advice: measured vs. manu' what to expect :vifa M13SG-09-16

[jquin]

Hi All
I built a MTM center from two M13SG-09-16 using someone elses design for M13SG-09-08, which is meant to be almost identical.
However it had massive problems with distortion.

So I decided to start from scratch, built a Wallin Jig and used speaker workshop. First results were noisy so I upgraded soundcard and the result were awsome, textbook impedence curves on both drivers.

Problem T/S parameters were way off data sheet be quite consistant between drivers.

Results
Driver A_____Driver B____Manufacturers
Qts=0.72____0.70______0.35
Qms=2.2____2.0_______1.42
Qes=1.1_____1.1_______0.64
Vas=3.7L____3.5L______12 L
Fs=99Hz ____107Hz_____60Hz

Now these drivers are 16Ohms are and supposed to be driven in parallel in an MTM. So are they supposed to be tested in parallel?

The drivers have had the dust caps pushed in by my children which I pulled out again with glue. I removed some of the doping from the dust cap with solvent while removing excess gule. Not sure if this would have altered the T/S much.

Any Ideas?
Are Manufacturers T/S parameters usually this way off?

Regards
John Q

[madinoz]

Yes, and no.
VIfa has a history of overstating the resonant frequency (i.e showing it as lower than it really is) and overstating the VAS (it is usually measured as a good bit smaller than it really is). Peerless is not so bad. Scanspeak are usually pretty on the money (but still a little bit off).
Having said all that the XLS series of Peerless subs have spec' VERY close to the stated ones so it can vary from series to series from one manufacturer. This is only based on my direct experience with these manufacturers.
Does that help?

[Dave Bullet]

Testing in parallel? - no point. Measuring in parallel for an MTM design in SW? - yes.

Did you use delta mass or compliance for measuring Vas. Getting either volume or more likely added mass wrong will lead to wild Vas estimation. I presume your impedance curves and peaks were without "noise" since you said you had near textbook curves. Inaccuracy of an impedance sweep (environmental noise, cone reflections etc..) will lead to inaccurate T/S params.

If the existing -08 designed crossover was for an MTM - then 2 16 ohm drivers will not make 8 ohms and therefore any sense.

When you said you believed distortion was bad - this has nothing to do with how close T/S params are, but either a driver operating out of its intended operating range, an inferior driver or scraping or some other mechanical problem. However, the impedance curve would indicate mechanical issues.

What does your amplitude response look like? (FR measurement).

David.

[djarchow]

Edit - I see David posted while I was writing my response so there is some overlap here.

Did you break in the drivers before measuring them? While I wouldn't expect to see the fs drop by 40 hz, I would expect it to drop some after break in. I looked on Madisound and they post different t/s parameters for this driver. with fs aound 52 hz, qts - .432, qms - 1.535, and qes - .602.

Please note that measuring t/s parameters, especially VAS is tricky to get right and the various methods Delta Mass and Delta Compliance can yield different results.

In any case this driver (no matter which specs you use) can only handle about 5 watts per driver before hitting it's excursion limits below 100 Hz. So depending on where you high passed the driver this might be part of the distortion problem too.

Also since you went with the 16 ohm version, did you change the XO to match?

Kind regards,

Dennis

[sreten]

Originally posted by jquin

Any Ideas?
Are Manufacturers T/S parameters usually this way off?

Regards
John Q

Hi,

They are what you would get if a stiffened version specifically
to be used in AV filtered 80Hz to 100Hz was made. note that :

Qms 1.42 + Qes 0.64 does not = Qts 0.35, Qts = 0.44.
Which is what you expect for the 16 ohm version, /= 8 ohm version.

/sreten.