• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Teardown, few basic measurements and personal thoughts of the Focusrite Solo Gen3

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,965
Likes
2,516
Location
Iasi, RO
I recently purchased the Scarlett Solo 3rd Gen and couple of mics, mostly for podcasting and audio measurements.

The new generation seems to be better than the previous one regarding on how it measures; feel free to compare http://prosound.ixbt.com/interfaces/focusrite/scarlett/solo/gen2-2444.shtml with http://prosound.ixbt.com/interfaces/focusrite/scarlett/solo/24-44.shtml and you'll see that background noise is lower on the 3rd Gen and harmonics are all below -100 dB.

thd.png

Solo Gen2

thd.png

Solo Gen3

Despite what @amir said in another thread, I found that Focusrite ASIO drivers seem to be working fine on my desktop and also on a Lenovo T440p. Probably the 3rd Gen of Scarlett interfaces are somehow better or the newer drivers are better (or maybe it’s because of the latest Windows 10 update from September, 2019). The ASIO control panel is locking the bitrate to one single value chosen by the user, so if the software is trying to change this automatically, then it may crash, although sometimes it does it well (depends how well is accessing the ASIO drivers). However, from the Focusrite ASIO control panel the bitrate and bitdepth can be easily adjusted, but Windows 10 control panel will always see the max values that were previously setup in the Focusrite control panel!

The software bundle is very generous, including “Ableton® Live Lite™ and Pro Tools® First Focusrite Creative Pack – Softube’s Time and Tone Bundle; Focusrite’s Red Plug-in Suite; and a choice of one of four of XLN Audio’s Addictive Keys virtual keyboard instruments”.

I find the Ableton Lite 10 DAW not so easy to use, but very good for podcasting or mixing. However, I prefer to use this interface on my iMac with GarageBand, although the Ableton application works perfectly fine on both latest Windows 10 and latest Mojave as well, but GarageBad is having an easier visual interface and more built-in plugins and controls (I actually use the compressor, stereo expander and optionally the EQ as high-pass filter).

The bundle also includes Pro Tools | First Focusrite Creative Pack that asks for an iLok license management account that can be easily created within couple of minutes.

The RED PLUG-IN SUITE seems to contain a large suite of EQ and compressor plugins.

Also Gen3 has the AIR mode, introduced 30 years ago by Focusrite, that emulates an “airy” sound that increases the clarity, emphasises the voice and makes the sound more upfront and...”airy”. Focusrite has a graph with the airy effect here:

clarett-frequencyresponse-graphs-body-980.png

It’s easy to see the gradually volume boost with about 4 dB from 100 Hz to 10 kHz.​

Just to double check if my Solo behaves in the same mode when AIR is pressed, here are my measurements with AIR button not pressed vs. pressed:

Mic_freq_response.png

Original MIC-input frequency response

Mic_freq_response-AIR.png

AIR-mode pressed MIC-input frequency response

Pretty close to what manufacturer states in their documents, like expected actually. Based on manufacturer wrote, here’s what the AIR effect is intended to do:

“There are three main elements that contribute to the classic ‘Air’ effect.

Coupling – the interaction between the microphone and the ISA’s mic pre input transformer.
Clarity – created by the low distortion and high linearity of the transformer and preamp design.
Frequency – a subtle enhancement created by the transformer resonance resulting in an emphasis in the higher frequency content of the sound.


Engaging the setting on the Clarett Mic Preamp switches the impedance of the preamp to that of the original ISA and enables the ‘transformer resonance effect’, giving your microphone recordings the air and clarity of an ISA transformer-based mic pre recording. ‘Air’ is useful with any microphone, but in particular listen to the benefits with vintage microphones for which this input impedance was originally intended”.

In my tests, without AIR mode enabled and volume setup for -1dB peaks, when I pressed the AIR button it starts clipping a lot per the increased gain from AIR. However, after recalibrating the volume, ARTA is showing an increase in harmonics from less than -100 dB to about -80 dB. So, there's something else that AIR-mode does to the output sound.

I also have an audio recording done by me today: https://drive.google.com/open?id=14HOs8jNQ0S3Rrt_XkfZ8caydzzZekRZH. The first speech is done normally, the second one is done with the AIR button pressed. You should notice the 1 dB increase in volume, but also the more “airy” and brighter sound when in AIR-mode. It’s like a proximity effect of the mics, but on the mid-trebles instead of the mid-bass, so quiet a pleasant effect.


The PCB seems well done and clean enough, with good-looking soldering, although some flux residue can be seen on the backside of the PCB, but not sure anyone should actually care about this.

WP_20190916_19_56_45_Pro.jpg

Front of PCB

WP_20190916_20_10_03_Pro.jpg

Bottom of PCB

The USB transport is done via a XMOS chip. The USB plug is an USB-C small format, but the USB cable is USB-C to USB-A, so it can get connected to any regular computer. The ADC/DAC processing is done by a $10 CIRRUS LOGIC CS4272-CZZ chip able to record/playback two 24/192kHz audio signals with a dynamic of 114 dB and a THD+N of -100 dB (a SINAD of about 17 dB). Datasheet can be found here: https://statics.cirrus.com/pubs/proDatasheet/CS4272_F1.pdf.

WP_20190916_19_46_45_Pro.jpg

Close-up with the PCB
On the PCB bord I can spot a MC74HC595A chip which is a 8-Bit Serial-Input/Serial or Parallel-Output Shift Register with Latched 3-State Outputs, a HEF4053B - Triple single-pole double-throw analog switch, couple of NJM2122 - ultra low noise dual operational amplifier and several NJM8065 - dual operational amplifier.

I see that NJM2122 is quite a low noise opamp and given their positioning on the PCB layout I think these opamps are used on the MIC and INST inputs, just in front of the ADC chip, while the NJM8065 looks more like a general-use opamp, probably used on DAC outputs (the two TRS plugs) and probably for the AIR-mode effect and in the gain-stage of the headphones amplifier (not sure what the output buffer is, as the chip is scratched a bit with two parallel lines/scratches). I wonder how much the use of NJM8065 will influence THD+N? Maybe swapping those with OPA1612 will improve the final specs, although I’m not sure it really worth the trouble, given the risks caused by the very close proximity of opamps to the surrounding SMD passive components.

WP_20190916_19_48_56_Pro.jpg

JRC/NJM2122 and 8065 can be spotted on the board

On the backside of the PCB I identified a chip that might gather the BIOS/firmware: https://www.mct.net/download/macronix/mx25l8005.pdf. Inside there should be the built-in firmware code, most likely upgradable via the USB connectivity from a computer running Window.

Nearby the headphones stereo plug there’s a boost-converter voltage regulator chip marked as MKPC 3425 894. This might be similar with the datasheet found here: https://www.monolithicpower.com/pub/media/document/MP3425_r1.1.pdf. This is most likely used to boost-up the USB provided +5V power up to +48V to supply the “phantom power” to dedicated dynamic microphones.

Given the rather low maximum output power provided on the headphones-out and the lack of dedicated output buffers, I really think that the headphones are directly driven by the NJM8065 opamps, so sensitive headphones are required for monitoring purposes. I’ve used sensitive AKG K550 closed headphones, in a very quiet environment (quiet bedroom at night with no audio pollution sources) and the volume was very loud to my ears when I got the knob volume over 2 o’clock (couldn’t pass over 2 o’clock, because I care about my hearing). However, when using Audio Technica ATH-20X monitoring headphones the knob volume could be maxed out without feeling that my hearing will get damaged. I also tested the sound with Hifiman HE-560, but I felt like the volume is around 25% then needed, so low indeed, but it was clean and free of audible distortions.

The polarised capacitors are branded YUSCON and some of them seem to be rated at 85C, but some others to 105C (probably the ones used as bypassing the audio opamps are at 105C). Given that the audio interface is completely cold, even after hours of operation, I see no reason to pick up 105C capacitors inside, so I see a frugal choice here. Given the very low ripple that can exist in such audio interfaces that take the power from the USB’s +5V, I don’t think there will ever be issues with these caps. Well, the phantom power capacitor nearby the dedicated SMSP boost-converter is probably the most “stressed” one.

WP_20190916_19_49_55_Pro.jpg

Caps nearby the ADC/DAC chip, probably used to bypass its voltage rails

When I touch the outer shell of the XLR plugs from MIC/INST I get a tiny/little noise if no mic & no instrument are connected (max. input impedance) when maximum gain is used. That makes me thinking that the audio ground might be separated from the USB/computer’s ground. If that’s true, then this means that the probability of getting ground-loops is extremely low.

I’ve tested both Instrument and Microphone inputs and I see that the self-noise is about the same for both when set to minimum, with all the noise kept well below -140 dB, while when setting the gain to the maximum the noise increases a bit above -120 dB on the MIC input and a bit above -100 dB on the INST input. This is probably due to the gain difference between the two inputs, although the better THD+N I was able to get it on the INST input: 0.0015% on INST vs. 0.0038% on MIC.

Focusrite_Scarlett_Solo3-Noisefloor_mic_min.png

MIC input self noise when nothing connected and gain is to the min

Focusrite_Scarlett_Solo3-Noisefloor_mic_max.png

MIC input self noise when nothing connected and gain is to the max


ARTA_loopback_mic.png

MIC input used, 1 KHz sinewave

ARTA_loopback_mic_AIR-mode.png

MIC input with AIR button pressed

ARTA_loopback_mic_AIR-mode_Direct_Monitor.png

MIC input with both AIR and Direct Monitor buttons pressed
You can easily see the effect of harmonics increase when AIR button is pressed, but also the odd distortions and decrease of output volume caused by the Direct Monitor. Not sure what is causing such kind of effect when Direct Monitor is pressed, but this is usually not pressed while doing recording, so it shouldn't matter much.

Note: Measurements were done in loopback mode at an output level of 1.145V RMS (best signal/noise ratio I found), so the THD+N figures represent both ADC + DAC distortions and noise! Sinewaves were taken from audiocheck.net, 16 bit/48 KHz, because these are lower in harmonics than ARTA's built-in generator. However, the actual measurement's results for this audio device will definitely be better if Audio Precision equipment will be used.

I also tested the MIC input with an AKG D5S dynamic mic (no phantom power needed) with a self-noise of 18 db(A) and also with a SUPERLUX ECM999 measurements condenser mic (needs 48V phantom power) with an estimated self-noise of 22 dB(A) (if it's using the same capsule as ECM888 mic). Recording’s noise with the AKG dynamic mic is very low, barely audible on the final mix/podcast, while the background noise of a recording with the ECM999 microphone is way much higher, although this is not related to the audio interface itself, but to the microphone’s internal self-noise.

Worth mentioning that interface’s gain needs to get setup way much higher when a dynamic mic is used, so, theoretically, the internal noise of the audio interface may only be audible when dynamic mics are used, but this really depends on the gain used which depends on mic’s sensitivity. In this particular case, Superlux ECM999 has a sensitivity of 14mV/Pa, while the AKG D5S has 2.6mV/Pa.

ARTA_loopback_INST.png

INSTRUMENT input used, 1 KHz sinewave

ARTA_loopback_inst.png

LINE input used, 1 KHz sinewave

Overall, I find this audio interfaces very good for use in a home/small studio, having a really low internal noise and a very good THD: below -100 dB on INST input and about -85 dB on MIC input (loopback, measured just before the input RED light to show the clipping, with output signal of 2V RMS on the TRS plugs).

L.E.: I've added INST and LINE graphs, separately.
 

Attachments

  • 1569188281427.png
    1569188281427.png
    71.8 KB · Views: 738
Last edited:
OP
trl

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,965
Likes
2,516
Location
Iasi, RO
I did some measurements to the built-in headphones amplifier as well, it's like I've heard it with my headphones yesterday: lacking power, but good enough for monitoring purposes with sensitive cans.


Focusrite_Scarlett_Solo_Gen3-Headamp_infinte_02.jpg

Around 1.65V RMS with no headphones connected
(about the same if using 600 Ohms headphones)


Focusrite_Scarlett_Solo_Gen3-Headamp_30_Ohms.jpg

About 1.15V RMS when 30 Ohms resistors are connected
(a load similar with 32 Ohms headphones)


ARTA-Headphone_amp.png

Headphones amplifier without any load
(about 1V RMS)
From the above graphs I see that the built-in headamp is able to deliver over 1V RMS with a good audio quality (THD+N of 0.004%), so it should pair nicely with any sensitive headphones having an impedance between 32-64 Ohms. In my case I found a good pairing with Audio Technica ATH-M30X and AKG K550, but not with Audio Tehnica ATH-M20X, despite the identical sensitivity and impedance with ATH-M30X (from manufacturer specs). Basically, the sound was cut in half with ATH-M20X vs. the ATH-M30X, not sure why.

Based on http://www.sengpielaudio.com/calculator-InputOutputImpedance.htm I found that output impedance is somewhere close to 12-13 Ohms, so not quite great for headphones having an impedance lower than 32 Ohms.
 

Attachments

  • Focusrite_Scarlett_Solo_Gen3-Headamp_10_Ohms.jpg
    Focusrite_Scarlett_Solo_Gen3-Headamp_10_Ohms.jpg
    163.7 KB · Views: 367
  • Focusrite_Scarlett_Solo_Gen3-Headamp_infinte_01.jpg
    Focusrite_Scarlett_Solo_Gen3-Headamp_infinte_01.jpg
    144.4 KB · Views: 395

mkawa

Addicted to Fun and Learning
Joined
Sep 17, 2019
Messages
788
Likes
695
my guess is that numbers will go up slightly with name brand (nothing fancy, just panasonic polymers) cap swaps. easy and cheap to replace and will definitely help eg supply ripple. a cheap instrumentation power supply over the included brick would also be an easy upgrade.

and huge thanks for the pictures and results!
 
OP
trl

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,965
Likes
2,516
Location
Iasi, RO
Well, numbers might improve a bit (by lowering down a bit, actually) by upgrading the capacitors, but I'm not 110% sure, because inside the Scarlett interface there are no SMPS regulators to cause any noises. However, I need to measure the AC ripple & noise across opamp's voltage rails and if I'll find >5 mV RMS ripple then your recommendation might worth trying. Although the measured THD+N numbers are already very close to ADC/DAC's specs of 100dB: 0.001% on the paper vs. 0.0015% from my measurements (it's actually ADC's + DAC's THD+N, because it's in loopback).

I think that swapping the NJM8065 opamps with OPA1612 or with some other low-noise opamp that works well on +/-5V if there's a rail splitter inside the Scarlett card, or with LME49726 if the opamps are powered directly from USB's +5V.
 

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,156
Location
Tarragona (Spain)
Based on http://www.sengpielaudio.com/calculator-InputOutputImpedance.htm I found that output impedance is somewhere close to 12-13 Ohms, so not quite great for headphones having an impedance lower than 32 Ohms.

http://nwavguy.blogspot.com/2011/02/headphone-amp-impedance.html

FINAL WORDS: Hopefully I’ve made it clear the only way to get consistent performance between headphones and their source is to follow the 1/8th Rule. While some may prefer the sound using a higher output impedance, that’s very specific to each particular headphone, the particular output impedance, and the person’s own subjective tastes. Ideally a new standard should be developed and manufactures should be encouraged to design headphone sources with an output impedance below 2 ohms.

-> Gen3: Headphones with at least 100 Ohms.


BTW, Gen3 has better numbers but the harmonic profile is worse than Gen 2. And without H4 (adds body), but at these low levels ... How it sounds like a DAC?
 
Last edited:
OP
trl

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,965
Likes
2,516
Location
Iasi, RO
[...] How it sounds like a DAC?

I don't think I can spot two DAC's in an A/B test, unless one of them are really sounding different...in a measurable way (background noise maybe). The only one that was sounding a bit different was Burson PLAY when I used V6 Vivid inside (brighter sound, lot of details), but in all other 99% of cases...with a perfect volume match (even the channel balance) I'm not sure I can make a difference with my ears. However, I will try an A/B test and get back to you, but from my perspective this DAC sounds perfectly.

I know about the 1/8 rule, I respect that, but with my AKG K550 I found the sound being OKish, same applies to ATH-M30x (47 Ohms). However, headamp inside is really weak, you can only use it for monitoring, not for audiophile stuff.
 

Fregly

Senior Member
Joined
Dec 12, 2018
Messages
336
Likes
264
I wish they would do a better headphone amp that is usable with everything.
 
OP
trl

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,965
Likes
2,516
Location
Iasi, RO
Not in this price range, but most likely their higher end studio interfaces are having a beefier headamp inside.
 

mkawa

Addicted to Fun and Learning
Joined
Sep 17, 2019
Messages
788
Likes
695
isn't the HP amp meant to be an audio monitor for basic mixing tasks? if so, they don't need to put a lot of power in to drive what would most likely be mdr-v6 in the studio
 
  • Like
Reactions: trl
OP
trl

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,965
Likes
2,516
Location
Iasi, RO
Exactly, monitoring headphones are specially manufactured for such kind of low-powered headamps, so these headphones are usually having:
- a high sensitivity
- low impedance (usually around 32 Ohms)
- all are sealed back

So, Solo Gen3 has a good enough headamp for such sensitive low-impedance sealed headphones.
 

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,156
Location
Tarragona (Spain)
And it is an inexpensive ADC to make good vinyl rips, I think.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,451
Likes
36,880
I just saw this today. Good job on the testing and pictures. Thank you.
 
  • Like
Reactions: trl

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,528
Likes
38,124
Location
Gold Coast, Queensland, Australia
I have a Focusrite Scarlett 2i2 version 2 here and find the headphone amplifier more than adequate my various AKG headphones.
 

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,156
Location
Tarragona (Spain)
[Russian] http://prosound.ixbt.com/interfaces/focusrite-scarlett-solo.shtml

to English:

* https://translate.google.com/translate?sl=ru&tl=en&u=http://prosound.ixbt.com/interfaces/focusrite-scarlett-solo.shtml

* http://www.translatetheweb.com/?from=&to=en&ref=TVert&refd=www.bing.com&dl=es&rr=DC&a=http://prosound.ixbt.com/interfaces/focusrite-scarlett-solo.shtml

Optimized

Focurite-Solo-Gen3-inside.jpg

In the third-generation device, the analog part was redesigned and the wiring was significantly improved - now the signals to the analog controllers do not stretch across the entire board, but are placed on a separate loop. As a result, the passport parameters have grown significantly in terms of signal-to-noise.


asio-latency.jpg

The delay with a buffer size of 40 samples in 44 kHz mode is 5 ms, just like in the previous model. This is one of the lowest latency for USB2.0 interfaces. Moreover, the mode is absolutely working and can be used for recording and playback without problems.


Sound

We directly compared the SOLO generations of Gen 2 and Gen 3. The subjective difference is very insignificant, with Gen 2 the output sound is slightly sharper. In general, both generations have a completely normal sound, suitable for professional work. The changes are small, they are rather evolutionary. The new model [Gen 3] is very slightly preferable to the old...


Conclusions

The new generation of Focusrite Scarlett SOLO 3rd Gen looks more interesting and attractive than the previous generation. The redesigned filling has significantly improved the parameters of the device, and the emerging mode of emulation of branded AIR preamps distinguishes the interface from other budget solutions. Taking into account the balance sheets, you can safely recommend a device for recording and making music in your home studio. Focusrite Scarlett SOLO 3rd Gen features are enough for use with inexpensive microphones and entry-level studio monitors.
 
Last edited:

Guillermo Luijk

New Member
Joined
Dec 22, 2019
Messages
4
Likes
15
Location
Madrid
Superb review, thanks. Purchased the Scarlett Solo 3rd Gen over specific soundcards or DAC's to prevent latency problems when producing music on a DAW (this is a pain on Dell laptops), and do not regret the choice. Great to know performance measurements and inner technical details (specially interested in knowing the specific ADC/DAC IC: CIRRUS LOGIC CS4272).

Regards
 
  • Like
Reactions: trl
Top Bottom