FIREPOD S/PDIF Problem SOLVED!

dgatwood

is out. Leave a message.
As a bit of background, my first FIREPOD really did have dead S/PDIF. It couldn't lock to anything, including the Behringer in standalone mode, so that was definitely a hardware failure. It never worked after I performed a firmware upgrade, and the reversioner didn't help. However, it seems my second FIREPOD actually works... well... mostly.

That issue notwithstanding, here's how I fixed my current FIREPOD's S/PDIF input issues....

So I started to wonder about the FW1814 when the second FIREPOD could sync to my Behringer as long as the Behringer wasn't repeating sync from the FW1814. That was a bit of a red flag.

When I tried to make the FW1814 work with my pair of 8Pre interfaces (which are beautiful, by the way---just freaking beautiful), it wouldn't lock to them and vice-versa, so at that point, I concluded that the FIREPOD might be okay. Now this is particularly bad because 96 kHz ADAT is really only running at 48 kHz, splitting each audio channel in half and sending alternate bytes down each of two audio channels at half the sample rate. So this means that I was seeing sync problems at a mere 48 kHz.... This was worse than even the S/PDIF situation with the FIREPOD.

The piece of advice that ended up helping me find the bug was actually a somewhat erroneous piece of advice: the M-Audio tech support person said that the two interfaces should generally be used on separate computers... but that got me rearranging FireWire cables....

After lots of experimentation and testing ADAT optical between multiple computers (and never being able to reproduce the failure except when they were both connected to my G5 tower), I finally discovered that the M-Audio FW1814 is hyper-sensitive to FireWire power line bus loading and/or noise. Basically, its S/PDIF and ADAT hardware are unstable unless it is the very first device on the bus.

Note: the PowerMac G5 puts out tons of juice on its FireWire line (way beyond the required voltage), so I suspect that some engineer cost cut a critical filter capacitor in the FW1814's supply circuit, but maybe that's just me being cynical. :D

Ironically, the reason I had the M-Audio interface at the end of the chain was because they screwed up the hardware design and powered the FireWire PHY from the switched power supply instead of from the bus lines (or, ideally, with diode protection and power from both of those sources). The result is that if it is in the middle of the chain, when you switch it off, the devices beyond it go away, too. *grrrrrr*

Anyway, at this point, the M-Audio interface was slaved to the 8Pre interfaces. Just on a long shot, I decided to patch its output to the FIREPOD by S/PDIF coaxial and... it locked to it (after several seconds of soul searching). The FIREPOD did lose sync at one point (after several minutes of solid blue), so I have a feeling the FW1814 is right on the fringes of working right now, power-wise. This is supported by how hard the FIREPOD had to work to lock to the signal. I'll get a powered hub tomorrow morning and put the FW1814 on its own port. I think it should be solid from now on with those changes....

What's odd about this is that bus loading couldn't have been the problem with the FW1814 in my previous tests with the FIREPOD. I couldn't get the FIREPOD to sync to the M-Audio FW1814 even when two separate computers were used, which should completely rule out the additional bus power jitter as the cause for that issue. At least I'm pretty sure I tried that with this replacement FIREPOD....

So apparently, assuming I'm remembering my previous testing correctly, even when bus power is clean enough, the FW1814's internal sync generator isn't stable enough for the FIREPOD to lock to it at higher speeds. This is arguably a bug in both devices---the FW1814 for generating poor quality S/PDIF output and the FIREPOD for not being more tolerant of jitter in its S/PDIF input.

In any case, I have a MOTU 8Pre chained by ADAT lightpipe to a second 8Pre, which in turn is chained by ADAT lightpipe to an M-Audio FW1814, which in turn is chained by coaxial S/PDIF to the FIREPOD. With sync coming from the MOTU and the FW1814 properly powered, I show solid sync across all four interfaces, and I'm utterly rolling on the floor laughing at the realization that I probably could have had the FIREPOD and M-Audio interfaces working together several weeks ago if I had bought a powered hub for my audio rack like I originally intended.

Regardless, tonight, I finally have -everything- working. My PowerMac G5 can currently see 32 analog inputs and 16 analog outputs of simultaneous I/O, at least in theory, assuming nothing chokes software-wise. Wow. If this stuff actually works, I could just about mic every pair of players in our entire freaking wind ensemble. :D

I'm pretty sure I've solved the great mystery, though, and the answer to the question of which interface was to blame... they BOTH WERE!

I've dropped follow-up notes to both M-Audio and Presonus to let them know what worked and what didn't. Hopefully, Presonus will fix syncing in their firmware, and hopefully M-Audio will fix a whole host of hardware design flaws....

So here's what we learned today:

* The M-Audio FW1814 PHY is miswired and won't pass data with the device turned off. *Aaaaaagh!*
* The M-Audio FW1814 sync isn't as stable as it should be (or at least mine isn't). *Grrrrrr*
* The Presonus FIREPOD is unusually picky about the stability of the S/PDIF signal. *&^&^@^$*
* Upgrading the firmware on a FIREPOD will likely kill your S/PDIF ports (see note at top). *You have been warned.*
* The M-Audio FW1814 optical ports go severely unstable (unable to reliably handle what amounts to a 48 kHz optical signal) unless it is the first device on the chain. Indeed, it should be the only device on a port if at all possible. *WTF?*
* The pair of 8Pre interfaces still work flawlessly, both with each other and as a stable sync reference. *:D*
* The 8Pre interfaces are just about the first pieces of hardware I've owned in several years where I haven't found at least one significant hardware design flaw nearly out of the box.... *Congratulations, MOTU*

Any questions?
 
I'm really interested in that Motu 8Pre...

I was never a fan of the Motu pres (owned an 896), though they were fairly usable. Are the pre's on the 8Pre along the lines of the other Motu stuff, or perhaps a little improved. Seeing as they are "kinda" touting the ability to have the 8 pres and analog to optical out, I'm wondering about the quaility. Don't think I'd go that route for just 8 pres from Motu, but as an audio interface, maybe it's a pretty decent offering?
 
I haven't had a chance to really A/B the pres with anything yet.

As a follow-up, the FW1814 blues continue. I tried a powered FireWire hub. In theory, the FW1814 should have full bus power at this point. It still isn't happy. It really does want to be on its own FireWire bus without any other traffic on the bus at all. I've never seen a piece of FireWire gear that is so thoroughly misbehaved before. Frankly, I think we need to give M-Audio a bad mark on its report card. It should include the words "Does not play well with others."

This morning, I talked to someone at work who owns a FW410. He, in turn, knows a third person at work with an M-Audio FireWire interface of unspecified model. All of us have one thing in common: our M-Audio FireWire interfaces all behave badly if they share a bus with other FireWire devices, experiencing all sorts of fascinating failures.

At this point, since I have a workaround in mind that's guaranteed to work, it's not worth my time to try to figure out what's wrong inside the FW1814. Either way, it's something folks should be aware of if they're looking for a FireWire interface and intend to use the ADAT connections.

So the guaranteed workaround: I'm going to return the FireWire hub to Fry's tomorrow, and I have a PCI Express FireWire card on order. Since the FW1814 is so incredibly flaky, I'll keep it attached to my computer's FireWire port (where it works as long as it doesn't have to be sociable) and hook everything else to the card. Then, I get to decide whether to sell the FIREPOD or whether there are any reasonable uses for having 32 audio inputs. :D
 
Back
Top