive units and can get reduced distortion by traditional brute force methods 🙂
So you mean they are insulted? You mean, they are not imagining it? I get insulted since I don't actually sell any product. Where is my missing loot?
I have been looking at this driver: PTT6.5X08-NFA-01 (PTT6.5W08-01) - PURIFI
The surround problem often shows up in FR plots in even the most expensive drivers around 1KHz for a 6.5" driver and it just isn't there in the Purifi driver. It is caused by multiple of things both mechanical and acoustic.
PS: He says something very interesting and that plot... 1hr 29min 40 seconds. This is touching on the area that I am working on. Watch it.
I have been looking at this driver: PTT6.5X08-NFA-01 (PTT6.5W08-01) - PURIFI
The surround problem often shows up in FR plots in even the most expensive drivers around 1KHz for a 6.5" driver and it just isn't there in the Purifi driver. It is caused by multiple of things both mechanical and acoustic.
PS: He says something very interesting and that plot... 1hr 29min 40 seconds. This is touching on the area that I am working on. Watch it.
Last edited:
My tortured prose or the video? 😛 Would be interested in any opinion you have on the motor design and their measuring setup.
Or if the video is too hard on you then https://purifi-audio.com/wp-content/uploads/2019/12/PURIFI-AES-9607.pdf are they onto something dusting off some 1940s reasearch into force factor modulation?
Er no. Not even close.
The video was painful.
The paper is easier. This "force factor modulation" was known long before Cunningham in 49. It's bog standard from AC rotating machinery... the IdL/dt term I have been speaking of.(edit, fixed it from LdI/dt).
I have also been stressing that an AC shorting ring can only flatten inductance in a range of frequencies, they at least show the different frequencies plotted vs position in figure 2, the blocked vc tests. What they do not do is look into how a shorting ring lowers inductance in a different method from in front of the front plate. Going forward, L falls as a consequence of reluctance increase, the shorting ring lowers it by eddy dissipation. It was odd that they suddenly realized that a shorting ring can flip the sign (section 3.4), what did they think it would do?
They also should have repeated figure 2 but instead of a single tone, superimpose a hf on a lower tone being used to dynamically alter the vc position. I do not know what their positioning system bandwidth was, but they should have been capable of maybe 200 hz.I suspect they cannot measure the hf inductance due to the lf confounders however. Nor have they even considered the velocity dependence of the inductance.
In figure 5, they show D.C. Displacement for tone bursts. They do not consider how they gate the tone burst, if the burst always starts positive going, the D.C. Displacement will always be in the same direction, they should have tried inverting the bursts to verify. This is a common problem in x-ray undulators, the initial magnetic lobe can offset the e-beam so the first half period is usually of smaller field. (Remember, the tone burst signal is not a position signal, it is an acceleration signal).
In figure 7 they show 70 hz bursts, note the displacement is opposite direction to the 7khz and 400 hz. This is indeed consistent with their analysis, as examination of fig 2 shows 70hz and 7khz have opposite slopes at zero displacement. However, 70hz is also within the cabinet domain.
In the conclusion, they say this can be generalized to all motor structures. Well, yes, it was done so long before Cunningham.
Jn
Ps. It's not a bad start, but they have a long way to go in terms of understanding and test methodology.
Last edited:
So does that mean that Loudspeaker motors have caught up with where other things were 70 years ago yet, or at least a step in the right direction?
A step.
I did chuckle when they mentioned Keyence. We use their equipment to measure undulator gap positions. While I can position repeatably to under 5 nanometers, that's encoder position. Try mounting an encoder to sub micron accuracy, never mind in the 10 nanometer range. We use Renishaw absolutes, and have to establish an offset by measurement. And then try to keep the temperature rock steady. While the scales are invar, the entire structure is a combo of aluminum, iron, and neodymium.
Jn
I did chuckle when they mentioned Keyence. We use their equipment to measure undulator gap positions. While I can position repeatably to under 5 nanometers, that's encoder position. Try mounting an encoder to sub micron accuracy, never mind in the 10 nanometer range. We use Renishaw absolutes, and have to establish an offset by measurement. And then try to keep the temperature rock steady. While the scales are invar, the entire structure is a combo of aluminum, iron, and neodymium.
Jn
Last edited:
Oh, forgot to mention. In 3.2, they mention a mechanical resonance at 4khz affected the 2khz data. But then a 6th order polynomial fit was used to match that data in 3.3.
Above resonance, the load decouples from the support structure. All data taken above resonance has to be reconsidered in light of that, as well as data approaching resonance from below. When I do inductance frequency scans, I actually extend the measurement frequency at least an order of magnitude beyond requirements to look for data corruption as a result of resonance. I throw out data if a resonance is nearby, and use only the well behaved data.
Fix the test problem, don't include it in the modeling..who does that?
With a system resonance built into the math, it's no surprise that the force factor modulation transfer function goes all wacky (fig 3) at 3khz and above, the phase shift data shows consistency with bog standard speaker phase plots through resonance.
They could at least redone the blocked measurements with a slit front plate to check plate eddy current effects, that would have been a no-brainer and an advancement.
Jn
Above resonance, the load decouples from the support structure. All data taken above resonance has to be reconsidered in light of that, as well as data approaching resonance from below. When I do inductance frequency scans, I actually extend the measurement frequency at least an order of magnitude beyond requirements to look for data corruption as a result of resonance. I throw out data if a resonance is nearby, and use only the well behaved data.
Fix the test problem, don't include it in the modeling..who does that?
With a system resonance built into the math, it's no surprise that the force factor modulation transfer function goes all wacky (fig 3) at 3khz and above, the phase shift data shows consistency with bog standard speaker phase plots through resonance.
They could at least redone the blocked measurements with a slit front plate to check plate eddy current effects, that would have been a no-brainer and an advancement.
Jn
Last edited: