REW Impedence measurement QMS and QES are way out -Please help

[Peregrine]

I am experiencing a strange problem. I have measured 4 drivers, 3 of them have unknown specs and the 4th I have specs for. The QMS and QES are way out but Fs is spot on.
What I have tried so far is rebuilding the test rig/harness 3 times and have used different jacks front and back of PC still get the same results.

Looking at the graphs the spike in impedance at Fs seems to be very high, from my limited knowledge this could be the reason but I don't really know what to check further.

Attached Mdat files for review.

Any help would be greatly appreciated.

 

[John Mulcahy]

99g looks like way too much mass to add to that driver, 20g would probably be plenty. The free air measurement has a strangely shaped peak that looks wrong. Running the sweep at 0 dBFS leaves no headroom in the digital path which could cause issues, better backing that off a few dB.

 

[Peregrine]

I tried with 15g and got very similar results, I tried the 20hz to 20000hz but REW resets the values when I press start. Btw, thanks for the prompt response.

 

[John Mulcahy]

I tried the 20hz to 20000hz but REW resets the values when I press start

Sorry, don't understand what you are referring to there.

 

[Peregrine]

In the Measurement screen you have the option to change the sweep start and end. normally 0 to 24Khz, When I set it to 20hz to 20Khz it defaults back to 0 - 24 when pushing start.

 

[John Mulcahy]

Not something I have seen or can reproduce, what build of REW are you running?

 

[Peregrine]

REW 5.20. RC13b I will download latest version and try again.

 

[John Mulcahy]

If you make an impedance calibration measurement the range will be set to the full span since the cal file must cover the full frequency range, perhaps that is what you are referring to?

 

[Peregrine]

Ok, apologies, I misunderstood your previous comment, I will try with a lower DBFS, and downloading latest version. What is the min/max range QTS should be in so I know the results are accurate or is this a silly question?

 

[John Mulcahy]

Slightly silly, drivers can vary a lot, but a bug in the Qms value was fixed in 5.20.1 so all may be well when you change to the current release.

 

[Peregrine]

So I ran through the whole procedure again, with the new downloaded software and using 20g of weight. Could you take a look at the files and comment please, do you think thy are accurate?

 

[John Mulcahy]

You seem to have attached a measurement made in SPL mode rather than the impedance measurements.

 

[Peregrine]

Uploaded wrong file.

 

[John Mulcahy]

Free air measurement looks good, the added mass measurement looks odd and the model fit to it is poor. The added mass needs to be symmetrically placed (one method is to use 4 equal masses) and best being around the Mms value or a bit lower, so less than 10g might work better. You also need a very accurate scale to measure the mass, the result is only as accurate as the measurement of the mass itself. The SpeakerBench manual has good tips, it is for the dual added mass method (which REW also supports) but the measurement setup advice is good for either.

 

[John Mulcahy]

As an aside, the Q values look sensible.

 

[Peregrine]

Thanks for the info provided above, I have ordered a new scale that goes to 0.01g hopefully that will improve results and secondly the weight being balanced around the cone makes sense. But my concern is assuming the weight value is wrong adjusting it up or down by 5g in the T/S calculator has no effect on the Q values, is this correct?

 

[John Mulcahy]

The Q values are mostly (though not entirely) driven by the free air measurement, the added mass is needed for Mms and related values.

 

[Peregrine]

I am not sure the Q values are correct, when I measured the driver with the known values the Q values were very different. Listed spec from manufacturer Qms 3.8011 Qes 0.39089 the recorded specs were Qms 14.360 and Qes 1.922. Fs was 25.4 and 25.9 very close

 

[John Mulcahy]

Did you measure those with REW version V5.20.1 or later that includes the Qms fix?

 

[Peregrine]

You are correct the measurement was done with the previous version, but looking at the results of the 6 in driver measurements from old version vs new the QMS went from 4.5 to 6.135 QES from 1.477 to 1.848 and Qts from 1,112 to 1.420, I was expecting it to go the other way. I can pull the 18 inch out its enclosure again and run more tests but from what I see it looks like it is going to move further out from the manufacturer specs. Is it possible that the onboard soundcard of my PC is messing up the results or is the accurate Fs and indication that it is working correctly?

 

[John Mulcahy]

You cannot extrapolate from a different driver, but you can load the old measurements in the current build and re-run the TS param calculations. I haven't seen anything in the measurements to indicate a problem with the soundcard.

 

[Peregrine]

I loaded the old measurements into V5.20.2, Qms went down to 10.601 from 14.360. Spec sheet is 3.8011.

 

[sm52]

Peregrine
if you calculate the average Mms value according to your measurements, how much will it turn out?

 

[Peregrine]

Peregrine
if you calculate the average Mms value according to your measurements, how much will it turn out?

In the results for the 6 inch the Mms is 9.55g, does that answer your question?

 

[sm52]

You have made several attempts to measure T / S parameters. I'm talking about the average between them. If it is 9.55 g, then the weight of all weights should be 8.3-8.6 g. If we divide it into 4 weights, we get 2.15 g, one out of four.