What caused the failure of the Christie ACT. My Theory.

[Devin McNamara]

While these units are probably not very common anymore. If you’ve got a few sitting around, a gas stove, a heat gun, some aluminum foil, and you are bored keep on reading. I acquired a couple of these from a closing theatre. Neither of the units booted. They powered up but went no farther. The green buttons lighting up suggests the power supply is functioning. A simple meter test confirmed this. Upon looking at the motherboard inside the ACT one will see the relay section at the left, the main processor section in the middle, the NIC section at the right, and the power supply section at the bottom left. The problem, at least with both these units, is occurring in the CPU section. The CPU itself in fact is the problem. Both the CPU and NIC chips are mounted to the board using a form of electronic connection known as Ball Grid Array, a format where the connections of the chips are underneath the IC in the form of small beads. The chip is coated in flux then heated. A machine puts the beads on the chip and it is heat smelted onto the board. These birds are RoHS compliant which means that they use lead free solder which in the electronics world is an absolute nightmare. These units were manufactured in the early to mid 2000s around the same time the XBOX 360 suffered the RRoD. Essentially what is happening is heat cycling is causing these crap BGA solder beads to crack. This causes the CPU chip to lose connection to the motherboard and stall the boot loader when powered up. Looking at the board, the chip in question has the word “coldfire” on it. Place a piece of aluminum foil on a stove burner, remove the battery from the board and place it on top of this and light the burner and keep it as low as possible. You are only get the board decently warm. Now you want to cut a hole in some
more foil just slightly bigger than the size of the CPU. Place that over and around the CPU and proceed to heat it with a heat gun. I had luck after about 20 straight minutes of heat with my little 300 watt RadioShack heat gun. You don’t want to over cook it, the goal is to get the solder under the chip just molten enough to reflow. Usually surface tension will keep the chip in place as long as it’s level and you don’t bump it. Let the board cool before attempting to power it up. Once I re-powered it I was surprised to see the ACT spring to life. So far, after a couple of weeks of tinkering with it, it has continued to function. Your mileage may vary. It may take a few tries to make it permanent. It took me about four attempts each time I heated it longer each time.

[Harold Hallikainen]

Congratulations! That's pretty amazing. BGAs are really something. We had boards go through x-ray inspection to ensure the BGA soldering was good. Again, congrats on the success!

[Marco Giustini]

it's recently turned out that the issue with XBOX and PS3 at the time was not the BGA under the chip but the connections between the actual die and the interposer which is then soldered to the board. What usually happens is that heat will temporarily fix those connections but thermal cycles will break them again - it's apparently a design flaw with the material used to keep the die solid with the interposer.

Make sure you power on and off the ACT daily for a while to confirm that's working fine.

Obviously, I wish you that's not the case with your unit!

If you allow me, gas stove and heat gun are usually fatal for a PCB! Definitely you want to warm them for a few minutes, not 20. The lead-free reflow profile allows 30 seconds over the melting temperature (220C for Lead free solder), no more. The whole reflow process should last a few minutes. That said, if that worked for you, then happy days

[Bruce Cloutier]

I to this day avoid BGA mountings, more or less for fear of issues in trying to rework or even develop confidence in the connections beneath. Bad enough that I have had to resort to QFN packages (no pins, just pads under edges) to actually get parts given the component supply mess.

The ACT that I have seen have had their switches completely worn. But, yes, early lead-free solder formulations were too brittle and would fatigue easily with repeated thermal stress. This failure mode is very common. And stories of products brought back to life in toaster ovens are common as well. I had a quality large rear projection TV that succumbed to this problem. In this case luckily it was the solder connections for large components such as filter caps and transformers. Repair required re-soldering those connections with the iron. Under magnification you could clearly see the cracks. Those had to join up to create a complete separation and probably then corrode a bit before causing failure.

I don't think that this issue is the [only/key] reason that Christie no longer makes the ACT.

It is one thing to bring something back to life and quite another to put that back in service to where you are relying on it. But I bet it is gratifying to figure it out and get it going again.

[Steve Guttag]

If you are interested in rework, in general, I highly recommend Louis Rossmann's channel on YouTube. His company does Macbook repairs but his YouTube channel has hundreds of videos on troublehooting and rework, including BGAs. And, to echo Marco's point, he has shown that while "reballing" which reheats a component may cause it to work again, most of the time, that is short lived as the failure point was likely within the chip, not the balls themselves and it was inconclusive if the "fix" was heating of the component or the reball, for which time will tell.

As for the Christie ACT. It really was only good for a particular type of customer. One really had to invest into its scripting to make it worthwhile. If all one wanted was a relay to respond to an ASCII command...there were other, cheaper, options out there (JNIOR and eCNA come to mind). If you did do the substantial time investment, then you could create your scripts such that each ACT was identical through out one's chain and only the IPs of the particular site need be changed. The LEDs on its face...while cut that it could animate them...just who is looking at them?

There definitely was power in the ACT but I found it overly complicated for what it was and would charge a premium to work with them due to the excessive time it took just to make customer requested changes.

[Leo Enticknap]

One really had to invest into its scripting to make it worthwhile.

Agreed completely. The one and only occasion I had to f*** with one was last summer, when I was part of a team that installed an LG direct view LED videowall to replace a conventional projector and screen in an 8-plex in Orange County. That project also included replacing the audio processor with Q-Sys. There was a pre-existing ACT automation, which I had to reprogram to interact with the Q-Sys core (both sending commands to it and receiving them from it) and the LG screen's control board. The time investment into its scripting language just to make those tweaks was basically a day. The impression I got is that it is an extremely versatile unit, but a lot of what it does duplicates the macro and cue functions that a DCP server has anyways. I haven't had to deal with one since than, and enough time has now passed for me to have forgotten most of what I had to self-teach myself that day. Grrrr....

[Steve Guttag]

Hence...anyone that wants me to mess with one pays extra. I just don't have the time to invest in such an oddball. And that was before it was discontinued. Now, I'd probably charge double just for the frustration. With an eCNA, I'd be in and out in moments! I'm sure others would say the same with the JNIOR (and that IS one that I'll invest the time in someday, when the job calls for it...it is so pervasive).

[Bruce Cloutier]

We had insight into the production status of the ACT before it became unavailable. These are being replaced with a JNIOR and one of our control panels by some (including Christie). We had reviewed the ACT I/O count and user interface with recommendation that we offer a panel more synergistic with the ACT but with a JNIOR core. I was recommending a new product version as opposed to an assembly with a current JNIOR model inside. The former being less expensive to manufacture, likely more reliable and having more functional value. This was some time ago and the concept did not get traction. I believe it would have had we had a discussion about it with Christie and then surveyed some ACT users. Steve points out the I/O count advantage of the eCNA. This would have offered an alternative.

Anyway, I am hoping that things loosen up in 2023 allowing us to bring new things to the table. We are still sweating the supply chain crap.