#72 MOTU 828 mk3 teardown

Yet another eBay score - caught this one with a blown power supply. Blown, as in, skidmark inside the casing, circuit board scorched, tracks vaporized. But maybe i'll make that into another article on its own, someday. In the mean time, let's take a peek under the hood of one of these arguable "classics".

Now, as you may or may not have guessed from the name, there's obviously(?) more than just a passing similarity between this and the older 828 mk2. The insides are a wee bit of a different story, though. But then again, given all the extra features, you'd sort of expect that.

Yeah, there's quite a bit of stuff in there. It's sort of much, but at the same time, it's not even all that crammed full. 

The power supply is the same we've seen in the 8Pre, the 896mk3 Hybrid and the 8M. "If it ain't broke, don't fix it", although in this case, it pretty much was. But i swapped it out with the one from a comatose 8Pre i had, due to not wanting to waste time, so there.

Due to the construction of the thing, and in the interest of logic(?), we'll take a quick peek at the piggyback-daughterboard, that carries the ADAT connections as well as the line outputs and the SMPTE timecode in/out.

At one end, there's an AKM AK4358 8-channel DAC, as well as as a FairchildSemi 74VHC04, probably buffering some clock signals. Output low-pass filters are JRC NJM4580's, and there's a tiny voltage regulator, likely powering the DAC; probably a Texas Instruments TPS73001, but the photo is a bit too blurry to make out. 

Output drivers are several more NJM4580's.

The view doesn't change much upon removing the daughterboard, but at least now there's nothing in the way anymore.

Orchestrating the entire operation is an Atmel / Microchip AT91SAM7S256 microcontroller. You'll be forgiven for the sensation of deja-vu, but we have indeed encountered the same chip inside the recently-inspected Audio Express. And in the 896mk3 Hybrid as well, in fact. The firmware is (mostly) contained in the neighbouring Atmel / Microchip AT45DB161 16Mbit / 2Mbyte flash memory.

The heavy lifting is done by two big fat chips, this time around. On one hand, a Xilinx Spartan-3E XC3S500E FPGA, and on the other hand, a Texas Instruments TMS320C6722 DSP. The latter also employs a Samsung K4S641632K 64Mbit / 8Mbyte SDRAM chip - i would guess, mainly for the reverb. The FPGA's core is powered by, i'm suspecting, another TPS73001, but don't quote me on that. Up "north" you can just see the "classic" Texas Instruments TSB41AB2 Firewire PHY.

Also, interestingly enough, in the shot above we can see a couple unpopulated connector footprints near the rear edge of the board, that look remarkably similar to an Ethernet socket, as well as a USB-B socket. Even though the Hybrid version has Firewire 800 connectors... In any case, intriguing.

Also, off to the side, there's a clocking solution i hadn't seen before - a dedicated Analog Devices AD9834 direct digital synthesis chip.

Not a whole lot else regarding power supply circuitry on the main board, apart from the phantom power supply being handled by a Texas Instruments TPS61040 boost converter, complete with a voltage multiplier tacked onto the output (yet again). In the near vicinity are a pair of Analog Devices ADM709 voltage supervisors, aka reset controllers.

On to the audio circuitry - the two microphone / instrument preamplifiers are the nearly-ubiquitous Texas Instruments PGA2500's, aided by a few ST Microelectronics TS922 opamps. Over to the side, a pair of NJM4580 drive the two headphone outputs, and something like an LM7905(?) regulates the -5V rail for the preamps.

An AKM AK4396 stereo DAC is in charge of the main stereo output, along with a few NJM4580's on the way.

On the analog-to-digital conversion side, we have two Cirrus Logic chips, a CS5368 and a CS5364, an 8-channel and a 4-channel converter, respectively. All 12 channels seem to be in use, although i'm not 10000% sure what the last pair is used for, really. Betwen eight line inputs and the two mic/instrument inputs... Regardless, a veritable army of TS922's assists them.

A bit further "downstream" there's a Cirrus Logic CS3308 8-channel volume control, with some further NJM4580's driving the outputs. Also, a Texas Instruments TPS79850 regulates the 5V rail for the converters.

And last but by no means least, an AKM AK4382A DAC handles the second (independently assignable) headphone output, with a couple of more NJM4580's after it.

And that just about wraps this up. Testing all the ins & outs revealed no issues, so the toasted power supply was all that was wrong with this.