#83 Focusrite Scarlett 18i8 gen2 teardown & repair

I know it's been quite a while, but i haven't given this up! Life just gets in the way, i'm sure you know how it is... Regardless, this time we'll be taking a dive into yet another "as faulty" acquisition - an utterly pristine-looking Focusrite Scarlett 18i8 gen2.

Feature-wise, at least on the hardware side, it's virtually identical to its gen1 "older sibling", so no news there.

Nothing terribly startling to see after the insides are liberated from the aluminium enclosure either. Only the assembly method differs a bit, hence the new metal plate running along the rear panel.

No surprise to find an Xmos controller in the middle of it all - this time, an Xmos XU210-256-TQ128. At least that's my best guess, given the "new" cryptic "XFR001C" marking, which says nothing. Not unlike the 2i4 gen2, but the Focusrite spares department were kind enough to confirm the actual part number in that. This XU210 is the smallest 128-pin part with no built-in flash memory (because there's the 8-pin chip next to it). Speaking of flash memory, the firmware is stored in a Macronix MX25L8006E 8Mbit/1Mbyte memory, while the clock generator is once again a Cirrus Logic CS2100 (marking "2PCC").

Converters are the now-familiar Cirrus Logic CS4272.

The opamps also offer a healthy dose of deja-vu - one NJM2122 per mic/line input, the rest being NJM4565's.

Well, except for the headphone outputs, driven by a pair of "classic" NJM4560's.

Apart from a lone pair of Samxons, the rest of the electrolytic capacitors are an assorment of CapXon and Jamicon (i'm guessing, going by the series-name-in-a-box marking on those smallest ones).

Moving on to the power circuitry, it only took me about an hour to track down just what the hell the 3.3V converter could be, and only after manually trawling through a couple dozen datasheets on Mouser, did i finally manage to stumble upon it - a Monolithic Power Systems MP2227 synchronous buck converter. Really odd that, especially since it's not even an obscure far-eastern manufacturer, at least this particular marking showed up in no searches i could figure out. 

Next, a bog-standard ST Microelectronics L7805 regulates the 5V rail (likely only the glue-logic and possibly the analog supply for the converters).

And last but not least, a pair of Monolithic Power Systems MP3425 power the +/6.9V analog supply rails, and the 48V phantom power rail, respectively.

Right then, with the teardown & documentation out of the way, up at the top i mentioned i had bought this "as faulty". The problem being that, upo power-up, all the LEDs would flicker continuously at a rate of ~10Hz or something, quite fast-ish, and of course, the device would not get detected by the computer.

Power issue, clearly. Especially since one could hear some internal ticking, at the same pace as the LEDs flashing. Very well then; what about a plugpack different than its own stock one?

Drum roll, please...

Yup, as you might've guessed, powered up fine, got detected by the driver, no problems. Well, apart from first needing to update the Scarlett MixControl package on my test-laptop, that is.

A few well-placed but "delicate" hammer blows later... Can you spot anything wrong in this picture? Here, i'll give you an even better angle.

Surprise-surprise, the two "KSG"-brand electrolytics on the output, gave up the ghost. Needless to say, after being replaced with a couple of 680uF / 16V Rubycon YXF's, it once again works just fine. But since we cracked it open (quite literally) anyway...

Pretty tidy otherwise, clean soldering, no humongous red flags to speak of.

Only noteworthy(?) thing on the bottom being the main controller. Markings say "Z10G23", but i couldn't tell you what it could be. Even if it blew, i doubt i'd bother trying to track down a replacement for the chip, instead of just getting another plug-pack.

Later edit: A kind reader has just informed me the PWM chip is likely an On-Bright Elec OBGZ10MP.

 

And there you have it, first post in two and a quarter years! You may or may not want to keep an eye out for more stuff later in the year - i might "even" do a teardown on the 18i20 gen2 that's been my "daily driver" for the last couple of years, plus there's some microphone-related stuff "in the oven"...