Speaker Impedance measurement ,circuit and sound card output

[icbcodc]

The high sense resistor together with the input impedance of the soundcard is making a voltage divider that is reducing the input, hence the channel difference.

Thank you John.

I will replace the RSENSE which should have been delivered today and calibrate again tomorrow.

 

[jtalden]

I don't know how to do the conventional setup,sometimes I am also confused by the bad result. I just read the help file for impedance measurement and ask somebody who did that before and watch some Youtube videos.

We normally use the same channel for input and output for each purpose as:
Measurement Channel; output = R, input = R
Reference Channel; output = L, input = L
I just hadn't seen them crossed over before.

 

[icbcodc]

We normally use the same channel for input and output for each purpose as:
Measurement Channel; output = R, input = R
Reference Channel; output = L, input = L
I just hadn't seen them crossed over before.

Sorry, It's the first time that I've measured speaker impedance using REW, I don't have an idea about normal and unnormal way.

But I just get the right result today replacing the 1000 Ohm resistor with a 100 Ohm resistor,the problem is solved.

 

[icbcodc]

The high sense resistor together with the input impedance of the soundcard is making a voltage divider that is reducing the input, hence the channel difference.

Hi John,

There is a good news. I replace the RSENSE 1000 Ohm with a 100 Ohm one, channel level problem is solved and the curve looks pretty good.

I just put all level adjustable to "Max".

Do you think I should keep this sound card for further measurement?

 

[icbcodc]

The high sense resistor together with the input impedance of the soundcard is making a voltage divider that is reducing the input, hence the channel difference.

I did a full circuit calibaration and a first measurement of speaker impedance,maybe there is something wrong?

Level down to -6dbfs

Check level:

Open circiut cal

reference cal: a 49.95 ohm resistor replace the speaker load.

There is no curve in the result,why?

refenece cal result , looks good.

Ready to measure the loudspeaker impedance curve.

low signal level warning message.

first measurement curve,impedance is too high?

Is there anything I input wrong in REW?

 

[John Mulcahy]

You have the input channels the wrong way around, usually the input is on the R channel if you made the connections as shown in the help.

 

[icbcodc]

We normally use the same channel for input and output for each purpose as:
Measurement Channel; output = R, input = R
Reference Channel; output = L, input = L
I just hadn't seen them crossed over before.

I check the help again.

According to the help, soundcard measurement output channel should be RIGHT, input channel should be RIGHT also.

But for impedance measurement, soundcard reference output channel is not connected, input channel should be LEFT.

I know why I was wrong,I just copy the settings from the video,but I make the circuit referring the diagram in the help file, so the output channel was setting wrong to L, in my circuit it's Right output channel.

Here are the youtube video information.

In the circuit,output is LEFT.

In REW, output should also be LEFT. But in my REW,output should be RIGHT channel.

 

[icbcodc]

You have the input channels the wrong way around, usually the input is on the R channel if you made the connections as shown in the help.

Yes, I'll change it.

And the Asus Xonar U3 Plus should have been kept.

 

[icbcodc]

You have the input channels the wrong way around, usually the input is on the R channel if you made the connections as shown in the help.

Hello John,I'm back.

I modified the output channel and input channel settings in REW, and made a calibration, then I made Impedance measurements,but something is still wrong.

Could you help me?

The procedure and settings are shown below:

Max Volume, Right Channel Output and Right Channel Input( Right input channel will be connected to speaker positive in impedance measurement)

Clear Calibration:

Open circuit cal
Range setting: 20hz to 20000hz (Is it necessary?)
Level: 0 dBFS
Rsense value :not input (It is set to 0 by the software last step,I don't input any data)
Should I input the actual value of RSENSE which is 99.85 ohm in to Rsense field?

Open circuit cal result( 99.355% of reference),looks good.

Is it a correct result?

Short circuit cal
Range: 20hz to 20000hz
RSENSE: 0 (default)
RINPUT:1000 (default)
Rleads:0 (default)

Impedance short circuit cal result: 113 mOhm (around),

Is it correct?

Reference cal
Reference resistor value:49.85(around)

The curve looks weird from 4khz to 10khz,
something wrong?

Twice impedance measurements(added mass method)

T/S Parameters error,my goodness!

Low point on impedance curve causes DC resistance too high.

Did I do anything wrong?

 

[John Mulcahy]

Should I input the actual value of RSENSE which is 99.85 ohm in to Rsense field?

Yes, of course.

 

[icbcodc]

Yes, of course.

Alright, I'm trying.

But there is a warning message,should I ignore it?

But the curve and description looks good below,do you know why?

 

[John Mulcahy]

You have a very low value entered for the input resistance, is that figure correct? V5.20 beta 50 has a wider tolerance for level differences between channels.

 

[icbcodc]

You have a very low value entered for the input resistance, is that figure correct? V5.20 beta 50 has a wider tolerance for level differences between channels.

Hi John,

I just input 1000 ohm as the input resistance,because I don't know how to calculate input resistance.

How can I get the correct input resistance value?

BTW, should I update Rsense after "Open circuit cal" with the value calculated by REW in the curve?

 

[John Mulcahy]

How can I get the correct input resistance value?

Look for the figure in the specification sheet for the soundcard. If it isn't specified use 20,000.

BTW, should I update Rsense after "Open circuit cal" with the value calculated by REW in the curve?

REW is not calculating an Rsense value. Do not change it.

 

[icbcodc]

Look for the figure in the specification sheet for the soundcard. If it isn't specified use 20,000.

REW is not calculating an Rsense value. Do not change it.

Hi John

I've tried to find out the input impedance of this SOUNDCARD (Creative Soundblaster Digital Music LX SB0271),but I failed. Even for the latest Creative soundcard, I also couldn't find the input impedance information.

Creative Soundblaster Digital Music LX ,it's more than ten year old.

SB0271:

So I have to input 20000 ohm into Rinput field.

Here is the result (improved) :

Open circuit cal (better)

In description, 99.877%
In curve, 99.4ohm

Which number should I refer to?

Short circuit cal
In the curve, value is 182mohm,is it meaningful?

Reference cal

Is there anything wrong in the curve between 2khz to 20khz?

Ready for impedance measurement

No mass impedance curve

Is this curve correct between 3khz to 20khz?

Added mass impedance curve

Is this curve correct between 3khz to 20khz?

T/S parameter calculation:

a new warning message below

T/S parameter calculated:

Should I keep this very old second hand soundcard (it's more than ten year old) ?

 

[John Mulcahy]

In description, 99.877%
In curve, 99.4ohm

Which number should I refer to?

Those numbers are for REW, you don't have to do anything with them.

Short circuit cal
In the curve, value is 182mohm,is it meaningful?

Looks OK.

Reference cal

Is there anything wrong in the curve between 2khz to 20khz?

Yes, that noise should not be there.

Have you tried to do a normal soundcard cal with that card? Creative interfaces sometimes have unusual monitoring arrangements that cause feedback loops during measurement, it can be difficult to locate the controls to disable monitoring. See if you can get a good loopback response measurement from the card without noise.

 

[icbcodc]

Have you tried to do a normal soundcard cal with that card? Creative interfaces sometimes have unusual monitoring arrangements that cause feedback loops during measurement, it can be difficult to locate the controls to disable monitoring. See if you can get a good loopback response measurement from the card without noise.

No.

I'm just trying to do a soundcard cal with it in latest REW,and the result looks bad and crazy.

Check level

Right output to Right input channel loop back

bad result curve

Try again, red is left input channel loopback, green is right input channel loopback.

Both are bad.

circuit is like this :

I just short connect the right input to the point in front of the RSENSE,so the Rsense is bypassed.

So I should get refunded and buy a new soundcard?

 

[John Mulcahy]

That is the appearance of monitoring being active, so some of the input is being fed back to the output. It needs to be turned off. If you go to the Windows sound Recording tab and look at the properties for the Creative input on the "Listen" tab make sure "Listen to this device" is not selected. Sometimes it is only possible to turn off monitoring using the Creative control software

 

[icbcodc]

That is the appearance of monitoring being active, so some of the input is being fed back to the output. It needs to be turned off. If you go to the Windows sound Recording tab and look at the properties for the Creative input on the "Listen" tab make sure "Listen to this device" is not selected. Sometimes it is only possible to turn off monitoring using the Creative control software

Thank you so much, John. I found the problem.

It's not here "Listen to the device" ,it was unchecked.

It's here in the Creative control panel. CMSS-3D was checked maybe by default. This caused the noise at high frequency part.

I just unchecked it

Normal soundcard calibration is fine.

And I do a full calibration + impedance measurement + T/S parameter calculation

Open circuit cal, looks perfect.

Short circuit cal, looks better

Reference resistor cal, looks better.

NO MASS impedance measurement,looks better.

ADDED MASS, looks better.

T/S PARAMETERS without any warning message.

Maybe I should keep this soundcard for further measurement.

 

[John Mulcahy]

Well done on tracking that down, looking good now. You could clean up the measurements a little more by using the "Noise" filter option on the Measure dialog, just below the impedance calibration buttons.

 

[icbcodc]

Well done on tracking that down, looking good now. You could clean up the measurements a little more by using the "Noise" filter option on the Measure dialog, just below the impedance calibration buttons.

Thanks a lot, John.

I tried with Noise filter meduim and without Noise filter. It's a beautiful function.

With filter on result: smoother

Without filter: good but not that smooth.

If I don't use this filter,the little noise will affect the T/S parameter?

 

[John Mulcahy]

The parameters will probably be OK with or without the filter, unless there is some background acoustic noise - the filter is helpful in rejecting background noise which could otherwise affect the values since drivers are microphonic. If the driver has a very sharp resonance the filter might affect the peak, however, so it is worth checking that.

 

[icbcodc]

The parameters will probably be OK with or without the filter, unless there is some background acoustic noise - the filter is helpful in rejecting background noise which could otherwise affect the values since drivers are microphonic. If the driver has a very sharp resonance the filter might affect the peak, however, so it is worth checking that.

Yes, a quiet environtment is needed for measurement.

The filter is designed for me because my environment is a little bit noisy(trucks engine and tire noise,children's shouting and other machine noise)

Everytime I run the measurement,I have to wait for a few seconds till environment is quiet temporarily.

Without this Noise filter,more accuracy for measurement means more patienc.

 

[icbcodc]

The parameters will probably be OK with or without the filter, unless there is some background acoustic noise - the filter is helpful in rejecting background noise which could otherwise affect the values since drivers are microphonic. If the driver has a very sharp resonance the filter might affect the peak, however, so it is worth checking that.

And I have another question，free air environment means the speaker needs to be far away enough from the ceiling,floor and walls, but actually I only put the speaker on a table and keep it's hole in the megnet unblocked by the table surface.

Do you think the free air environment is important to the impedance measurement?

 

[John Mulcahy]

Free air means the driver should not be in a box or baffle or very close to a surface. It is best if it is supported vertically, as it would be when in a speaker.