The difference between factory numbers & DATS are likely due to context and completely consistent. I have learned to use factory numbers 1st when modeling a box and use my local measures (Smith & Larson WT2) to match drivers.
Here is my “editorial” on the subject: http://www.diyaudio.com/forums/soft...rements-book-win-mac-linux-4.html#post4657538
dave
Here is my “editorial” on the subject: http://www.diyaudio.com/forums/soft...rements-book-win-mac-linux-4.html#post4657538
dave
I'm not sure it will help, but the document mentions break-in of about 10 minutes at *maximum* displacement.
This is NOT linear displacement.
They specifically mention just below CLIPPING for the driver.
..basically this means pushing it until you start to hear slight clipping effects and then back-off from that.
It currently "sounds" as if the driver's surround resistance (from inadequate break-in) is factoring-into the mms of the driver, giving a false higher mms value.
-just a guess.
(..also, don't forget to measure the diver in a *vertical* free-air condition and of course wait for the VC to cool properly after break-in.)
..basically this means pushing it until you start to hear slight clipping effects and then back-off from that.
That's exactly what I did. Increase the level until they were rattling, then back off a bit. The voice coil was then moving approx 8mm pk-pk, and for at least an hour. As far as I know, SB measure their driver responses and TS parameters at 2.83V, so they should be linear and the results from DATS should be valid. I agree the parameters will change at higher level, but then you can get any answer you like.
There was absolutely no change in the parameters before and after break-in, so I'm sure that further break-in would be futile.
That's exactly what I did. Increase the level until they were rattling, then back off a bit. The voice coil was then moving approx 8mm pk-pk, and for at least an hour. As far as I know, SB measure their driver responses and TS parameters at 2.83V, so they should be linear and the results from DATS should be valid. I agree the parameters will change at higher level, but then you can get any answer you like.
There was absolutely no change in the parameters before and after break-in, so I'm sure that further break-in would be futile.
Last edited:
The issue that you describe, I've found it's true of every relatively large driver that I've measured using DATS.
According to what people told me on here, the issue is because the signal level of the DATS is so low. IIRC, it's about one watt. You need a higher voltage to get a reliable measurement of a driver.
For instance, if I measure a 4" midrange using DATS, the T/S is close to spec.
If I measure an 8" woofer, the QTS is quite a bit higher than expected.
If I measure a 12" subwoofer, the QTS will be off by as much as 100%.
According to what people told me on here, the issue is because the signal level of the DATS is so low. IIRC, it's about one watt. You need a higher voltage to get a reliable measurement of a driver.
For instance, if I measure a 4" midrange using DATS, the T/S is close to spec.
If I measure an 8" woofer, the QTS is quite a bit higher than expected.
If I measure a 12" subwoofer, the QTS will be off by as much as 100%.
But 1W = 2.83V into 8 ohms, so the level used by DATS is correct - it should match SB's parameters. Driving the woofer harder will just make it non-linear and the fs and Qts will be pushed down, giving a false result.
However, I do agree that many drivers do not meet their TS parameter specs, although it tends to be the Cms which is too low (too stiff), whereas the Mms is normally about right.
However, I do agree that many drivers do not meet their TS parameter specs, although it tends to be the Cms which is too low (too stiff), whereas the Mms is normally about right.
The SB17NBAC35-8 does actually exist on the SB Acoustics website which I found after a second search that's why I corrected my initial post.
SB Acoustics :: 6" SB17NBAC35-8<br><font color="#C70039">SHIPPING</font>
C.M
SB Acoustics :: 6" SB17NBAC35-8<br><font color="#C70039">SHIPPING</font>
C.M
I got four SB drivers work well, sound fine, look good, seem well made ...but the Qts measures high with DATS.
Mine were 12" (SB34NRX75-6)
listed as Qts 0.4 & Fs 19Hz on their website
The Qts I measured:
0.6015
0.5495
0.5729
0.5795
Fs ~25Hz
One driver got experimented with:
Qts 0.58 (start)
Qts 0.52 (after about 20 hours of break in)
Qts 0.47 (with additional magnet)
My break in was much gentler than what the OP in this thread describes.
I have noticed a similar trend. It could be because:
a) DATS is unreliable as driver size increases
b) Manufacturers are unreliable (use cheaper magnet materials) as driver size increases
c) a bit of both
I've measured some small drivers that were very on spec. The biggest thing that was a perfect match to the specs was a pair of Peerless 6.5"
Most of the larger drivers I've tested have been higher Qts than advertised.
However, most of the larger drivers I've tested have been made in Asia.
The outlier: a pair of old Electro-Voice woofers from 'murica. These have very low Qts / high efficiency, and these are the biggest drivers I've owned (18")
Mine were 12" (SB34NRX75-6)
listed as Qts 0.4 & Fs 19Hz on their website
The Qts I measured:
0.6015
0.5495
0.5729
0.5795
Fs ~25Hz
One driver got experimented with:
Qts 0.58 (start)
Qts 0.52 (after about 20 hours of break in)
Qts 0.47 (with additional magnet)
My break in was much gentler than what the OP in this thread describes.
USB power supply is only 5V, so the DATS signal must be low-ish. About 2 watts is the most it could deliver. I should actually measure it and see.
I have noticed a similar trend. It could be because:
a) DATS is unreliable as driver size increases
b) Manufacturers are unreliable (use cheaper magnet materials) as driver size increases
c) a bit of both
I've measured some small drivers that were very on spec. The biggest thing that was a perfect match to the specs was a pair of Peerless 6.5"
Most of the larger drivers I've tested have been higher Qts than advertised.
However, most of the larger drivers I've tested have been made in Asia.
The outlier: a pair of old Electro-Voice woofers from 'murica. These have very low Qts / high efficiency, and these are the biggest drivers I've owned (18")
Last edited:
DATS doesn't measure the impedance at 10Hz; it does a sweep from 20Hz to 20kHz. The 10Hz sinewave is just used for break-in, as it is well below resonance and so the cone excursion is maximised for a given input power.
The drivers are SB17NBAC35-8 and not SB17NRXC35-8.
They are quite new metal coned models, whereas the NRXC is an older paper coned design.
The output from the DATS is in differential form, so it should easily be able to produce 2.83Vrms (1W/8ohms). I did actually try measuring the level, but the sweep is too fast for my DVM to respond. I'll dig out my old 'scope and try that.
Looking at some of Zaph's measurements, he also gets very high values for the Qts of SB drivers. Compared to other brands, his measurements suggest that in terms of TS parameter accuracy, SB are among the worst, Peerless are the best, and SEAS and Scanspeak somewhere in between.
OK that's not too bad. Both the NRXC and NBAC are supposed to have Qts ~0.35, so 0.46 should be usable.
When experimenting with the MJK worksheet for my (MLTL) cabinets, I found that provided Qts was below 0.5 I could still tune the port successfully. But my drivers are Qts=0.55, and I couldn't get a flat low end response (except by using a sealed box).
There are tolerances on magnets, and spiders, but mass is mass, and if the manufacturer states that mass is about 12g and you measure an outstanding 19g, it's obvious that i prefer to trust the datasheet than the measurement...
DATS might be ok for impedance measurements, but my chinese bazar next door sells much better precision kitchen scales...
Last edited:
Definitely no, manufacturers do not change moving coils and cone materials allowing mass increases of almost 100%!
A loudspeaker is much more than a pitta mass spring system, and the overall response ( linear and non linear distortion) is far more important than small details like exact Fs , Qts or Vas knowledge. For this reason mass have to be mass, and if it is not, the driver will be another totally different driver.
Btw, the "appropriate" signal level is an interesting concept, can you elaborate on this?
PD: Same might apply to "appropiate" break in process, "appropriate" added mass, "appropiate" measurement signal and generator, etc...
My 2 cents, build the bloody box according to the damned manufacturer spec sheet and tweak if necessary. Theoretical discussions on TS parameters and their measurement are a dreadfull waste of time!
Last edited: