#63 M-Audio M-Track 2x2M C-series teardown & repair

Boy, that's a lot of M's up there... But yeah, here's another eBay score, bought as faulty. Basically, because it was too cheap to pass up, even if it's unrepairable - a paltry 30eu. The ad said it will no longer power up, save for the phantom-power indicator LED, and won't be detected by a computer. But before troubleshooting, it warrants a thorough inspection, inside and out, wouldn't you say?

I must say, the construction seems / feel a bit more substantial than i expected. The unit is a bit more massive than i (would've) thought, anyway. They might've drawn some inspiration from Mackie's "Big Knob" line-up - not that there's anything wrong with that. 

"Business as usual" around the back as well, although as the "C-series" on the front / top might've hinted, the USB connection is a Type-C this time around. No-name combo XLR's though...

Right, off comes the bottom cover (made out of ~1mm thick steel, which gives the unit a decent part of its weight and sturdy feel), and... Not a whole lot is revealed. Maybe except that fat electrolytic capacitor, sitting in its own little cutout in the board.

Two-tier construction, the two boards being linked through a multi-way ribbon-cable. Some might argue, excessively long, but on the other hand, i suppose that facilitates easier production-testing?

 Bottom board mechanically removed, and we get a first look at what makes (or rather, used to make) this tick (digital clock puns, anyone?). Probably not shocking to see an Xmos controller, on one hand, and on the other hand, not a whole lot of much else. Then again, how much circuitry could, would, or should you cram into a diddy little 2in/2out interface, anyway?

And now, a closer look at the main board in this thing. Reasonably well laid out, plenty of space... And yet, i'm puzzled by the choice of locating what looks like the phantom-power boost-converter, so close to the (what should be) high-impedance instrument inputs, which would be that much more prone to noise pick-up. But hey, what do i know? I don't work at M-Audio...

The brains of the operation, according to the "6U6C5" marking, is an Xmos XS1-U6A-64-FB96. 

 

"XS1-U6A-64-FB96 devices include two DC-DC buck converters which can be con-figured to take input voltages between 3.3-5V power supply and output circuitvoltages (nominally 1.8V and 1.0V) required by the analog peripherals and digitalnode." That explains the two 4.7uH inductors in the vicinity, as well as the lack of any 1V regulator in the area (as seen in other Xmos-equipped interfaces). 

 

Next to it there's a Macronix MX25L4006E 4Mbit (512KByte) flash memory chip. In-between the two audio clock crystals (24.576MHz being a multiple of 48kHz, and 22.5792 being a multiple of 44.1kHz) there's a Texas Instruments SN74LVC3G04 triple inverter gate, as well as a Nexperia 74LVC2G240 dual inverting buffer. A Cirrus Logic CS4270 handles the audio conversion in both directions, while the OnSemi MC33078 dual opamp next to it acts as the low-pass filter for the analog output. Next to that, an OnSemi NCP2811A (adjustable gain version) is employed to drive the headphone output. An interesting choice, when in so many other interfaces, plain opamps would be used (at best, with some discrete transistors for added current capability).

On to the power section. More or less, anyway. Right next to the USB-C connector there's a FairchildSemi (now OnSemi) FSUSB30, a two-port USB switch. I'd reckon that's to account for the reversible nature of the Type-C connector, "both" of the USB2 data line pairs going into this chip, and it selecting between them, whichever is active. The DPAK-packaged component is some version of a linear 3.3V regulator, going by the "330" marking. There's also an LM393 dual comparator right in the corner, but i'm not quite sure what exactly that could be used for.

Apart from a couple (single and dual) transistors, given away by the "Q_" board designations, U11 is, as far as i can figure, a Texas Instruments LP2958-N low-noise linear regulator, the 4V fixed version.

The 5-pin "PFNI"-marked part is a Texas Instruments TPS60403 charge-pump, for creating the negative analog rail for powering the opamps.

The phantom-power is handled by a National Semiconductor (now Texas Instruments) LM3578A switching converter. Next to it is another MC33078 for buffering the instrument inputs.

Moving on to the top board - most of the analog circuitry seems to be located here. The two mic preamps (on the right side of the photo), kinda-sorta look like the "classic"(?) discrete-plus-opamp design, often encountered in small-to-medium-size mixers; Presonus calls them "XMAX". "North" of those is a pair of LM339 quad comparators, one for each input channel's level indication. The NJM4558 and the two diodes in the vicinity, going by the traces snaking towards the aforementioned comparators, is likely the precision rectifier, to provide the "DC" signal that mimics the audio level (comparators not really working in AC).

Interestingly enough, they were kind(?) enough to actually list the pinout of the ZIF connector / ribbon cable that runs between the two boards, on the board silkscreen itself. Between that and the traces, it really helps figuring out what's what. With the aid of that, my hunch is reinforced, that the two relays have something to do with connecting the (buffered) instrument inputs to the preamps, given the "INS1(SNS)" and "INS2(SNS)" signals. I haven't read the manuals, but i've got a strong suspicion that plugging a jack into an instrument input overrides whatever's plugged into the respective combo-XLR on the back. Other interfaces sometimes offer a manual switch between either line or instrument (if there's no separate 1/4" instrument input), or between the front and rear inputs, etc.

The two MC33078's near said pinout diagram are part of the "USB / Direct" blend circuitry. Not quite sure what the one near the two 220uF electrolytics does exactly, though. Perhaps also part of the instrument input circuitry? Your guess is as good as mine, but at this point, i'm not quite THAT curious.

Right then - after hacking together a little adapter, with a couple wires and a USB socket, i powered the thing up from my bench supply set to 5.1V and limited to 500-odd mA. Flick the switch, straight into current-limiting, dragging the voltage down to 2.8V or so. And the Xmos chip gets kiiiinda warm. Upped the limit to 1A, still hitting it, and the chip is getting even hotter. Welp, time for a new processor then, i suppose... 

[Later update] Yup, replacing the Xmos did indeed do the trick!