Poor measurements of Scarlet Solo, setup issues??

[paul8]

I'm trying to measure Scarlet Solo USB audio interface internal noise and THD, and it shows very poor values no matter what I do.
This is on Macbook Pro M1, BTW. Scarlet itself is not clipping judging by the indicators on front panel.
I basically just connect balanced input with right channel output, and then play 1 KHz tone, sample rate is 96KHz, THD-N is 99%, something is off here, but what?

Thanks for tips.

 

[sm52]

very low fundamental signal.

 

[trobbins]

Play with it - turn the signal level up and down. Do you have a meter and can it measure a Vrms level?

 

[John Mulcahy]

Have you selected the correct input channel? Looks more like crosstalk than an input signal.

 

[paul8]

Have you selected the correct input channel? Looks more like crosstalk than an input signal.

I did, when I check levels in REW everything looks good, I'm able to match out and in around -12 dB, but then when running actual test REW reports low input level. My guess is REW doesn't play well with Scarlet drivers on Mac...

 

[John Mulcahy]

Unlikely, plenty of folk using Focusrite interfaces with REW on Mac. REW can't control interface sample rate on macOS so you need to make sure the rate you want to use in REW is also configured in the Mac's Audio Midi Setup, but beyond that REW on Mac works well

 

[linearphase]

I did, when I check levels in REW everything looks good, I'm able to match out and in around -12 dB, but then when running actual test REW reports low input level. My guess is REW doesn't play well with Scarlet drivers on Mac...

Might this be case of the signal source turned off or a open cable?

 

[paul8]

OK, I switched to creative SB usb card, the card itself plays very well with REW and shows reasonable 0.05% THD-N, I'm even able to calibrate it.
However trying to measure my receiver THD by sending 1KHz signal to it using card's TOS output, and feeding pre-outs into the card's input I get some unreasonable measurements again, like 50% THD-N with pre-outs sending only 1.04V. Looks like only if I send 0.2-0.3 volts in pre-outs then I get reasonable SINAD.
Any hints here?

 

[trobbins]

Perhaps you need some more tools to self-diagnose your issues. Do you have an alternative sinewave generator, like a 555 powered from a 9V battery, or a USB music player that can play a sinewave, or a test record or tape or cd that can input to your receiver.

 

[sam_adams]

You are clipping either the DAC on the AVR or the soundcard input. You need to reduce the level.

 

[paul8]

You are clipping either the DAC on the AVR or the soundcard input. You need to reduce the level.

Yes, I played with levels and was able to get reasonable measurements, but it was rather tricky, basically I had to reduce input gain/level on the USB card side to almost 0 to prevent clipping.
Regarding original Scarlet Solo issue, once I switched from balanced input to TRS issue got resolved, likely cable problem.

 

[John Mulcahy]

I had to reduce input gain/level on the USB card side to almost 0 to prevent clipping

Sounds like there may have been some monitoring active causing a feedback loop.

 

[paul8]

Sounds like there may have been some monitoring active causing a feedback loop.

Could be, "direct" mode was on though on Scarlet Solo.
I'm mostly surprised that noise levels were much higher compared to measurements done by other people on same receivers (Denon and Yamaha), they used pro tools though.

 

[trobbins]

There are a few ways to modify the 'noise floor' in RTA. I'd suggest focussing more on ensuring you are seeing valid harmonic levels, and in doing so then appreciating how your view of the 'noise floor' evolves with experience. For starters, its worth doing a loopback to act as a benchmark, and even a loopback with input and output connecting to the same points on a powered amp to confirm your system environment benchmark, and then to play with the averaging and window settings, and coherent averaging.

 

[paul8]

There are a few ways to modify the 'noise floor' in RTA. I'd suggest focussing more on ensuring you are seeing valid harmonic levels, and in doing so then appreciating how your view of the 'noise floor' evolves with experience. For starters, its worth doing a loopback to act as a benchmark, and even a loopback with input and output connecting to the same points on a powered amp to confirm your system environment benchmark, and then to play with the averaging and window settings, and coherent averaging.

Yes, I did loopback calibration for Scarlet Solo, if that's what you mean.

 

[trobbins]

Did you then make a measurement with input and output cables/connections connecting to same point (input or output) of a powered amp? That may show up environmental noise pickup due to power loops etc.

Then perhaps make an RTA measurement with amp with all connections in place and amp on, but with no signal at amp input - that will then identify if you have any ground loop noise present from your soundcard connections going through the 0V link of your amp.

 

[paul8]

Did you then make a measurement with input and output cables/connections connecting to same point (input or output) of a powered amp? That may show up environmental noise pickup due to power loops etc.

Then perhaps make an RTA measurement with amp with all connections in place and amp on, but with no signal at amp input - that will then identify if you have any ground loop noise present from your soundcard connections going through the 0V link of your amp.

All good suggestions, will also try to disconnect my Mac from power outlet for measurement session, thanks!