I want to be able to do distortion analysis, examining the distortion products and the distortion waveform, for the BA-3 I am building. From what I've read, I can get my feet wet with a Scarlet 2i2 and Room EQ Wizard, or other available software. I assume I can wire a mono TS jack to the speaker outputs on my amplifier and connect them to the Scarlet. Is any of this incorrect?
You'll need some attenuation if feeding a high level signal to a sound-card interface, otherwise it risks blowing it up. Often people put an attenuation network and dummy load in a box for this sort of testing.
So I’d put the dummy load inline with the positive lead, and connect the rest as normal? Do I need a dedicated attenuator after the dummy load to bring it down to line level?
I guess the following is a heads-up, as the general topic can get technical.
You need to be aware of how the input signal is electronically connected to the test amp with respect to the measurement system ground, and similarly how the output signal is connected. That is because a USB soundcard like the 2i2 uses a common ground for its output signal and input signal. Certain amps, like with class D output stages need special attention as one end of the speaker is not connected to amplifier ground.
It's also good to link to details on the amp you are using, as not everyone knows what a BA-3 is. It looks like the BA-3 may have a grounded speaker output, but that needs to be confirmed.
If the BA-3 has a direct connection between input and output grounds then that simplifies the interconnections needed to get you started.
HD measurement is normally based on a static resistive load, so not a speaker for a load. Normally that means a load resistor that can easily handle the output power of the amp and is the nominal resistance (impedance) of the speaker and the resistor is not prone to significant distortion itself.
The peak output voltage from the amp can be calculated using P=Vsq2/R to give rms voltage and then converting rms voltage to a peak level from ground. Are you ok with those calculations? Then you need to resistor divider attenuate that output voltage to bring it under the max peak level that the 2i2 can handle - so that the 2i2 input doesn't clip. Can you have a go at determining those voltage levels and resistor values for a voltage divider?
The other aspect is that REW needs a reasonably modern PC to run on, and preferably a laptop (so as to alleviate the chance of annoying hum/noise artefact signals showing up in the spectrum). Also you shouldn't be shy about technically detailed software programs like REW - there are help tutorials and threads on how to use it, so you need to put in the effort before crying for help if something doesn't quite work and you have only just started to use the program.
You need to be aware of how the input signal is electronically connected to the test amp with respect to the measurement system ground, and similarly how the output signal is connected. That is because a USB soundcard like the 2i2 uses a common ground for its output signal and input signal. Certain amps, like with class D output stages need special attention as one end of the speaker is not connected to amplifier ground.
It's also good to link to details on the amp you are using, as not everyone knows what a BA-3 is. It looks like the BA-3 may have a grounded speaker output, but that needs to be confirmed.
If the BA-3 has a direct connection between input and output grounds then that simplifies the interconnections needed to get you started.
HD measurement is normally based on a static resistive load, so not a speaker for a load. Normally that means a load resistor that can easily handle the output power of the amp and is the nominal resistance (impedance) of the speaker and the resistor is not prone to significant distortion itself.
The peak output voltage from the amp can be calculated using P=Vsq2/R to give rms voltage and then converting rms voltage to a peak level from ground. Are you ok with those calculations? Then you need to resistor divider attenuate that output voltage to bring it under the max peak level that the 2i2 can handle - so that the 2i2 input doesn't clip. Can you have a go at determining those voltage levels and resistor values for a voltage divider?
The other aspect is that REW needs a reasonably modern PC to run on, and preferably a laptop (so as to alleviate the chance of annoying hum/noise artefact signals showing up in the spectrum). Also you shouldn't be shy about technically detailed software programs like REW - there are help tutorials and threads on how to use it, so you need to put in the effort before crying for help if something doesn't quite work and you have only just started to use the program.
Last edited: