Notch filter with REW, cannot understand the use of [FS sine Vrms]

jesperlykke

Registered
Thread Starter
Joined
May 26, 2020
Messages
20
Hello all.

A long time user on diyaudio, lead me to here to get hopefully answer to a quistion(s) i have. (User lykkedk @ diyaudio.com)

Recently i "build" a setup to measure amplifier etc. harmonics, noise and so.
It's been some journey, but i'am getting good results.
The build is based on a modded Behringer soundcard, which is modded with a balanced input, acepting amplifier input directly, without use of voltage dividers.

I'am trying to use a notch filter, along with my internal generator / external 1kHz sine osc. and / or measuring amplifier's.
Reason i build the notch is to have much lower input's on the Behringer, hopefully getting even better result's.

I will explain what i did for using it for now.

* With input-level at MAX.
* Set genreator at known voltage, and calibrate REW to that in the "Generator" dialog.
* Verify, using RTA with "V" that the voltage is correct.
* Make a sweep and save this as "SoundCard" (Generator out ---> Behringer in)
* Then i connected my notch (Generator out ---> Notch ---> Behringer in) - Make a sweep and save as Notch1KHz
* Next trace arethmetic (Notch / SoundCard), and Save/Export as mic.Calibration file.
* Load mic. calibration file,
* With external osc. set at a known voltage (E.g. 1008mV) feed like this (1KHz-external ---> Notch ---> Behringer)
(This could be at -3dB, just befor the soundcard starts to clip)
* Plug in the fundamental in the dialog at RTA to this known voltage (1.008V)
Now REW show's the correct number's and everything seem's allright.

It's seems to be very good, also when measuring through an poweramplifier. (External osc. ---> Poweramp. ---> Notch ---> Behringer)
If i set the output of the poweramp. to around the same voltage as the fundamental (1008mV) the dB shows correct numbers (Rightside, Top on RTA window), and again everything seems perfect.

But if i then set the amplifier output to say 2.800V for measuring at around 1w, the dB is going positive. (Could end up at maybee 5.6dB or so) (EDIT :: and setting the fundamental to this value ofcause)
This is because the FS sine Vrms is set to the calibrated level at -3dB i guess.

How do i handle this? I cannot figure it out guy's
I think it's the only missing parameter i need to know for having it under control :T

Attached some pictures, with text to describe what it is.

Jesper.

Don't mind if the numbers does not fit with my description. (The USB-PC-6dB is a typical "loopback" with external osc.)
 

Attachments

  • USB-PC-6dB.png
    USB-PC-6dB.png
    38.4 KB · Views: 78
  • FS-sine-Vrms.png
    FS-sine-Vrms.png
    9.7 KB · Views: 87
  • NOTCH_in_action.png
    NOTCH_in_action.png
    52.7 KB · Views: 92
  • Modded_Behringer_Setup.jpg
    Modded_Behringer_Setup.jpg
    285.6 KB · Views: 94
Last edited:

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
Are you making any change to the gains along the input path? If the relationship between input voltage and input dBFS values changes you need to repeat the level calibration to show correct voltages.
 

jesperlykke

Registered
Thread Starter
Joined
May 26, 2020
Messages
20
Are you making any change to the gains along the input path? If the relationship between input voltage and input dBFS values changes you need to repeat the level calibration to show correct voltages.

EDIT :: Dont mind my 1KHz is off to 1088Hz (sometimes it is showing this in REW, donno why)

Hi John... Thank's for the reply!

Yes, i'am making changes in the input path.

If the relationship between input voltage and input dBFS values changes you need to repeat the level calibration to show correct voltages.

I'am a bit lost here i think.
Is this correct what i do :
WITHOUT NOTCH :
* Set external oscilator to 1055mV - Just before soundcard is clipping
* Plot this into the FS Sine Vrms calibration (1055mV)
* Confirm in RTA that it shows 1.05V
WITH NOTCH:
* Load mic. calibration
* Set manuel fundamental to 1055mV
* Confirm this in RTA (Showing 1.05 / -3dB)
- This tells me that my notch & mic. calibration is working as expected

When i then measure an amp, with around 1055mV@8ohm - I can also confirm that everything is looking right.
What i still really don't understand is what to do when turning my amp. up til say 2800mV
Entering 2.8V in manuel fundamental then gives me +5.5dB

Should i simply just plugin 3950mV to have -3dB into calibration (or use preset, which is a nice feature)
Is it correct (I see that THD numbers are changing, so i would really like to do it correctly)

The picture i attached is having text as explanation.

I really appreciate the help here.
And it possible please correct my spelling in the title (undertand = understand :))

Jesper.
 

Attachments

  • Calibrate_FS-Sine-Vrms.png
    Calibrate_FS-Sine-Vrms.png
    46.1 KB · Views: 69
  • Verify_NOTCH.png
    Verify_NOTCH.png
    46.7 KB · Views: 71
  • CloneAmp_Confirm_Notch.png
    CloneAmp_Confirm_Notch.png
    50.6 KB · Views: 65
  • Notch_5.5dB.png
    Notch_5.5dB.png
    51.2 KB · Views: 70
  • Adjust_FS-sine_Vrms-3dB.png
    Adjust_FS-sine_Vrms-3dB.png
    51.9 KB · Views: 68
Last edited:

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
With your 1.05 V input the ADC peak is -3 dBFS. If you were to put 2.8 V as an input the ADC would clip heavily, since there is only 3 dB of headroom. 2.8V corresponds to a further 8.5 dB so the equivalent digital level would be +5.5 dBFS. You are notching out the fundamental, so the ADC never sees such levels and doesn't actually clip, but the display is correct. If there were 2.8V coming in (the pre-notch level) that would be 5.5 dBFS and that is the correct fundamental level to use for the distortion calculations, FS sine should not be adjusted.
 

jesperlykke

Registered
Thread Starter
Joined
May 26, 2020
Messages
20
That's a very good explanation John.

What i needed to fully understand it.

The values you describe also fit the actually values i have.
Also the THD number's fit much better when leaving the [FS Sine Vrms] at the calibrated one. at 1055mV.

This also tells me, that 0dB is around 1100mV which is "bitwise" at 0dB scaling from there to the 24bit the Behringer have teoretically at -144dB.(dBFS) - My noisefloor is around there, eventhrough i think REW is showing the averaged one here, so could be lower... but well its very good as is!

Jesper.
 

jesperlykke

Registered
Thread Starter
Joined
May 26, 2020
Messages
20
Hello all here...

With the new option to select the AES17-2015 notch, is there something to know, regarding the way i use my passive notch ?
Also the way it's working is to "create" a skirt where the fundamental is, having this area not used when calculating. - Is is also prefered to use this without the use of a notch filter (passive or active) ???

If the Use AES17-2015 standard notch option is selected the fundamental power for THD and THD+N calculations will be the power within a one octave span around the fundamental frequency. If that option is not selected it will be the power in the main lobe of the fundamental. When the fundamental level approaches the noise floor using the standard notch will produce a much higher figure than the fundamental main lobe, in those cases it is better not to use the standard notch.

52653


Jesper.
 

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
7,212
With the new option to select the AES17-2015 notch, is there something to know, regarding the way i use my passive notch ?
Also the way it's working is to "create" a skirt where the fundamental is, having this area not used when calculating. - Is is also prefered to use this without the use of a notch filter (passive or active) ???
That has nothing to do with whether you use a notch filter or not, "standard notch" is the AES-17 terminology for the way the fundamental power is to be extracted from the response.
 
Top Bottom