Tools and techniques to measure THD, Response etc

Member
Joined 2005
Paid Member
Most scopes won't do that. You'd need one that can do Fourier transforms. And you'd need a very clean sine input to your amplifier.

Frequency response
Assuming you have a sine generator with adjustable frequency connected to the amplifier input, just sweep the frequency and observed the amplitude change to determine the frequency limits at the output of your device.

Output Level
Read the amplitude of the test signal on the screen of your scope. Some scopes also provide measuremens of RMS, peak, and peak-to-peak levels.

DI? Dispersion Index? That's a loudspeaker thing, so I guess I don't know what you're referring to.
 
Thank you for your immediate response.
I have a Picoscope Oscilloscope which has many measurements including THD, Signal to Noise Ratio etc. and also signal generator.
So I will try it with this one to see.
As for the DI, I mean Direct Injection Box which convert the unbalanced signal to balanced. :)
I think with the same way as measuring preamp I will take the measurements also for the DI.
 
Hello there,

I would like to measure the THD, Frequency response, Output Level etc of a preamp and DI that I have made with the help of an Oscilloscope.
Please can you help me how to do it?

Thank you very much.
George
Hi George,

What bandwidth are you hoping to cover, and how low of distortion are you hoping to measure?

The higher the signal level, the greater SNR that can be characterized, as it allows you to stay above your source/sig gen’s noise floor longer. An audio interface can meet many goals of the DIYer both for signal source and measurement inputs. The QA401 measurement system also takes it further. It uses BNC terminals for low capacitance coaxial probes, and has inputs protected from overload.
 
Most oscilloscopes are very limited in distortion measurement due to the 8 bit precision (and poor input stage linearity - 'scope inputs are designed for high bandwidth not high precision, generally.
You are right, mine has 8 bit but it goes until 12 bit. I thought that it could be good to use it as my bandwidth is limited into khz.

Whatever equipment you use, your first step should be a loopback measurement to find the inherent limitations of the measuring system.
You are correct.

@kouiky
The bandwidth that I would like to have is from 20Hz until 30kHz.
I have the Picoscope 2204 which can go until 10MHz. It has 8 bit Resolution but can be enhanced until 12 bit.
Do you think that it's ok to use it or to go in a better solution as QA401 that you suggest me.

Thank you guys!
 
Last edited:
Sadly
@kouiky
The bandwidth that I would like to have is from 20Hz until 30kHz.
I have the Picoscope 2204 which can go until 10MHz. It has 8 bit Resolution but can be enhanced until 12 bit.
Do you think that it's ok to use it or to go in a better solution as QA401 that you suggest me.
Sadly, I’ve never got to use Pico-scope. A $150 audio interface however can easily do 2Hz-50kHz with 18 bits effective resolution using their onboard 24-bit ADCs and DACs.
 
Member
Joined 2010
Paid Member
I use ARTA and REW for all audio measurements.

ARTA will not be supported any more by its creator but in its current status, it is an excellent program.

REW is very good for loudspeakers and room measurements, but will do the job, and it is free.

With a good soundcard you could have excellent results with these programs.
 
  • Like
Reactions: 1 user
Hi George,

What bandwidth are you hoping to cover, and how low of distortion are you hoping to measure?

The higher the signal level, the greater SNR that can be characterized, as it allows you to stay above your source/sig gen’s noise floor longer. An audio interface can meet many goals of the DIYer both for signal source and measurement inputs. The QA401 measurement system also takes it further. It uses BNC terminals for low capacitance coaxial probes, and has inputs protected from overload.
These days its the QA403, from QuantAsylum, though this is not cheap (but it is cheaper than "professional" audio measurement units...
 
Hello,
My apologies for not replied you earlier, I was ill.
@gdan thank you for your suggestions.
I tried the REW and it's very good, I took some measurements and please I would like to ask you how to measure the dynamic response of an audio unit, etc the sound card or a preamp?
Here are the measurement of my SSL sound card. The -110.5 dBr is the dynamic range?

@Mark Tillotson thank you, I will check it. Yes It's not cheap, I agree with you.

Thank you guys and sorry again for my late response.
 

Attachments

  • SSL THD dB.jpg
    SSL THD dB.jpg
    132.5 KB · Views: 19
  • SSL THD Percent.jpg
    SSL THD Percent.jpg
    123.1 KB · Views: 17
You are correct.

@kouiky
The bandwidth that I would like to have is from 20Hz until 30kHz.
I have the Picoscope 2204 which can go until 10MHz. It has 8 bit Resolution but can be enhanced until 12 bit.
Do you think that it's ok to use it or to go in a better solution as QA401 that you suggest me.

Thank you guys!
For a number of years the 2204a was my only scope. It works great for testing and trouble-shooting most audio circuits and has enough bandwidth to observe loop stability problems with many amps (my 2204a bandwidth is more like 25 MHz). That includes square-wave testing.

Your 2204 can also do automated frequency-response measurements up to 100 kHz using FRA4Picsocope:
https://bitbucket.org/hexamer/fra4picoscope/wiki/Home

But it isn't useful for measuring THD much below something like 1%.

I concur with the REW+soundcard/audio interface approach for budget distortion measurements. Your soundcard seems to be pretty good - certainly suitable for assessing distortion of many DIY audio projects. You probably want to plot more harmonics, though. There is no guarantee that H2 is the dominant distortion, and it is useful to see how fast the harmonics fall-off with frequency.

jason
 
Last edited:
Member
Joined 2010
Paid Member
The dynamic range is defined as the ratio of the amplitude of the maximum possible signal (without distortion) of the device to the noise floor of the device.
So, first find the maximum signal that can output the device without distortion (for example +15 dB with a distortion of 0.1%).
Then find the noise floor, for example -110 dB.
The dynamic range is then, 125 dB (+15 - (-110) = 125dB).
 
For a number of years the 2204a was my only scope. It works great for testing and trouble-shooting most audio circuits and has enough bandwidth to observe loop stability problems with many amps (my 2204a bandwidth is more like 25 MHz). That includes square-wave testing.

Your 2204 can also do automated frequency-response measurements up to 100 kHz using FRA4Picsocope:
https://bitbucket.org/hexamer/fra4picoscope/wiki/Home

But it isn't useful for measuring THD much below something like 1%.

I concur with the REW+soundcard/audio interface approach for budget distortion measurements. Your soundcard seems to be pretty good - certainly suitable for assessing distortion of many DIY audio projects. You probably want to plot more harmonics, though. There is no guarantee that H2 is the dominant distortion, and it is useful to see how fast the harmonics fall-off with frequency.

jason
Hello Jason,
Thank you very much for your post.
2204 is a very nice tool, I'm using it to measure the output of amps but I will give it a try with the FRA4 Picoscope. Until now I'm using the standard Picoscope.

But as I checked with REW I have very good results.
I didn't expect it to be honest but the engineers of it surprised me! Very nice software!

@gdan thank you very much.
Very helpful your advice about REW software!
I will check it next days.
Can I take measurements both unweighted and weighted?