Some Sadhara Drivers

[ChocoHolic]

Hi Guys!

I' d like to get some opinions (or even better experiences, if available) about my following observations.
Some days ago I received two SADHARA drivers and noticed that their resonance frequencies were different.
I ran them for about 100 hours at 20Hz/0.5A in order to get stabilized parameters.
No change.
Driver A shows 23Hz. Driver B shows 28Hz. (Data sheet 26Hz).
The reason for this is a difference in the stiffness of the suspension.
Driver A shows 2.1 N/mm. Driver B shows 3 N/mm.
Considering that the resonance frequency is going with the root
of the stiffness (equal moving mass assumed), these results match very good to the observed difference in their resonance frequencies.
But I wonder if this remarkable difference is normal... ?

Furtheron I checked the force of both motors depending on the current.
Both speakers showed similar values around 10 N/A.
(A: 9.1 N/A , B: 10 N/A)
I was not able to measure very accurate, but the expected 13N/A seem definitely not be here...
(WinISD alpha calculated 13N/A from the values in the data sheet)

And of course I looked at the possible excursion:
For this I applied a 15Hz sine wave, speaker in free air condition.
A:
+/- 10.5mm at 20Vpeak/15Hz
+/- 20.5mm at 40V peak/15Hz
+/- 25mm at 48V peak/15Hz (limit for bottoming)
B:
+/- 9mm at 20Vpeak/15Hz
+/- 18mm at 40Vpeak/15Hz
+/- 25mm at 60Vpeak/15Hz (still margin against bottoming)
Bottoming at about 75Vpeak and +/-28mm

So also the excursion measurements indicate a different suspension.

Well, all together the drivers seem to be not that bad.
I will run them in a small sealed box (equalized by LT).
The small box will dominate the stiffness of the suspension anyway
by the higher stiffness of the small air volume.
Up to somewhere above +/-20mm, both drivers seem to be
more or less linear...
I did not check all Thiele Small parameters, but I would expect different Qms.
On the other hand Qes will be similar as the motors are matching quite good. And Qes dominates over Qms in these drviers, so for Qts can be expected similar values of both speakers ... I hope.....

But again: Even I could live somehow with all this...
Are such differences within normal limits?
I would have expected better matching of both drivers, some difference against the data would not make me wondering. But such differences between two drivers in one delivery???


Cheers
Markus

[markp]

For a 'quality' driver I would expect much better matching between the two. I'd call them up and see what the deal is.

[DanWiggins]

Hi all...

Suspension stiffness is the least controlled parameter of a driver. Spiders are typically spec'd with 20% tolerance; both drivers are within tolerance, if you can believe it. The reason is that spiders are essentially cut pieces of cloth, permanent pressed, then dipped in plastics. Lots of variables in the system, so that even spiders from the same production batch can vary by 20% without much you can do.

More important, especially to see how they work in-box, is that the SPLs of the drivers are close (really - this is a good "overall" check parameter). When the spider is stiffer, you'll find Fs higher, and Qes higher as well, with the net result being the low end F3 and response is extremely close to a looser spider version.

This is one of the dirty little secrets of speaker building. Apparent big changes in drivers all come from production tolerances of spiders, but when all is said and done, these changes are actually complementary in a given implementation.

For the BL, did you directly measure the DC resistance of the voice coils, and use that for calculation of BL? Also, many programs use the Sd to indirectly assist in calculation of BL (from compliance, which uses Sd), so what did you use for the Sd?

Thanks,

Dan Wiggins
Adire Audio

[ChocoHolic]

Hi Dan,
thanks for your kind support.
Interesting to know that 20% within one batch of spiders is
normal. With this I can assume that over 40% difference,
which I am facing, are also normal if they are not from the same
batch... Right?

I did very simple checks only.
Of course not very sophisticated, but better than nothing.

First I measured the stiffness of the suspension by putting
about 3.3kg weight to the cone and found that one cone
moved 11mm and the other one moved 16mm. So I got the
2.1N/mm and 3N/mm.

For BL measurement, I applied 1.5A DC and found around 5mm and
6.5mm excursion. These excursions correlate to 13.7N and 15N, if we consider the stiffnesses which I measured first....
I would have appreciated to measure with higher current and higher excursion, because my mechanical excursion measurement is quite poor
But I had no suitable power supply by hand...
By the way: How much DC would be the allowed limit for such
a measurement, if I consider 20s to do the mechanical measurement?

For the programm I did not measure anything,
I simply entered the values, which you offer on your WEB Site
to that software "WinISD alpha" and somehow it also calculated
BL. Sd was 436 cm^2.
I did some simulations with this software when I started to look around fro a suitable driver some months back.....

Bye
Markus

P.S:
What's the should-value for BL of a SADHARA?

[Paradise_Ice]

Hi ChocoHolic
Yes I think such differences are normal,
It is very hard to make a perfect matching drive unit,
Did I say hard I meant almost impossible!
There will always be subtle difference and your measurements have very little difference between them to be honest,
And besides our ears have a large difference between them,
And am not talking about the brain,
I understand that we want all things to be equal in a perfect world, but this is not that world, this is Earth.
Unless the people who make the woofer take the greatest care in getting very tightest tolerance for each component then and even then they will not be a 100% matching pair, more like 98% to 99% matching at best, I am happy to get a 5% variance between each driver we make, but can get 2%
Kef I am sure could get 0.5dB in there B101 I think?
Even then that’s a sound pressure measurement,
It’s not a T&S measurement as you have performed.
Nobody am sure could say the have a perfect matching set of speakers and remember that speakers deal with air pressure that is a variable also, it is mechanical but not in the same way a Engines pistons are mechanical, the finesse and detail we can hear is astounding, I hope this puts you at easy.

[ChocoHolic]

Hi Ice,
seems like I simply fetched a "unlucky" couple of drivers, within normal tolerances.
In fact all other woofers which I had earlier (i.e. Dynaudio or EV, but also low cost types), showed nearly perfect matching, so I am just not used to such tolerances.
On the other hand I am also convinced that the resulting resonance frequency of the sealed box, will be dominated by the stiffness of the air, because I want to go for a small sealed design somewhere in the range between 20l...30l.
So I don't care to much about the difference in the stiffness of the speaker's spiders.


@Dan:
By the way: What's your opinion about the smallest suitable sealed
box for a Sadhara if I use a LT for equalization? (Required power of the amp is no issue.)
I plan to equalize down to 12Hz. Below 10 Hz I plan to cut off by 6th order / 1db ripple / Tschebycheff. In first step I plan this in OP amp technology. May be I will learn in some years to substitute that lot of analog filters with a DSP.....


Bye
Markus