Speaker Impedance measurement ,circuit and sound card output

[John Mulcahy]

The author of VituixCAD has provided a measurement setup summary that is appropriate for an ECM8000.

 

[icbcodc]

The author of VituixCAD has provided a measurement setup summary that is appropriate for an ECM8000.

Thanks a lot.

This is a good news. I've downloaded and installed VituixCAD.

 

[xorcz]

Hello, I have a very basic question. I did the cable with 100ohm resistor, using the latest REW, notebook headphone output and microphone input. Both working, set to 100%, sweep level -8dB, which is just below clipping.
But whatever I do the first step shows: The level difference between the channels is larger than it should be...86,8%.. Should be the left output channel connected somewhere? Should I change the resistor value?

Thanks

 

[John Mulcahy]

The message is about a difference between the input channels, since both are used for impedance measurements. One is fed directly from the output, the other after the sense resistor. When doing open circuit calibration (so no load connected) the input channels should see the same level. In this case the measured channel is much lower than the reference channel, suggesting there is some load connected after the sense resistor. About 600 ohms would drop the level by about that amount.

 

[xorcz]

The problem was the 4 pin jack (TRRS) on the notebook. It does not have stereo mic input, just mono. I will use an older notebook..

 

[icbcodc]

The message is about a difference between the input channels, since both are used for impedance measurements. One is fed directly from the output, the other after the sense resistor. When doing open circuit calibration (so no load connected) the input channels should see the same level. In this case the measured channel is much lower than the reference channel, suggesting there is some load connected after the sense resistor. About 600 ohms would drop the level by about that amount.

Hello John,

I made a shelf to support the JL 12W3V3 subwoofer, and would like to measure its T/S parameters, but its cone is special,and where should I put the 4 pieces of masses?

 

[John Mulcahy]

Probably at the diameter where the voice coil former attaches, since it should be stiffest there.

 

[icbcodc]

Probably at the diameter where the voice coil former attaches, since it should be stiffest there.

Thank you John. I will try it today.

 

[icbcodc]

Probably at the diameter where the voice coil former attaches, since it should be stiffest there.

Hi John

I measured the subwoofer twice today.

It looks like the mass should be increase to around 40g and 80g, should I increase the mass?

first time with too small mass

Second time with more mass

 

[icbcodc]

Probably at the diameter where the voice coil former attaches, since it should be stiffest there.

Hi John
I made a third measurment with 43g and 86g mass, and the result looks better.

but the simulation and measurment for 4 mass is bad.

I put 4 mass like this.

Is there something wrong?

 

[John Mulcahy]

That's not so unusual, probably the mass measurements are not quite accurate. Ideally they should be accurate to better than 0.1g, which requires a very precise scale. The Bl looks quite flat through the resonance region though so you should be able to use those TS results.

 

[icbcodc]

That's not so unusual, probably the mass measurements are not quite accurate. Ideally they should be accurate to better than 0.1g, which requires a very precise scale. The Bl looks quite flat through the resonance region though so you should be able to use those TS results.

My scale is accurate to 0.01g， and I tried again with these mass ,and the new result is worse with varing BI value more than 10%.

Is this possible caused by the mass not fixing very well on the cone?

 

[John Mulcahy]

Maybe, but it isn't unusual for there to be variations with large drivers as even the act of fitting or removing a mass can alter the results if the cone moves much when doing it and the suspensions have memory. I think you are already following the recommendation to make the measurement with the largest mass first, with allows least disturbance when going through the steps.

 

[icbcodc]

Maybe, but it isn't unusual for there to be variations with large drivers as even the act of fitting or removing a mass can alter the results if the cone moves much when doing it and the suspensions have memory. I think you are already following the recommendation to make the measurement with the largest mass first, with allows least disturbance when going through the steps.

Yes, I put 4 mass , 2 mass and no mass sequently. I also inserted a pink noise playing between two measurment.

and I tried once more with more fixed mass, and the result looks good again.

 

[John Mulcahy]

Looks very good.

 

[icbcodc]

Looks very good.

Thank you very much.

But there is still a problem, I export the data and calculated in VituixCAD, can not pass, something wrong.

 

[John Mulcahy]

 

[John Mulcahy]

If you attach the mdat file I can look at it. You could also try your measurements at Speakerbench.

 

[icbcodc]

If you attach the mdat file I can look at it. You could also try your measurements at Speakerbench.

FYI.

 

[icbcodc]

If you attach the mdat file I can look at it. You could also try your measurements at Speakerbench.

Speakerbench is still not available for me, the same problem is still there when I upload file to it's server. I tried it a few weeks ago.

 

[John Mulcahy]

That looks fine to me. The model fit at the added mass peaks isn't great, but I have seen that with some driver types (with REW and Speakerbench). I think you can use the parameters REW has derived.

 

[sm52]

If you have Mms 150-160g, make one mass 100-120g, and the second 50-60g. Weighing accuracy is very important. I took 8 identical weights, weighed everything together, then divided by 8. I received, for example, 5.55 g. The graph turned out better than using 5g or 6g. It is better to fix the weights closer to the edge of the membrane at the ends of the letter X.

 

[icbcodc]

If you have Mms 150-160g, make one mass 100-120g, and the second 50-60g. Weighing accuracy is very important. I took 8 identical weights, weighed everything together, then divided by 8. I received, for example, 5.55 g. The graph turned out better than using 5g or 6g. It is better to fix the weights closer to the edge of the membrane at the ends of the letter X.

Thank you for your explanations, I will try new weights as you recommanded.

 

[icbcodc]

If you have Mms 150-160g, make one mass 100-120g, and the second 50-60g. Weighing accuracy is very important. I took 8 identical weights, weighed everything together, then divided by 8. I received, for example, 5.55 g. The graph turned out better than using 5g or 6g. It is better to fix the weights closer to the edge of the membrane at the ends of the letter X.

I read the paper of Dual added mass method again.

In the paper, m1 should be around MMS/2 and m2 should be around MMS, they werer using amplifier. I'm using a sound card, but I don't know whether should I change the m1 and m2, cause there is no more guidance.

and about how to mounting added mass,they do it like this

and the position will affect the result? I don't know. I think balance is more important.

 

[icbcodc]

That looks fine to me. The model fit at the added mass peaks isn't great, but I have seen that with some driver types (with REW and Speakerbench). I think you can use the parameters REW has derived.

Hi John
The Extended z parameters should be input to solve the high error value. The initial value are all 0, that cause high error value.

And Kimmosto update the latest version , these values will be calculated automatically.