Digidesign 192 cutting out.

[CalavoBob]

I've got a protools 10 HD system running off a Mac pro 5.1 into 2 Digidesign 192s which are loop synced to a Digidesign Sync. The first 192 is loop master and has 2 analog input cards, one analog output card and one digital output card. The second 192 has one analog input card, one digital input card, and two analog output cards.The second 192 is fed a short DigiLink cable from the first 192's extender port to its DigiLink input. This system ran great for several years until about 6 months ago when after about an hour or 2, the meters on the second 192 would begin to flash intermittently. Usually it's all 16 output meters that flash, but sometimes all the meters start flashing intermittently. I head a srackle, and then the outputs on the second 192 all stop working and remain dead until I reboot the system. I swapped the short DigiLink expansion cable, and even replaced the entire second 192 with another identical unit (with all different i/o cards), but the problem is still there. Thinking that it might be heat related, I separated the two 192 units, put perforated rack panels above and below them, and added a muffin fan between them in the rack box, but it still craps out after a few minutes. Has anyone seen this before? Any ideas? Here's a video of the whacky meters.

 

[CalavoBob]

When this happens, the first 192 still works fine, but it only leaves me with 8 outputs to work with.

 

[ccaudle]

after about an hour or 2, the meters on the second 192 would begin to flash intermittently.

Have you swapped the two 192 units to see if the extender port output on the first unit could be bad?

 

[kags]

Do you get any error messages when the interface stops passing audio? Have you run Digitest? It could be the HD card. Have you trashed prefs and/or reinstalled the HD drivers? Cleaned and reseated PCIe cards? Has anything in your setup changed- like new drivers or apps like UAD updates or ??

I’ve even seen bad usb cables on hard drives cause ProTools to show -6001 clock errors.

Do you have a disk image of your system drive from when everything worked fine? If so, boot from that.

192s are getting pretty old and run warm, so maybe also check the power supplies or even the analog expansion cards. Used 192s sell for almost nothing, maybe buy one to test.

These issues are frustrating and usually result from some seemingly unrelated.

 

[CalavoBob]

you swapped the two 192 units to see if the extender port output on the first unit could be bad?

I haven't tried swapping out the first unit. That's a good idea to doubt the extender port on the first unit.
Kags came up with a number of good ideas too. I hadn't changed the configuration for quite sometime when this happened though. Oddly enough it did not cut out yesterday or today. I'll have to try some mixing on it tomorrow to see if it still screws up. I cleared my head out this evening by tuning up a nice analog 24 tk. It was a swell machine. Satisfying to work on.

 

[kags]

Preferences and drivers can get corrupted over time. Cables can also misbehave over time.

 

[Quantum J Audio]

I’d go back to basics. Strip the system back to the most basic level. Remove the digi sync. Try one 192 see if it works, you just have to do good old fashioned fault finding and see what introduces the fault.

If the one unit still errors, replace the interface lead still errors, try the second interface.

Could also be an HD card issue, you a lot of variables to rule out.

 

[DaveDC]

The DigiLink interface consists of a number of LVDS serial connections transmitted over twisted cable pairs. The connections are organized into two 16-channels transmit lanes and two 16-channel receive lanes (X_IN, Y_IN, X_OUT, and Y_OUT.) Each lane consists of two 8-channel LVDS pairs plus an associated LVDS clock pair. All lanes are presented to an FPGA in the 192 where “patchbay” logic routes data to and from the various I/O cards. The FPGA also monitors the lanes for audio levels and drives the meter LEDs accordingly. The X lanes are used for the internal I/O cards only. The Y lanes are also sent to the Expansion connector. In the case of data coming from the DSP (Pro Tools outputs), the Y lane connection goes straight to the Expansion connector. In the case of data going to the DSP (Pro Tools inputs), the FPGA first selects the data source between internal and expansion.

Your unit seems to be (mostly) having problems with the Y data lane coming from the DSP. Because all sixteen channels are affected, I suspect the Y clock signal. A few directed tests could help narrow down the problem:

1. Is the input (ADC) path from the Expansion 192 okay? This will tell us if the FPGA in the Expansion 192 is okay. You could swap roles between the two 192s and see if the problem follows the box.

2. Is the audio messed up along with the meters? If the audio is okay then maybe just the meter drivers are bad. This would be uncommon but not impossible.

3. On both interfaces, check the internal 50-pin ribbon cables that connect the rear panel with the front panel. These carry the raw LVDS signals. It is possible that a lane clock has a dodgy connection. BTW, there is nothing special about this ribbon cable, so you can replace it with any standard 50-pin cable of the same length.

Let us know what you find.

 

[DaveDC]

And Quantum J Audio is correct that the DSP card (specifically its Y data lane clock) could also be going bad. If you have multiple DSP cards, you could swap them around. If you do find that one has a flaky DigiLink connection, I don't see any problem with just running it as a DSP card without I/O. I doubt it will blow up the rest of the system.