#37 Presonus Audiobox 1818VSL teardown and troubleshooting (part 1)

Together with the pair of Konnekt x32's (one of which is dealt with in the previous write-up here), i also got one of these Presonus Audiobox 1818VSL's. Also sold as untested, but powered up and allegedly detected by a Mac, so i figured it couldn't have been too far gone (or fubar'd).

The reality of it turned out to be a wee bit more complicated than one might've hoped...


I went for this because for my home setup, i really could've used the more comprehensive I/O (ADAT and more analogs) than my Focusrite Scarlett 18i6 provided.

While the thing did indeed power up, and got detected by the computer and its control panel, the latter wouldn't really let me do anything without a mandatory and unescapable demand to update the firmware. The unit says it has v1.18; the latest version of the control software wanted to update that to 1.49 - fair enough. Except that after a minute or so, it threw a "failed with Error 2001" at me.

M'kay... In some cases, laptops or other devices have been known to not want to update straight to the latest firmware version all the way from a very old one, but require one or two intermediate version "steps". With this in mind, i downloaded and installed the earliest software i could find on the Presonus website. Sadly, "even" that said it needed to update the firmware to v1.48, and subsequently failed, but with "Error 2011". Really helpful...

After opening a ticket on the Presonus support page, a rep (likely as part of a somewhat "canned" response, after asking for an msinfo32 export of a system summary) at least gave me a link to an even older control panel. Surprisingly enough, that actually seemed to work. I even managed to get the meters corresponding to the first output pair from Windows to show some activity (upon playing some YouTube videos), but alas, no actual audio output, neither on the main outs nor the headphone outs. Fudgesicles... Ah well, just sitting and twiddling thumbs never got anything done, has it?

With that said, it's time for a little peek under the skirt, woudln't you say? Right off the bat, interesting and pretty hefty construction - thick cast aluminium side-plates (containing the rack ears as well), and the faceplate is an extruded aluminium U-channel. The top, bottom and rear of the case is a one-piece stamped and bent steel(?) sheet.

And what awaits us inside? Well, apart from a "secondary" front panel (to which the XLR's and all the potentiometers are fastened to, not a terrible whole lot, as it turns out. Compared to stuff like the M-Audio Profire 2626 (and, i suspect, the Fast Track Ultra 8R, which likely shares the big analog board), this thing is half-barren.

Almost nothing on the bottom - some bypass capacitors, a few resistors and that's about it.

Oh, but what's this? Bloated and burst mid-tier (Jamicon WL - i suppose they were at least kinda-sorta trying) electrolytics in a switchmode power supply? How could this be?

And hello - someone's definitely been (quite deep) in this before. Oddly(?) enough, right on the flash chip... The puzzle pieces are kiiiiiinda-sorta starting to fall in place, right about now.

I wonder what're the odds the firmware update might actually go through, after a re-capping of the power supply. I'm somewhat skeptical about it being quite THAT easy, but i guess only time will tell. Failing that, i wonder if Presonus support might actually be willing to send me a firmware dump of the chip's contents, in case this reworked(?) one turns out to be somewhat dud.

But since it's all "disrobed" already, might as well do the inventory and eventually do some reverse-engineering. They've been bragging about these "XMAX class-A preamps" in their interfaces for years, and i've been curious just what's so uber-mega-special about'em anyway.

Quite some fiddling about (16 screws fastening the XLR sockets, plus 10 potentiometer nuts and washers), but the plastic front panel's off too. Pretty comprehensive RFI-shielding, though - the piece of copper foil below the headphone output was soldered to the "sleeve" pins of the jack, on the underside of the board. The XLR sockets have a pair of little barbs near the bottom, that poke into the big shielding foil.

Power supply section first [insert shudders here]. Not a lot of "chineesium" to speak of. Two OnSemi SS26 (2A 60V) Schottky diodes in series with the DC input (anti-reverse protection), one towards each "side" of the power supply. The (quite sizeable) transformer is the coupled inductor from the flyback supply for the analog, and the smaller inductor is for the digital side.

The former is handled by a TI TPS40210 controller, driving a Vishay / Siliconix SI4480DY N-channel MOSFET. The latter is run by a Fairchild Semiconductor (now OnSemi) FAN8303 step-down regulator (with integrated switching MOSFET).

The analog secondary is rectified by a(nother) pair of OnSemi SS26 Schottky diodes. Since the marketing blurb said something about the XMAX preamps running on "30V rails", my guess is these are putting out a plain old +/-15V, pretty much the standard for most opamps. Coincidence? You decide...

Regarding the output of the regulator for the digital side, i've got a hunch that's putting out 5V, and this DPAK doohickey with a sizeable heatsink-copper-plane and marked with a "330" is further regulating that down to the 3.3V that the faster digital circuitry needs. The markings ("45" and "330") are a bit too vague to find the actual datasheet for the part, but the style of them kinda-sorta rings a bell.

Running the whole show is an Xmos XS1-G04B-FB144. I think. It's the closest i could find that pretty much corresponds with their uselessly cryptic markings. The 32-core setup would indeed make quite some sense, given the pretty extensive DSP capabilities this little box has.

A quartet of Cirrus Logic CS4272 stereo CODECs run all the analog ins and outs.

A tiny little SMSC (now Microchip) USB3318 handles the USB interface for the Xmos chip.

Save for a lone TL082C that's used for the instrument inputs, all the other opamps used are quad ST MC33079's.

And now we get to the fabled(?) XMAX preamps. Each channel has a pair of 2N4403 PNP transistors, followed by a pair of "2C"-marked BC849C NPN's. This is all starting to look strikingly familiar - a pretty "classic" and very widely-used microphone preamp circuit. You'll find it in loads of Mackie, Soundcraft, Behringer, you-name-it mixers. I don't think i'll even bother actually reverse-engineering this - apart from potential variation in resistor values, it's nothing to write home about, really.

Yet more MC33079's acting as (most likely) output low-pass filters for the DACs.

Yet a few more driving the outputs 3-8. 1-2 are mirrorring the "Main Out" pair, with the difference that the "Main Out" goes through the volume pot on the front.

Aaaaaaand that's about that for the teardown and/or inspection. Now then... I'm thinking a quick voltage check, after which i think i'll try another firmware update, but while closely monitoring the main digital supply voltage. Just curious if it flakes out or something. I'm willing to bet money the ripple's all over the place anyway, with those bloated / burst electrolytics there...

Well, i was close... Sort of. The little FAN8303 buck regulator actually handles the 1V rail for the Xmos core voltage. I was right on the money with the 3.3V regulator though. Couldn't see any noticeable blip in the core voltage during a firmware update attempt, though.

Just for the hell of it, i read the contents of that flux-drenched Atmel AT25DF041A, then found a Winbond W25X40BV on a dead laptop mainboard, and wrote the "stock" data into it. No huge surprise, unit still works the same way, in the sense that a firmware update still fails. I really hope Presonus can provide me with a (known-working) dump of the contents of the flash memory, i'm more and more positive that's the main issue with this. Well, that and the bloated caps, obviously.

[To be continued...]