#89 MOTU Ultralite AVB teardown

Slight side-note to begin with - just a couple days before this write-up, this blog has reached no less than 500.000 views. That's half a million(!!!), quite mindblowing for this niche of a collection of topics, and the frequency (or lack thereof, rather) of postings. I'd though just like to take this chance to thank all of you who stumbled upon this little corner of the Internet, and i hope you found useful information, if not even learned a thing or two. Onwards and upwards!..

I had been eyeing one of these to act as my "daily driver" desktop interface for some time now, but couldn't justify shelling out 4-500 bucks for one of these, just on a whim, as it were. But when i was randomly browsing through Reverb (in the US) for one of these, as luck(?) would have it, this was listed for merely half that, by virtue of a broken-off encoder. Gimme!..

Arguably even moreso than its ancestor, this little bugger really is chock-full of features. Nevermind the two mic inputs and two separate instrument inputs (usually they're combined), there are also six other line inputs, eight line outputs, MIDI in/out, ADAT in/out and, as the cherry on top and one big reason i was gunning for one of these, AVB connectivity. The latter will very conveniently provide access to all the channels of the MOTU 8M in the recording rack in the studio downstairs.  

Popping the hood (MOTU's "classic" cast aluminium, which they seem to have given up on, on their newer releases) we're met with a populated board, although arguably almost sparse, given the feature set.   

First one to come off is the main processing daughterboard. Just as well, it likely requires more than the two (or maybe four?) layers of the main board. The brains of the operation is a Texas Instruments OMAP-L138, assisted by an Alliance AS4C32M16MD1-5BCN 512Mbit / 64Mbyte LPDDR SDRAM, while the firmware is stored in an Infineon / Cypress S25FL256S flash memory chip. The "heavy lifting" though (audio processing & routing) is done by an AMD / Xilinx XC6SLX16-3FT256I Spartan-6 FPGA.  

Nothing terribly exciting on the back, apart from the Microchip / SMSC USB3340 interface chip, and a Cirrus Logic CS2100 PLL clock generator.

Which then brings us to the uncovered mainboard. Yep, quite a bunch of stuff on there, including some (physically) pretty sizeable (likely) DACs. But let's take a closer look at all that then, shall we? 

Most supply rails are handled by only two (albeit complex-ish) chips, both from Texas Instruments - a TPS65580 for the analog rails (+/-5V and the 48V phantom power) and a TPS65581 for the digital side of things (the back of the processor board shows labels for 3.3V, 1.8V and 1.2V, at least). The lone 5-pin linear regulator nearby is a Nisshinbo NJM2846LD3 5V regulator, most likely supplying the analog side of the audio converters.

The one other power supply chip is a National Semiconductor / Texas Instruments LM2990S negative 5V linear regulator. 

 

Before moving on to the audio stuff, the tiny SMSC / Microchip LAN8720A Ethernet transceiver is worth mentioning. "Just" Fast Ethernet though, not even Gigabit, interestingly enough. 

 Now then, time for the fun stuff(?) - the two mic inputs are handled by a pair of Texas Instruments / Burr-Brown PGA2505.

The two instrument inputs are buffered by a Texas Instruments OPA1652 dual opamp, and then fed into a Cirrus Logic CS5346 ADC. Interestingly enough though, it's a stereo ADC, but with six(!!!) selectable stereo input pairs - go figure. They must've gotten a really good deal on these chips, otherwise surely they would've gone with a simpler solution, although they're likely using the internal PGA (programmable-gain amplifier; in this case, up to +/-12dB of gain or attenuation) for input level setting...

The six line inputs get fed into a Cirrus Logic CS3308 8-channel volume control (with a bunch of ST Microelectronics TS922's as buffers), and then along with the two mic inputs, into a Cirrus Logic CS5368 8-channel ADC.

 

The headphone outputs have an AKM AK4382A stereo DAC, and a pair of Rohm BA4580R dual opamps - albeit in two different packages. Yet the smaller one seems to connect straight to the output jack, so... I'm not quite sure what's going on there, but oh well...

The eight line outputs are fed by a pair of AKM AK4413 4-channel DACs, and more BA4580R's. And in the vicinity, a lone Texas Instruments PCF8574 I2C expander resides.

 

Aaaand that just about wraps this one up. I'll replace the encoder at some point (or all five, for consistency), but for now, this'll do. Looking forward to use this for audio measurements too - hooray for digitally-controlled analog levels that are repeatable!