Speaker Impedance measurement ,circuit and sound card output

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
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.
From the measurement data I have seen using Mms as an added mass can be a little too much for some drivers (that is why I recommend using half Mms for single added mass measurements) so you may get more reliable results using 1/3 and 2/3 Mms as sm52 suggests.
 

sm52

Member
Joined
Mar 14, 2019
Messages
863
but I don't know whether should I change the m1 and m2, cause there is no more guidance.
I only had an external sound card, no amplifier. But the drivers are smaller than yours. I have uploaded the results to speakerbench.com for verification. The best results were obtained when m2 was 90% of Mms.
 

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
From the measurement data I have seen using Mms as an added mass can be a little too much for some drivers (that is why I recommend using half Mms for single added mass measurements) so you may get more reliable results using 1/3 and 2/3 Mms as sm52 suggests.

I made a new measurement with 4 * 28.33g mass, is the result better? There is an error message when I input DC resistance to calculate the t/s parameters, cause the actual value is around 3.5ohm, but REW measured a value less than that.

I just input 2.58ohm to pass it and get the T/S parameters.

42979


Maybe that cause by the negative connector of the circuit is not very well, the resistor is around 0.1ohm to 0 ohm.:rolleyesno:
 

Attachments

  • JL 12W3V3-4 TS MEASURMENT with mass 4 x 28.33g.mdat
    8.2 MB · Views: 10

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
I only had an external sound card, no amplifier. But the drivers are smaller than yours. I have uploaded the results to speakerbench.com for verification. The best results were obtained when m2 was 90% of Mms.
Thank you for your information.
I tried to use new mass value (1/3 and 2/3 of mms) to measure the JL 12W3V3-4 SUBWOOFER, maybe the result is better.
 

sm52

Member
Joined
Mar 14, 2019
Messages
863
I have uploaded your measurements to speakerbench.com. Quote: 'The fit analyzes your data and the fit result and provides you with a quality rating: Excellent, Good, Fair or Sorry.' Your result is Fair.
 

Attachments

  • fit.PNG
    fit.PNG
    19.2 KB · Views: 22

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
I have uploaded your measurements to speakerbench.com. Quote: 'The fit analyzes your data and the fit result and provides you with a quality rating: Excellent, Good, Fair or Sorry.' Your result is Fair.

Thank you.

Electrical error is 0.97, too large, I need to check the circuit connection again. :dizzy:
 

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
The result in #139 looked slightly better, but there will be variations every time you measure.
Hello John
I checked the connection of the circuit today,I think it's fine.

And I made a whole new measurement, but there is still a warning mesage about BL.
42990


The BL is not that flat.

42991


And I still can not upload zma file to speakerbench.com:praying:

I think the mass is too long and heavy,it's not fixed very well on the cone(if I leave them on the cone too long such as 10-20 minutes,one of them will fall),and the suspension is afftected when I move 2 mass from the cone,cause it pasted hard.

I think the mass need to be optimized :-(
 

Attachments

  • JL 12W3V3-4 TS MEASURMENT with mass 4 x 28.36g=113.44g.mdat
    8.2 MB · Views: 8

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
It is worth checking the signal level you are using to make sure it isn't too high. I'm not sure why you keep measuring though, why aren't you using the earlier results?
 

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
It is worth checking the signal level you are using to make sure it isn't too high. I'm not sure why you keep measuring though, why aren't you using the earlier results?

The signal level is -6dbFS, when I play pink noise, REW generator shows clipping message like this.

42992


I just want to get more accurate T/S values of this subwoofer.
 

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
It is the voltage across the driver terminals that matters. It is probably OK if you are using a 100 ohm sense resistor or thereabouts, as a rough proxy the test signal should not be loud when you measure the impedance. The T/S values will vary with drive level and voice coil temperature, there is no one 'true' set of values. The results are only valid for the particular test conditions under which they were obtained. I haven't seen anything to suggest you should not trust the values you obtained in the earlier measurements.
 

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
It is the voltage across the driver terminals that matters. It is probably OK if you are using a 100 ohm sense resistor or thereabouts, as a rough proxy the test signal should not be loud when you measure the impedance. The T/S values will vary with drive level and voice coil temperature, there is no one 'true' set of values. The results are only valid for the particular test conditions under which they were obtained. I haven't seen anything to suggest you should not trust the values you obtained in the earlier measurements.
Thank you for your help,John.


I will use the values in my first measurement.

I hope I can find out why I could not get good simulation in the later measurement.
 

Bernard

Member
Joined
May 22, 2017
Messages
120
Hello icbcodc,

Last year, in june, Mr. Jeff Candy (Speakerbench) said :
By following his instructions and John's advice to lower the weight of the masses, I got this:
If you look at the Speakerbench quality test result, you can see that the result is "FAIR".
This did not prevent Mr. Candy from saying:
and
I have put together your Speakerbench results in the attached pdf document.
You should follow John's advice after your first simulation and consider the measurements as correct. Anyway, you can redo the T/S measurements as many times as you like and you will get different results each time.

We don't need the precision of scientists who send men to a space station! ;-)

Bernard
 

Attachments

  • icbcodc_JL12_3 model tests.pdf
    436.7 KB · Views: 18

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
Hello icbcodc,

Last year, in june, Mr. Jeff Candy (Speakerbench) said :
By following his instructions and John's advice to lower the weight of the masses, I got this:
If you look at the Speakerbench quality test result, you can see that the result is "FAIR".
This did not prevent Mr. Candy from saying:
and
I have put together your Speakerbench results in the attached pdf document.
You should follow John's advice after your first simulation and consider the measurements as correct. Anyway, you can redo the T/S measurements as many times as you like and you will get different results each time.

We don't need the precision of scientists who send men to a space station! ;-)

Bernard
Thank you Bernard.

I can not upload my exported file to speakerbench.com. It's a pity.

I just want to improve the result, but failed , actually I use VituixCAD2 to check the error level, as we see in the pdf you uploaded, the first measurement is slightly better.

Maybe for 12inch subwoofer , small mass is better.
 

Bernard

Member
Joined
May 22, 2017
Messages
120
If you feel like measuring again, try spacing the weights a little further apart with a distance of 80 mm between the edges.
You can also do a single mass measurement. You will see that the results will not be very different from the dual mass method.
Good luck!
 

sm52

Member
Joined
Mar 14, 2019
Messages
863
icbcodc,
I noticed that your short circuit calibration is not very good in all measurements. This can be seen in the GD plot. Yours and mine. Also during 2 * 28.33 measurement there is 4 kHz interference. Also during nomass 4 * 28.35g-56.72g-56.72g measurement, there is interference at 200-300Hz and 6KHz.
 

Attachments

  • short JL.PNG
    short JL.PNG
    134.3 KB · Views: 16
  • short.PNG
    short.PNG
    98.7 KB · Views: 14

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
If you feel like measuring again, try spacing the weights a little further apart with a distance of 80 mm between the edges.
You can also do a single mass measurement. You will see that the results will not be very different from the dual mass method.
Good luck!

If you feel like measuring again, try spacing the weights a little further apart with a distance of 80 mm between the edges.
You can also do a single mass measurement. You will see that the results will not be very different from the dual mass method.
Good luck!
Yes, I‘ll try it later.

The mass is a bit big and tall, so I have to paste it on the cone with force,but that means I have to push and pull the cone when I take it off from the cone. I don't know whether this action cause problem.
 

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
icbcodc,
I noticed that your short circuit calibration is not very good in all measurements. This can be seen in the GD plot. Yours and mine. Also during 2 * 28.33 measurement there is 4 kHz interference. Also during nomass 4 * 28.35g-56.72g-56.72g measurement, there is interference at 200-300Hz and 6KHz.

42999

Group Dely is weird in short circuit, what does it mean? I checked the connection of the circuit ,it looks fine.

The circuit is simple, and I know sometimes the connection between the line and the crocodile clip(positive connector) is not good, but It was good last time I measured the impedance. but GD shift is still there last time as you found..
 

sm52

Member
Joined
Mar 14, 2019
Messages
863
The sound card may not be able to handle the short circuit. The ASIO buffer may be too small. Maybe you need to play with the levels. Many reasons.
 

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
The sound card may not be able to handle the short circuit. The ASIO buffer may be too small. Maybe you need to play with the levels. Many reasons.
Hi sm52

I tried to minimize the short connection impedance,and change the position of mass, I got new measurements with more flat BL curve, I don't now whether this fit is better,could you help me to check it on speakerbench.com.

Thank you very much.

43008
 

Attachments

  • JL 12W3V3-4 TS MEASURMENT with new POSITION 113.46g.mdat
    8.2 MB · Views: 12

Bernard

Member
Joined
May 22, 2017
Messages
120
Hello,
Here is today's verdict :

Fit quality : SORRY :-(
Zero-mass error : 0.25
Full-mass error : 0.39
Electrical error : 2.05
Error (fail > 1) : 1.02

You might as well give up and take the values you found earlier and build one of the boxes recommended by JL Audio.
Don't let the T/S settings ruin your life! :cool:
 

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
Hello,
Here is today's verdict :

Fit quality : SORRY :-(
Zero-mass error : 0.25
Full-mass error : 0.39
Electrical error : 2.05
Error (fail > 1) : 1.02

You might as well give up and take the values you found earlier and build one of the boxes recommended by JL Audio.
Don't let the T/S settings ruin your life! :cool:

Thank you Bernard.

This subwoofer will be installed in a new closed box a few days later, and I've ordered the box tonight. :T
 

Bernard

Member
Joined
May 22, 2017
Messages
120
Here are the measurements made in June 2020 with one of my speakers: the results between REW, Speakerbench and Vituixcad2 are close as you can see.
I had spent some time on it because it was the Covid confinement but I won't do it again in the future. ;-)
The hardest part is not finding the right weight for M2 (and making 4 masses out of it) but sticking them on the membrane vertically and removing the masses as carefully as possible for the M1 and M0 phases.
The measurement with the single mass method has been added to allow comparison of the two methods.
The choice is yours !
Cheers.
Image1.png
 

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
Here are the measurements made in June 2020 with one of my speakers: the results between REW, Speakerbench and Vituixcad2 are close as you can see.
I had spent some time on it because it was the Covid confinement but I won't do it again in the future. ;-)
The hardest part is not finding the right weight for M2 (and making 4 masses out of it) but sticking them on the membrane vertically and removing the masses as carefully as possible for the M1 and M0 phases.
The measurement with the single mass method has been added to allow comparison of the two methods.
The choice is yours !
Cheers.
View attachment 43009

Great research, Bernard.

Actually, your research answers a question asked by someone else few days ago: "What are the benefits of the dual added mass? " ?

And the answer by Kimmosto the author of Vituixcad2 is like this:

* BL is a bit more accurate.
* No need to measure Re.

Difference to single added mass is usually insignificant so dual mass is not so important for diyers and box designing IMO. Knowing accurate BL and Mms would be more important for driver manufacturers assuming that individual test samples represent whole production. Unfortunately that is not common.

Knudsen LOG model and others including suspension creep is not the same as simulation with standard T/S parameters having constant Cms and Rms. Simulation with suspension creep needs simulator supporting creep parameters - R0, C0 and Creep Beta in this case. In practice it's not easy to use parameters calculated with 'LOG model' checked outside VituixCAD Enclosure.


Your research says it using data. :T
 

icbcodc

New Member
Thread Starter
Joined
Apr 29, 2020
Messages
187
The sound card may not be able to handle the short circuit. The ASIO buffer may be too small. Maybe you need to play with the levels. Many reasons.
I found the orignal data of measurement which was made last year, the GD is almost flat using this same soundcard and same circuit. This soundcard is too old , and is a second or third hand product. I need to update it.:)

43010
 
Top Bottom