CP750 - level drop disease presented shortly after power supply swapout

[Leo Enticknap]

I look after an 11-plex in which the CP750s (all but one manufactured between April 2011 and January 2012) are dying. They are gradually upgrading to 950s. As of now, four screens have been upgraded, and the remaining seven still use 750s.

The power supplies died in two of these remaining CP750s within the last month, and when we got new ones back in stock, the manager decided to replace the ones in the four remaining 750s preemptively (to avoid down time when they eventually crap out). This I did last Monday. On Saturday, they reported a fault with one of them, symptomatic of level drop disease.

I checked it out on site just now, and that's what it is, alright - only 66-69dB playing pink at reference. Cranking the channel gains all the way up gets it into the low 70s. They're having to play movies at 8-9 on the fader to get an acceptable level: previously, they were playing features on 5.5 to 6.5, depending on the movie.

My question / request for opinions is this: could replacing the power supply board have tipped the motherboard over the edge? Slightly increased DC voltage output from the new power supply, maybe? If there is a risk of this, then I will advise other 750 operators that replacing the power supply preemptively is not a good idea, and that they should wait until the existing one dies (at which point they'll have nothing, except the cost of a relatively cheap part, to lose).

[Steve Guttag]

While that is possible...I wonder if the connector for the power supply had issues (I've seen some that are cooked) and maybe you're not getting the full voltage now with a poor connection. Normally, the level drop is a gradual one, not a sudden one. If it is affecting multiple channels and all at once...again, I'd check the power supply connections and verify that all of the rails are at the right levels.

[Bruce Cloutier]

Maybe swap one of the not yet failed power supplies back in that 750 to see if the audio levels return?

[Bruce Cloutier]

When were the 750s manufactured? Maybe there's a bad solder joint on the board. Check if the audio level is sensitive to stress on the motherboard? For a bit at the beginning of the lead-free push the solders weren't quite yet tested. Solder on the PCBs would crack after thermally being stressed. I have repaired a number of things by simply reheating suspect solder joints. Maybe it just got disturbed during the power supply replacement?

[Carsten Kurz]

Dolby has commented on these typical failures - for the power supply, it is an inrush current limiting device that fails, while the level drop is caused by overheating capacitors. At least that's what I remember from the tech notes.

[Leo Enticknap]

I wonder if the connector for the power supply had issues (I've seen some that are cooked) and maybe you're not getting the full voltage now with a poor connection.

Some of the Molex-type connectors had gone brown and brittle at the power supply end: I can't remember whether this one had or not.

Normally, the level drop is a gradual one, not a sudden one.

I don't know for sure that it wasn't gradual. All I know for sure is that there was a customer complaint on Saturday, in response to which they had to crank the fader up to way above reference. It's entirely possible that it had been deteriorating for a few days prior.

Maybe swap one of the not yet failed power supplies back in that 750 to see if the audio levels return?

This is not the typical power supply failure mode. All the ones I've dealt with have either presented with the unit dying completely and utterly, or the front panel LCD and buttons going out (or dim and flickery) and sometimes a 60Hz hum/buzz being sent to one or more power amp channels, but LAN communication being OK and the meters displaying the input signal OK.

They've ordered a 950 for this screen. If I get time to experiment further when the install day arrives, I will (though reinstalling the previous power supply is not an option, because they all went to e-waste).

Many thanks for everyone's suggestions.

[Marco Giustini]

How much do you know this customer? Could it be that they never realised of the level issue until they swapped the PSU and decided to actually take a look at the auditorium? How many times an existing fault is highlighted only when someone touches the system for maintenance?

[Steve Guttag]

FWIW...while recapping CP750 power supplies is not enjoyable...if one runs into a shortfall of getting them, it is a viable means to keep most CP750s going. If the motherboard fails...it is an obscene amount of capacitors to recap though there are two diodes that have been identified as potential problems (particularly the surface mount versions) near the power supply feed to the motherboard, which was designed to take either SMT or Through-hole diodes.

[Marco Giustini]

and those diodes can cause the same level issue?

[Steve Guttag]

No, they normally will kill a rail.

[Leo Enticknap]

How much do you know this customer? Could it be that they never realised of the level issue until they swapped the PSU and decided to actually take a look at the auditorium?

I can believe that this CP750 had been gradually deteriorating for days before the fault was noticed, but not for weeks. This is why I suspect that the power supply board replacement may have triggered the start of the process. I do whatever projector planned maintenance is due (B, C, etc.) when I'm there to do lamp swapouts in a given screen (site staff do air filter cleaning; I do everything else), and also blow out the amplifiers with a Datavac and play pink and a sweep test on all the channels and check the levels at reference with a SPL meter. I last worked on the screen in question on August 13, and the CP750's output levels were OK then. We have a concurrent problem whereby the XC-1 crossover cards in the old MX1500a amps are now failing in significant numbers: usually the low goes, and the mid/hi survives (though in one case it was the other way round). If this happens on a center channel amp, they notice immediately (obviously, because of complaints about hard to hear dialogue) and call us, but if it happens on left or right, it's usually noticed when I'm there for maintenance.

Due to time constraints and the desire for guaranteed reliability going forward, attempting board level repairs on these 750s isn't an option: we're now upgrading to 950s as soon as a fault presents. Dolby does now have CP750 power supply boards back in stock, and I'm told that for the moment at any rate, there is no problem getting hold of them. I will experiment with this unit further (specifically, looking at the cable and measuring resistance, and if I manage to acquire a "used but OK" power supply between now and then, trying that) if I get a chance. This could produce useful information with which to help other customers, for one thing.

[Steve Guttag]

Are you sure it is the XC-1? I've normally found that the input attenuators are the evil ones.

[Bruce Cloutier]

I don't have experience with audio design but have had interest in getting into it. The sound quality must be impacted as well and not just the audio levels, right? You have to be losing dynamic range or something.

[Marcel Birgelen]

Analog audio electronics is fun... you need to actually know a bit about electronics, that's probably why it's often done so shitty those days. Most modern designs try to get the analog signal into the digital domain as fast as possible and back into the analog domain as late as possible...

Whenever you have to crank the volume to compensate for a loss of signal strength, you'll decrease the dynamic range and increase the noise level.

I've sucesfully recapped a CP750 mainboard once, but I used an ESR meter and replaced any capacitor that was borderline between bad and ok. The operation was sucessful. I didn't find one cap that was entirely out-of-spec, but a bunch that were just so-so. I guess that's also where the trouble lies with the motherboard. Overheating causes a bunch of caps to go borderline bad and the design simply has been too tight and doesn't allow for any degradation over time.

[Steve Guttag]

I'm impressed. There are a LOT of capacitors on there and I'm sure many are in parallel so did you remove them (say hot tweezers) measure and then put back? And that is just it...one would almost need to slather on the flux...tweezer off, tweezer on ...repeat, repeat...about a 100 times...then wash the board/ultrasonic clean that mess off. It is a daunting task. Something I could see as a 1-off if one owned it and had more time than cash to replace.

Even the power supply, it is cheaper to buy new than to pay the real costs of recapping (labor...the caps aren't bad and there are not that many of them). However, when you can't get the part, that is better than nothing.