I was starting on a new speaker design using a Fostex woofer from their NX6A powered monitor. This 6.5in driver has a unique pinched cone and wrinkled surround, similar to the Purifi woofers. I got the bare drivers about 3 years ago for a blowout sale ($15 IIRC). Since these are not currently in production anymore, this speaker design would be fully open source and was being made for my personal DIY pet project (not commercial). I started to describe it here. Anyhow, after installing the driver in suitable cabinet that I had lying around, I proceeded to make some measurements to collect the FRD files needed to simulate the XO. The speaker looks like this. Tthe cabinet was an early prototype for another woofer and tweeter so none of the cutouts had rebates or cutouts that fit the NX6A woofer and a Dayton AMT2-4 tweeter that I had planned on using. So, I had to make some 3D printed adapters for the tweeter and for the woofer, I settled for surface (non-flush) mounting and will live with it.
So after putting the speaker with drivers installed and I took the initial sweep, I was shocked at the high level of distortion from 150Hz to 4khz, especially the 3rd and 5th harmonic distortion (note that distortion levels in legend refer to distortion at cursor location - which I am not sure if I kept constant - I would go by the vertical dB axis value though, you can see distortion peak at 64dB :
I spent the next hour trying to see if there was a setup issue, or something wrong with the electronics, wires, frame rattle, etc. All to no avail. But what I did notice was that the level of distortion was slowly going down with each sweep. I was playing with removing the vent tube, plugging the tube with stuffing, or leaving the tube in, thinking it might have something to do with it. Here are subsequent sweeps and you can see each one, the level of THD goes down a bit. the sweeps are chirped sound lasting about 5 seconds.
Number 2 (vent tube removed):
Number 3 (with vent tube back in):
Seeing the THD go down a bit, I tried running 45Hz at about 2.8vrms for 1 min to see if this loosened up the suspension: And it sure did!
So I tried leaving the 45Hz on for 60 minutes - distortion at 300Hz went down to about 34dB an apprximate 30 dB improvement!
I wanted to be sure it was not caused by the vent or lack of one or by plugging the vent, so I took a few more sweeps in different configs and now the distortion remained low:
Here are all the responses overlaid and you can see the effect of the vent/no vent/stuffed vent:
So I have left the speaker on playing music continuously for 3 days now, hoping that maybe I can eek out a bit more improvement. I have not checked again yet, but it seems the majority of break-in for this deiver occurred in the first hour playing some bass to really move the cone.
Here is the Wago plank crossover that I ended up with:
Measured response:
Overall distortion at 2.0Vrms and microphone 0.5m away:
Measured phase variation:
Pretty clean now. It sounds quite nice. But I think I have seen it with my own eyes and instrumentation - the real effect of driver suspension break-in on harmonic distortion. Has anyone else measured something similar? You can listen to the current speaker here.
So after putting the speaker with drivers installed and I took the initial sweep, I was shocked at the high level of distortion from 150Hz to 4khz, especially the 3rd and 5th harmonic distortion (note that distortion levels in legend refer to distortion at cursor location - which I am not sure if I kept constant - I would go by the vertical dB axis value though, you can see distortion peak at 64dB :
I spent the next hour trying to see if there was a setup issue, or something wrong with the electronics, wires, frame rattle, etc. All to no avail. But what I did notice was that the level of distortion was slowly going down with each sweep. I was playing with removing the vent tube, plugging the tube with stuffing, or leaving the tube in, thinking it might have something to do with it. Here are subsequent sweeps and you can see each one, the level of THD goes down a bit. the sweeps are chirped sound lasting about 5 seconds.
Number 2 (vent tube removed):
Number 3 (with vent tube back in):
Seeing the THD go down a bit, I tried running 45Hz at about 2.8vrms for 1 min to see if this loosened up the suspension: And it sure did!
So I tried leaving the 45Hz on for 60 minutes - distortion at 300Hz went down to about 34dB an apprximate 30 dB improvement!
I wanted to be sure it was not caused by the vent or lack of one or by plugging the vent, so I took a few more sweeps in different configs and now the distortion remained low:
Here are all the responses overlaid and you can see the effect of the vent/no vent/stuffed vent:
So I have left the speaker on playing music continuously for 3 days now, hoping that maybe I can eek out a bit more improvement. I have not checked again yet, but it seems the majority of break-in for this deiver occurred in the first hour playing some bass to really move the cone.
Here is the Wago plank crossover that I ended up with:
Measured response:
Overall distortion at 2.0Vrms and microphone 0.5m away:
Measured phase variation:
Pretty clean now. It sounds quite nice. But I think I have seen it with my own eyes and instrumentation - the real effect of driver suspension break-in on harmonic distortion. Has anyone else measured something similar? You can listen to the current speaker here.
Last edited:
A few seconds at 450 Hz would probably have done the trick too. It looks like the break-in is basically done after a minute at 45 Hz. It would have been interesting to see 2,5, and 10 minutes but that's a bit late now.
That's pretty consistent with what I've heard from speaker designers who measure. Run the driver for a few seconds and it's broken in.
Tom
That's pretty consistent with what I've heard from speaker designers who measure. Run the driver for a few seconds and it's broken in.
Tom
I know I have heard a difference before and it sometimes takes days, but I never did a deep stroke bass exercise like this. Good to hear that others have also done this with new drivers. Makes sense that an extensively tested driver is naturally broken in.
For large woofers maybe it takes longer. I think pro audio pleated cloth surrounds take longer.
For large woofers maybe it takes longer. I think pro audio pleated cloth surrounds take longer.
You are saying that the suspension stiffened up when not used for 2 days? Almost like a memory effect. What kind of suspension was it? Doped fabric, rubber, or foam? Some old school suspensions even used leather.
Moat of this is because the box and driver were not designed for each other - I grabbed whatever loose spares I had lying around that were close enough to make a speaker. The cabinet is about 10.4L and the Vas on this driver is 8L, so cabinet is a little too large and the vent is 2in x 6in (default one that came with the cabinet). I can live with it though as I will probably use it with a subwoofer.
But preceeded them by decades. Saw this woofer first at AES 1999 in NYNY. A really clever design.
Geddes has shown that the measured distortion needs to hit something like 25% to correlate with the sonics, but it is really good to see objective measures showing changes with break-in… the timimgof this woofer means there is a chance it was designed by Tanaguchi-san, who has said he has done at least one design where the spider needed 1500 hrs to break in.
dave
Last edited:
Yes, and I do not think that this is an exception. I tested B&C 10HPL51.
Whenever I measured fc of a driver over the decades now, it was always significantly higher than specified.
Could be interesting to do more long-term testing of fc.
I have found Dayton drivers to be close to stated specs regarding Fs, Qts, Vas. No exaggeration. The Fs does go down a few Hz after break in but only off by maybe 3-5 Hz. Other brands (even the esteemed Danish ones) I have consistently found Qts to be optimistically stated lower than what is measured. That’s a value that won’t get lower with age or break-in.
I totally forgot to mention that Fostex did this a long time ago. Purifi did it only recently but they did get a flat as a board frequency response and very low distortion over the entire operating band. But for the price, one can get four conventional woofers that are pretty nice and would work a whole lot less hard. Size wise, won’t be as compact though of course.
Similarly, Dan Wiggins developed the XBL motor tech with linear BL and long 10mm xmax. Very low distortion motor and this was done a long time ago as well.
