How does the AMT2-4 compare to the PT2522? I happen to have them both and know that you have used them both.
It would be interesting to watch an FFT while playing to watch the harmonics fall as break-in takes place. It could be recorded. The test signal is a question though, you need a low frequency to approach Xmax but the most distortion occurs at 300Hz.
It is also interesting to me that the distortion is almost an inverse of the woofer impedance curve, as though it's current or force dependent (at 300Hz this would be compression at the former cone interface). I wonder if the funny shaped surround has some bistable behavior, possibly intentional in order to flatten the suspension compliance curve? Or maybe there is some residual magnetization in parts of the motor structure that gets worked out by low frequencies? Highly odd dominant distortion is often related to magnetics or weak connections to steel parts.
That woofer really deserves to be flush mounted what with the low-diffraction surround.
It is also interesting to me that the distortion is almost an inverse of the woofer impedance curve, as though it's current or force dependent (at 300Hz this would be compression at the former cone interface). I wonder if the funny shaped surround has some bistable behavior, possibly intentional in order to flatten the suspension compliance curve? Or maybe there is some residual magnetization in parts of the motor structure that gets worked out by low frequencies? Highly odd dominant distortion is often related to magnetics or weak connections to steel parts.
That woofer really deserves to be flush mounted what with the low-diffraction surround.
Last edited:
What I meant is: I presume LF measurement is nearfield, then merged with a 1 meter measurement above some 400 Hz?
I've certainly seen FS drop on new drivers and go back up a little bit after the break-in cycle. I've put that down to temperature. I've not seen FS rise over years, quite the opposite I tested some W12CY001 after maybe 5 or 10 years and FS was lower than spec.
This certainly fits with my experience with an Infinity car woofer that I finally installed after having sitting in my garage for >5 years. I figured the spider had really stiffened up and it was so bad initially that the driver would "squawk" depending on the bass excursion. Within a couple hours of use the sound disappeared, and without measuring I assume the driver finished softening up soon after. I have only installed one of the pair so I'm hoping to reproduce the effect when I install the other one just to prove the point. Maybe I'll put it on a baffle and record the behavior outside the car just so I can be certain.
I can do real time FFT using the RTA mode. Excite it with pink noise and record RTA. Or excite at 350Hz and monitor peaks vs time.
I still have one more new (non-broken in) driver.
It would be interesting to compare the "broken in" speaker to the non-broken in to see if the resting position of the cone is the same.
As the speakers were unused for three years (and may have been for years prior to purchase) if stored horizontally, the suspension may have sagged.
If so, the up or down (in or out..) offset would increase distortion caused by both Cms and Bl nonlinearities.
If the suspension offset returned to normal after some use, distortion would decrease.
The distortion was entirely odd harmonics, so a resting position shift does not explain the distortion as that would cause even harmonics. This points to electrical or mechanical soft hysteresis. If it were the spider, I would expect the distortion to be highest at Fs. There are no high order harmonics like there would be if it were a rub/buzz. The distortion is highest at 300Hz where the force on the cone and the voicecoil flux are highest. So I think the most likely explanation is either magnetic, or a break-in behavior of the cone.
Maybe the magnet has some complex properties that result in magnetic hysteresis until the voicecoil flux works it through a range. Maybe the funny shaped cone requires a special gluing job that has hysteresis until worked. Maybe the inner ring of the spider, which experiences more stress than the outer ring as frequency increases, has it's own break-in behavior separate from the rest of the spider.
I only have crazy conspiracy theories. But the common element here is hysteresis which goes away with excursion. To rule out magnetic hysteresis, you can work the suspension back and forth with your fingers and then check if the distortion has changed. To rule out mechanical hysteresis, you could lock the cone in place and run a ~30Hz tone for a while (at impedance minimum 20-60Hz). Mechanical hysteresis would respond best to exursion at Fs. Magnetic hysteresis would respond best to a signal well below Fs that is not at an impedance peak.
Maybe the magnet has some complex properties that result in magnetic hysteresis until the voicecoil flux works it through a range. Maybe the funny shaped cone requires a special gluing job that has hysteresis until worked. Maybe the inner ring of the spider, which experiences more stress than the outer ring as frequency increases, has it's own break-in behavior separate from the rest of the spider.
I only have crazy conspiracy theories. But the common element here is hysteresis which goes away with excursion. To rule out magnetic hysteresis, you can work the suspension back and forth with your fingers and then check if the distortion has changed. To rule out mechanical hysteresis, you could lock the cone in place and run a ~30Hz tone for a while (at impedance minimum 20-60Hz). Mechanical hysteresis would respond best to exursion at Fs. Magnetic hysteresis would respond best to a signal well below Fs that is not at an impedance peak.
Here is a study with break-in of 0, 1, 7, 13 and 28 hours break-in.
https://techtalk.parts-express.com/...changes-during-break-in?p=1474069#post1474069
https://techtalk.parts-express.com/...changes-during-break-in?p=1474069#post1474069
Mms obviously does not change with break-in, so in those tests the model should have changed Vas instead of Mms to account for Fs shift. The change in Le could either be due to the coil resting position changing with break in, or from the change in cone position due to the added mass method when Cms changes. It would be interesting to repeat these tests with better methodology.
I assembled the second speaker and made some measurements after various steps in time to see if the effect can be repeated. Here is the setup - for the woofer measurements, I went direct drive and bypassed the crossover.
Here is the baseline distortion sweep before any break-in, similar hump of distortion like the other driver:
The 30 seconds of 45Hz at 4.3Vrms applied:
After 60 seconds more and we see significant changes already:
Then 30 minutes of 45Hz at 4.3Vrms:
Now after a total of 50 minutes of 4.3Vrms:
I think we have reached steady state.
Here was the FFT RTA at the beginning with 247Hz excitation, H3 was 1.02%:
After 50min of 45Hz, the RTA at 247Hz excitation shows H3 at 0.28%, a significant reduction:
Here is the baseline distortion sweep before any break-in, similar hump of distortion like the other driver:
The 30 seconds of 45Hz at 4.3Vrms applied:
After 60 seconds more and we see significant changes already:
Then 30 minutes of 45Hz at 4.3Vrms:
Now after a total of 50 minutes of 4.3Vrms:
I think we have reached steady state.
Here was the FFT RTA at the beginning with 247Hz excitation, H3 was 1.02%:
After 50min of 45Hz, the RTA at 247Hz excitation shows H3 at 0.28%, a significant reduction:
My take is that the spider was "crunchy". The adhesive/damping compound on the spider likely became very stiff during the long storage period. When the the VC moved away from the rest position, individual fibers would break free, creating noise. I usually mechanically exercise the cones of woofers and mids prior to measurement. I have actually heard noise pushing the cone back and forth, especially on drivers stored for a long time. a kind of crackling noise. Just a guess, because I have not measured it with anything other than a DATs.
I heard that spiders begin with a solid coating which develops microcracks throughout as it breaks in, and the self-healing tendency of these cracks is responsible for the damping characteristics of the spider as well as it's tendency to stiffen when not in use.
I like the “crunchy spider” theory. I did not check specifically for “cooling down” effect but I doubt that’s it. I’ll test again today to see if it’s t remains low.
They are different since one is planar magnetic (pushes air like a regular membrane) and other is AMT (squeezes air). Although I think the AMT sounds clearer and more like a real organic sound. But GRS has ability to be open back.
I have the original big Heil AMT tweeter and I have to say, that is one of the best sounding tweeters ever. To me, it sounds better than even a ribbon.