Troubleshooting Lost Surrounds

[Martin McCaffery]

Trying to narrow things down before we can get a tech here.

Monday we had some loud popping and crackle in the right surrounds. I wasn't here for it, but it was enough to get the audience to complain. We've had random pop and crackle in the past, but so fleeting we couldn't hunt it down.
Today I tested (no shows since Monday) and discovered we have no surrounds at all now. When I run a digital sound loop, you can see on the CP650 there are no bars. When I switch the monitor to read off the amps or read off the processor, no sound at all in the surrounds.
When I use the pink noise generator on the 650 the surrounds are working loud and clear off both the amps and processor.
Therefore amps work and sound processor works.

So it appears the surrounds are not getting from the server to the 650, or the 650 is not processing the surrounds.

Any different conclusion and where do I go from here?

[Frank Cox]

What server? I wonder if there's a setting in that somewhere that told it to turn the surrounds off. "2.0 sound output" or something, maybe?

[Mike Blakesley]

Assuming you have rebooted everything....right? (That's always the first question)

[Martin McCaffery]

GDC 1000 and yes, everything rebooted/

[Leo Enticknap]

If internally generated pink within the 650 is OK, but AES coming in from the SR-1000 is not, my instinct would be to check the following.

• The two Ethernet cables taking AES from the SR-1000 to the CP650. Are they properly strain relieved at both ends? Try testing and/or replacing them (with shielded).
• You must have Odyssey boards and/or RJ45 to DB25 adapters on the back of the CP650. Disconnect, contact clean, and reseat.
• Is it just the rear surrounds that are giving trouble, or the side surrounds as well? If it's just the rear surrounds, the input to those goes through the option card I/O DB25 connector on the back of the 650. The Odyssey board for 7.1 is actually two boards, with a ribbon cable between them, the daughterboard being purely to get the AES for the rear surround channels into that I/O. If it's just the rear surrounds that are out, this suggests to me a failure in the Odyssey board.
• Other possibility is a bad cat790 card in the CP650. I did have one in which there was a random crackle on all channels. Opening up the CP650, pulling all the cards out, spraying DeOxit onto all the multi-pin connectors (plus a thorough internal clean in the process), and then reassembling it cured it. Worth a try if all else fails.

I'd be surprised if the fault is in the SR-1000, given the relative age of the two units at either end of the AES cables, but you never know.

[Marcel Birgelen]

It it started with a pop and crackle and pink noise from the CP650 still works, my suspicion is the Cat790 in the CP650. The pop and crackle indicate something in the analog part of the chain is failing. Leo's train of thought seems legit though.

If it turns out to be the CAT790, my advise is to replace the entire CP650 instead. Seemingly, there still are sources on this forum for them for reasonable prices.

[Jim Cassedy]

I'm with Leo on checking AES cables & Odyssey board connections. I experienced almost
exactly the same issue once and it simply turned out to be the Odyssey adapterthingy had
gotten loose. As long as i had the 650 out, I also pulled out ALL the cards, cleaned all the
contacts and took all the circuit board to the alley in back of the theater and cleaned several
years of accumulated booth-dust offa all the boards using a soft paintbrush & air compressor.

[Leo Enticknap]

The Datavac is my weapon of choice for creating an impressive dust cloud from a machine that hasn't been opened up in a decade. But yikes! They've gotten more expensive. When I bought mine in 2018, they were only $60.

If it turns out to be the CAT790, my advise is to replace the entire CP650 instead.

If Martin decides to go that route, the gotcha is that not all CP650s have that card in them, and so this is something to check. That having been said, if the nuclear option of replacing the entire audio processor is to be seriously contemplated, then unless it is desired to retain 35mm audio capability, I would opt for a currently supported model, e.g. the CP950 or Trinnov OV2, unless the budget for doing that simply can't be raised. A used but checked out and confirmed OK CP650 can be had for significantly less, and installing it would be a lot less disruptive; but you'd just be kicking the can down the road.

[Steve Guttag]

If there is no film capability left in the theatre, I wouldn't put another CP650 in. You are definitely kicking the can down the road and that road may be very short. I use CP650s for film-only now...sources to other systems. If a processor like the OV2 was installed, and you do have film capability, the OV2 does have a multichannel analog input to accept the CP650's feed. If no film, then I agree, the CP950, OV2 or Q-SYS, if you think the flexibility it brings is of value but that will be the most expensive option, most likely (but the most flexible, by far).

As to the problem, depending on how the AES gets to the CP650, pair 3 could have a break in it...as it was breaking, it would get you wild digital audio that could be interpreted as crackling (or even just outright noise)...and if broken, it would be dead.

Since it is an SR1000...with RJ45s. String a temporary CAT cable from the SR1000 over to the CP650 (presuming an Odyssey RJ board) and try that with your test content to see if it works...if it does, replace the cable, if not, check the Odyssey board...could have a cold solder joint or a problem with the transformer (the CP650 uses 75Ω while the SR-1000 uses 110Ω so there is some matching that should be done.

[Stefan Scholz]

Yes, SR 1000 is 110 ohms balanced line, CP 650 has sneaked the inputs at a later stage, and reused existing unbalanced 75 ohm connection. It is not a problem with the short route from a server to the processor to directly connect them. Even Dolby does state that in their DSS 200 manual.
Check odyssee board. You could also solder a direct cable from RJ 45 to DB, which works well. I did this more than once. Just a few feet, there are no problems with reflections due to mismkatchm, and we're not talking microwave transmission here.
Checking input connection, clean internal connectors and reseat boards would be my work of choice.
If still unreliable, I follow everyone elase here, and order a new (in box) processor, CP 950, DA 25 or Ovation 2, if you're not going into the networked approach, which I favor today.

[Marcel Birgelen]

If Martin decides to go that route, the gotcha is that not all CP650s have that card in them, and so this is something to check. That having been said, if the nuclear option of replacing the entire audio processor is to be seriously contemplated, then unless it is desired to retain 35mm audio capability, I would opt for a currently supported model, e.g. the CP950 or Trinnov OV2, unless the budget for doing that simply can't be raised. A used but checked out and confirmed OK CP650 can be had for significantly less, and installing it would be a lot less disruptive; but you'd just be kicking the can down the road.

In the end, you're always kicking the can down the road, the question is how far you want and are able to kick it. Martin knows best what his budget is, but the fact that he's still operating on a CP650 could give a hint regarding the budget.

There is a hefty price difference between a new Ov2 or CP950 and a refurb CP650, especially since the CP650 would be a drop-in replacement and anything else would be a new install. I'd always try to get the CP650 from a reputable source, one that would guarantee the presence of the Cat790 and not some Ebay auction, that's why I hinted at this very forum. If you established the defect is in the CP650 and unless you're doing this frequently, I think it's better to replace the CP650 in one go, instead of messing around with the components in it. Especially looking at what a refurb replacement CP650 would cost v.s. the potential labor involved in debugging individual components.

[Steve Guttag]

Even Dolby does state that in their DSS 200 manual.

Where? Where does it say that? Here is what it says below figure 3-15:

The DSS200 8 x AES OUT connector is balanced 110Ω, and the CP650 OPTION CARD I/O connector (with Cat. No. 790) is unbalanced 75Ω. If you are not using a cable manufactured by Dolby, be sure to use a similar cable that meets the AES audio specification.

That is a far cry from: eh...do what you want and just hook it up any old way. Note, even CAT cable isn't really 110Ω cable, it is (mostly) 100Ω. And yes, you'll get away with a LOT on shorter distances with impedance mismatch. CAT cable can go about as long as any cinema could want without issue with the slight mismatch (note...it isn't a perfect 100Ω either...depending on how it is routed, what bends it has...etc. One is also quite dependent on the tolerance at both ends. Note too, AES3 the 110Ω balanced runs at a substantially higher voltage (2-7Vp-p) than typical AES3-id (the 75Ω coax at 1Vp-p).

I would not advise someone that has a problem with a particular signal to start out by throwing caution to the wind and hook it up wrong because [to paraphrase]: "Its always worked for me."

My reasons for abandoning the CP650 for Digital Cinema use is that EVEN if you get a brand new one (and I think Sam Chavez still has some NOS CP650s)...Robosound and the like are just waiting for you...it isn't an if...it's a when. If the cost of a replacement sound processor is what would put a theatre into the red (and we are talking about an investment that should last over a decade and quite possibly 2-decades, I'd question the theatre's viability.

Again, before one starts replacing stuff, I'd check to see if the signal on pair 3 is getting to the processor. I believe that the CP650 just clocks off of pair 1 (so you can't just hook up pair 3, you have to have pair 1 as well).

[Leo Enticknap]

From Martin's description, it's looking like this system worked OK until one day it didn't, without anything having been done to it. So if there was an impedance mismatch on the day the SR-1000 was installed and the CP650 was able to cope with it then, then either this isn't the problem we're looking at now, or something has happened to make that mismatch worse (e.g. resistor failure on the Odyssey board).

If this is a hardware fault with the CP650, I think everyone has outlined the advantages and drawbacks of the two responses to that. If you decide to go the "the only replacement for a DC-3 is another DC-3" route and replace the 650, then you will pay less up front, and, assuming that you are able to save a settings backup out of the old one and upload it into the new one (a process that will require an RS232 to USB adapter, unless you have an ancient laptop that actually has an RS232 DB9 port on it), you won't have to hire a tech to install it, tune it, re-do automation cues, etc. etc. But even the final CP650s to be built are at least decade old, with all the reliability implications that involves. Replace the processor with a currently supported model and you likely have no risk of any trouble with it for 5-10 years, possibly subjectively "better" audio, and the full range of spares available from the OEM if something does go wrong (not to mention a time limited parts warranty that comes with it). But it'll be more expensive to buy, and will need to be installed, integrated, and tuned from scratch.

[Steve Guttag]

Leo, my discussion on the impedance mismatch was in response to Stefan comments (and what I think is bad advice). Whereas I've had bad RJ45s that have failed over time (I'm guessing that the crimp lossed contact or there was a break in the solid conductor) that is the fastest/cheapest troubleshooting...presuming he had an Odyssey board installed as the test, if the cable is in hand, takes a minute to plug it in.

Digital is deceptive because it works until it doesn't. That pair may have always been on the weak side for any one of a number of reasons (at either end, and including the cable) and over time it has fallen from good enough to not good enough. Digital works or it doesn't. To use a film analogy. If a Dolby digital reader can run a solid 7 and another runs in the 0-1 range...which sounds better? However, if something starts to fade or a connection gets a little oxidized...which one slips into failure faster? With AES3, there is no meter...there is, at best, a pass/not passed indication on some systems. Naturally, there are testers for the CAT cable itself in varying degrees from continuity, to qualification to certification (with geometrically increasing costs associated with each). If it were me, I'd plug around the existing cable and see if that "fixes" the problem first...it is the cheapest/fastest test he can do.

[Leo Enticknap]

I think we are broadly agreeing with one another: my point is that if increased resistance is the cause of this problem, then something must have happened to cause the "fallen from good enough to not good enough." It could be as simple as someone putting a heavy box down on the AES cable running along the floor next to the baseboard, crimping the cores inside, and then when the box is lifted...

IMHO, troubleshooting something like this is analogous to the classic risk assessment method that one is taught in the health and safety classes that megacorps subject their new hires to. You rank the likelihood that a risk materializes from one to five, and the impact/consequences if it does happen from one to five, multiply the two numbers, and that is your total risk factor. So a common cold would be a 5 x 1 (very likely to happen, but will not have serious consequences if it does), and a 737 spinning out of control and crashing onto your house is a 1 x 5 (very unlikely to happen, but if it does, you're buggered). The total number helps you to decide how much it's worth investing to mitigate the risk.

My adaptation of this is to change those two factors to how likely is a suspected reason to be the cause of the fault, and how easy is it to do the checks needed to confirm or eliminate that suspicion? IMHO, I would run this as follows in this case:

Loose or dirty contacts in external connection = 5/1
Bad AES cable = 4/1 if you have the means to test it to hand (e.g. a Fluke meter or another cable to swap in); 4/2 if you don't
Loose or dirty contacts inside the CP650 chassis: 3/3
Odyssey board or connector hardware failure = 2/4
Cat790 card failure = 3/4
Other CP650 hardware failure necessitating replacement of the unit or obtaining a discontinued part = 3/5