Calibrating new Focusrite Scarlett Sole 3rd Gen

[dch53]

Hi. After using a Juli@ soundcard to test amps and crossovers for a few years I decided I needed an external soundcard. I'm trying to create an REW calibration file.

So, I've found a couple of threads on calibrating REW with this device. One had a similar problem to mine but it was never resolved.

Here's my setup:

Can't find a way to get rid of the Timing Reference and Loopback Input settings. The Scarlett Solo only has one input so they're not relevant. Maybe I should have bought the model with 2 line inputs. Too late now.

I have the Focusrite and REW sample rates set to 96kHz. Here's the 1kHz tone with the Focusrite Monitor pot set so Out and In match:

The resulting frequency and phase look like this

1.2dB and 2.6dB down at 20Hz and 20kHz respectively. That's no good. What am I doing wrong?

Thanks.

 

[John Mulcahy]

Looks like a pretty normal response. Click the Make cal file button to generate the calibration file.

 

[dch53]

Thanks John. I decided to put my Pete Millett Souncard Interface in the loop. Much better result:

I've calibrated with that. 1kHz THC+N looks like this:

Noise is higher than I'd expect at -75.6dBFS.

The Focusrite control panel doesn't appear to have a the latency adjustment like the ESS Juli@ card did that I had to have set to 2048 samples so I have to keep redoing the THD+N plot to get a good one.

I tried turning Direct Monitor on on the Focusrite hoping to avoid a pot adjustment but I suspect REW was showing zero headroom and the frequency plot was a mess.

I'll try doing some measurements and see how they compare with the ESS Juli@ internal soundcard.

 

[John Mulcahy]

Direct monitor creates a feedback loop, sending the input signal back to the output.

 

[dch53]

Thanks.

 

[trobbins]

Reduce your measurement bandwidth to adjust the N noise level (ie. enable high and low pass settings in Distortion settings of RTA settings.

 

[dch53]

I did my first measurement with the Focusrite:

Terrible noise result so I redid it using the Juli@ soundcard:

Not flash either but it is a tube amp. A 21dB difference in the noise level.

So, I did some loopback tests. Juli@ first:

and Focusrite:

More than 25dBFS worse Noise level.

Then I noticed that the levels of the fundamental are different.

I've been using the Juli@ for measurements for a number of years but I decided to revisit its calibration. There's no way to adjust the input level to get it close to the output level. There's no REW control and the Juli@ level is set to full:

I don't know what I did about that when I originally created the calibration file.

 

[dch53]

Remembered that originally I calibrated with the Juli@ card looped back to itself. Not with the soundcard interface in the loop.

The soundcard interface provides range switching, a digital metre and protection.

More investigation required as to why measurements with the Focusrite give such poor noise results