Need some help for a Focusrite Scarlett 2i2 (1st gen) audio interface repair

[AlexP.]

Hi folks!

We had a scarlett 2i2 sound-card crap on us at work and I'm attempting to repair it. When I took it back home I could hear something rattling inside, and oh surprise one of the output caps popped out (the can went off and the dry insides were still standing)

All 4 output caps (two per output) are dead (bulged at the minimum, leaked or can separated for some) Before I blindly attempt to replace them with low ESR ones and potentially losing time and money for nothing, I'd like to identify the cause.

None of the caps were shorted, be it in circuit or after removing them. All voltages rails seem correct, the card is drawing 100mA when plugged on a 5V supply. Turns on with the caps removed so far so good... but...

When I measure the voltage where the caps should be, I got a negative voltage on all four. Something between -50 and -80 mV, relatively stable reading but slightly different for all four. Seems abnormal right and could explain why they exploded?

I probed all other caps and especially all ceramic ones, found nothing odd. But there are those 4 terminal ones which I believe are ceramic filters (X2Y decoupling capacitors according to wikipedia), I'm not used to seeing those often if at all in my day to day job. They beep in continuity mode between each of their two terminals pairs (they are laid in a cross pattern), but read in circuit something sub-ohm (0.3 maybe but can't trust that quick measurement) maybe it's perfectly normal maybe it's not?

So here I am at the moment. I have not ordered replacement caps yet, and I would prefer to be sure that the source of the problem is gone (or that they just died due to aging) before attempting a replacement. Any help or hints would be greatly appreciated!

[Swake]

How are those output caps related to the output? How are these things connected?

An audio line level signal shouldn't be enough to pop a can of a cap. Sounds more like abuse from wrongly connecting it on the output of an amp or a phantom power supply or something similar.

What is driving those outputs? OpAmps maybe? Are these ok?

[Audiorepair]

It is very common that output electrolytic capacitors fail.

Replace them with the highest spec ones that will fit.

Trying to determine the cause of failure is futile.

[trobbins]

Trying to determine the cause of failure is futile.

The failure cause may well be identifiable, as it is one company, and likely connected to one item of equipment, so there could be a dc bias on that equipment (of a bad polarity for the cap).

[David_AVD]

Failed output capacitors in audio interfaces is often a result of phantom power (usually 48V DC) being present on the inputs it was plugged into.

[Audiorepair]

Failed output capacitors in audio interfaces is often a result of phantom power (usually 48V DC) being present on the inputs it was plugged into.

It can also be due to being the output capacitors were shit to begin with.

[AlexP.]

First of all, thank you all for your answers, your help is much appreciated!

I'm sorry for the slow reply, I got sucked in a work wormhole, and since this repair is not urgent it was put on the back-burner.

I took a few pictures. On the first two (overview and caps) you can see how the crapped caps looked like when I removed them as well as an overview of the board region.

Then the interesting ones, two for each pair. Both come out from MAX4477 opamps as expected. C134 and C137 then go to C100 and C102 of the second picture (the ceramic filters I talked about) which go directly to the output jack socket. For pair 2 everything arrives from the left of the second picture which is 90° rotated compared to the 3 others, but same deal ends in C99 and C104 then down to the jack below.

Meanwhile I met some guy well versed in audio gear repair during some gig, and he told me to replace the electrolytics by film ones (non polarized) as it could eliminate the issue with the negative voltage I'm measuring altogether (as well as improving on the surely dodgy factory ones)

He also told me to inject or play 1kHz sine waves and check if the opamps outputs look correct or not, which I still have to do.

To answer some of your questions: I don't think it was a phantom power issue, but during the last operation where this card was used, it was receiving LTC timecode from a console in order to slave some video playback for the show. Both were plugged on the same outlets in the same technical area, but there were tons of other stage equipment (pyrotechnics, lasers, sound and lighting etc) and the driver froze down mid-show as well as our video playback as you can imagine (very embarrassing moment) so we can't indeed exclude that some potential difference or other bad stuff happened via the inputs.

Technically the outputs were not in use (not plugged to anything) but since it was using the ASIO driver, you always have both input and output active even if you feed the output buffer with zeroes.