"New" Alchemy shipped with almost flat certificate battery!

[Leo Enticknap]

I have just installed a supposedly "new" (antistatic baggie still sealed with Barco tape) Alchemy. When I opened it up, I found that the certificate battery was only five months short of the official replacement interval of five years!

image.png

After replacing it, I discovered that it was not long for this world:

image.png

AFAIK, Barco don't have an official recommendation of how low is a safe voltage, but I've measured batteries in Alchemies that have actually failed into FIPS lock in the 2.9s. I had a very lucky escape with this one (it was OK when I booted it after replacing the battery).

I find it disappointing that Barco is shipping supposedly new equipment that is this close to failure. The moral of the story is to keep several batteries in our field spares kits, and to check the batteries in all these things we install, including pulling and checking those that ship pre-installed in a projector.

[Gordon McLeod]

We got several GDC SC3000 that had dead batteries on delivery

[Steve Guttag]

In the current ICMP/ICMP-X service manual, they do discuss the minimum voltage value for units in storage and that is 2.8V. It is strange at how they word it too. They're almost cautioning you from prematurely changing a marginal battery. As in, if you measure the battery and it is above 2.8V...then it's best to let it go another year.

This step is recommended for ICMP units that have been in stock for more than 5 years: these units do not work but they are also not powered by the projector. You need replace the battery if the voltage is less than 2.8V. Otherwise, it is better to wait another year if the unit will not immediately used.

The other "helpful" tip they have is if you have stock batteries, check them before installing and ensure that they are at 3.2V or higher to ensure a 5-year life.

With the penalties so high (dead server) for letting that battery drain and the cost of the battery so low (relatively speaking), I've been on a more of a 4-year life-cycle for IMBs that use batteries to ensure I catch them before they drain because, again, the cost ratio of a dead battery to the cost of the battery itself. And, if you are pulling the ICMP to clean its edge connectors, as per the Info-T...you are already there.

My strategy also came in handy during the C19 shutdowns. i had a site were I just changed batteries out in 11/2019 (GDC SX2000AR)...the site fell victim to going out of business as not being able to sustain itself during the shutdown. After a year of no-power to the building before it was acquired by the next entity, we were brought back in...no dead IMBs...and, as part of my wake up procedure...all fresh batteries again...they were about 3-years old, at that point but also running without any projector power either.

Don't mess around...change the batteries.

And those of you with traditional "box" servers, if you haven't already, start thinking about changing those BIOS/CMOS batteries (you know, the CR-2032 or Leo's BR-2032). The hotter the booth the faster they drain. I've seen servers do very strange things on flat or mostly flat BIOS batteries.

The real big scare is going to be all of the ICPs out there when their certificate batteries die (soldered in BR-2032).

[Agnius Acus]

Which ones have soldered batteries?

[Steve Guttag]

All TI ICP boards (not the ICMP) as used in everyone's Series 2 projectors. It is right next to the socketed one. The socketed one is just for the RTC clock.

[Philip Jones]

When are they likely to start failing?

[Leo Enticknap]

If you mean the TI ICP soldered batteries, TI's official line is that they have a 10-year design life. However, I suspect that they are only discharging when the projector's card cage is unpowered. I service some very early (as in, manufactured in 2010 or 11) Series 2 projectors, still with their original ICPs, and those ICPs are still OK. However, they have been powered up 24/7 (apart from during the pandemic). Almost all the ICP battery failures that I've encountered have been in projectors that at the very least are powered down overnight, and in some cases in mixed use venues and powered down for days or even weeks at a time.

Barco ICMP batteries are officially supposed to be good for five years. However, I agree wholeheartedly with Steve that they should be replaced every four to be on the safe side; especially as the risk of losing an ICMP during the swapout process is extremely low. They're not like SX-3000s and cat745s, for which the risk of losing the cert during the battery replacement process is nontrivial.

Anyways, it seemed that I jumped the gun in ranting. Apparently that Alchemy was actually delivered to our shop in 2020 as part of an order for a sale that wasn't completed for some reason (I'm guessing to do with covid, given that it was 2020), and was only recently rediscovered during a clear out / stock take. So my apologies to Barco on that one. I'm told that there are some more, too, so I'll have to get back to do those batteries and test them sooner rather than later!

[Agnius Acus]

Here‘s what it says in the NEC NC900 manual:

To backup the authentication data needed to receive cinema video signals, a secondary battery is used inside the projector. If you have not used the projector at all for 6 months or more, the battery will lose power and the authentication data will not be able to be backed up. Always put the projector into standby mode for at least 48 hours once every 6 months to recharge the battery.

I believe we never changed the battery during almost a decade of service, although ICP was checked multiple times, so maybe it was. It seems to be a trivial thing to change compared to other things, like fans, but:

Although data is maintained while the battery is being replaced for approximately 3 hours by the sub-battery built into the ICP board (when the sub-battery is fully charged), please replace the battery quickly. To ensure that the sub-battery is fully charged, turn the projector power supply on for 30 minutes or more before replacing the battery.

[Marco Giustini]

This step is recommended for ICMP units that have been in stock for more than 5 years: these units do not work but they are also not powered by the projector. You need replace the battery if the voltage is less than 2.8V. Otherwise, it is better to wait another year if the unit will not immediately used.

Sounds more like an NEC bulletin

I think they suggest that an unpowered board is using a bit more power from the battery than one which is normally used - is there a supercapacitor in those cards? Maybe what they want to say is "power on the board before swapping a battery will give less chances of failure"? I am guessing here.

[Steve Guttag]

Since NEC uses a stock TI board, there is no special aspect to their batteries (speaking to series 2 projectors or, as NEC would consider them, early series 2 as they have the ICP...NEC considers even late model projectors series 2 even though they are not using a TI ICP). The 6-month issue is normally for the Enigma (aka Link Decryptor) board that uses a super-cap that will drain within 6-months of no charge. Just had one go in an NEC recently (NC2000C). One day it was fine...the next, it was toast. It had more than just a lost certificate error too.

As to the ICMP battery swap. It still gives me the hebegebees. The battery clip they use is not all that forceful so I'm amazed it makes good enough contact. On the up side, you can verify that you have the new one in properly before removing the old one.

I have not had the issue Leo has had on the SX3000 or SX2000AR (same IMB...one just has the OS part as a daughter board). They too use two batteries but in tandem. I think the biggest fear is if the rubber feet that push down on the batteries might stick to the batteries and "pluck" the batteries out of their sockets. I always hold down the batteries when removing the board stiffener. Now, I've had other SX3000 issues...in fact, I've had 100% failure out of them as none of the ones I've services are on the original. They tend to drink their batteries so I'm more touchy about battery changes. The SX2000AR does not seem quite as thirsty. The SR-1000, if you are on semi-recent software, will show you the voltage of each battery too...so the mystery is removed on when to change.

[Leo Enticknap]

The risk I've found with the SX-3000 is that if the batteries are left too long, one by itself is not enough to sustain the memory while you swap out the other. Their instructions tell you to measure both before starting, and then swap out the weaker one (the one reading the lower voltage) first, leaving the stronger one to hold the memory solo. If the stronger one is above around 2.9, it's usually fine. Below that, however, the risk of losing it becomes significant. I've lost a few that way, so much so that if I measure both batteries at 2.9 or below before starting, I advise the customer of the risk and ask if they want me to proceed, or come back when they've reactivated the warranty and/or having the screen down for a couple of days is less of a problem.

The ICMP (and ICP-D) is the model I'm least worried about when swapping batteries. I've likely done the best part of 100 battery swaps, and have only lost one. The biggest risk factor is that when you lever the old battery out, the insulating straw over the positive spring terminal goes with it, thereby risking a short, especially if the battery is a CR2477N, on which the negative terminal extends up to the plastic o-ring under the rim. I think that must be what happened with the one I lost.

[Steve Guttag]

Of course, now, they recommend using a plastic tongue on the bottom in addition to the straw on the clip so you are insulating on both terminals when moving the battery in or out. Knock on something, I haven't had a mishap on any server's battery swap.

[Marco Giustini]

Since NEC uses a stock TI board, there is no special aspect to their batteries

Mine was a joke about Barco's English on that paragraph

[Jason Raftery]

The risk I've found with the SX-3000 is that if the batteries are left too long, one by itself is not enough to sustain the memory while you swap out the other. Their instructions tell you to measure both before starting, and then swap out the weaker one (the one reading the lower voltage) first, leaving the stronger one to hold the memory solo. If the stronger one is above around 2.9, it's usually fine. Below that, however, the risk of losing it becomes significant. I've lost a few that way, so much so that if I measure both batteries at 2.9 or below before starting, I advise the customer of the risk and ask if they want me to proceed, or come back when they've reactivated the warranty and/or having the screen down for a couple of days is less of a problem.

The Tadiran TL-2450 batteries in the SX-3000 should be considered dead when they drop below 3.50V. Some early SX-2000AR IMBs shipped with Panasonic CR2354 batteries--those can be considered shot at 2.90V. < 3.40V on the TL-2450s will put the cert at risk. Once the voltage starts to drop below ~3.55V on the TL-2450, look out--they will crater rapidly (I've seen them go from ~3.5V down to < 3.0V in ~2-4 weeks if the IMB is not installed). See the discharge characteristics in the attached specs page.

If changed every 12-18 months (for IMBs in storage) and 3-4 years (for IMBs in daily use), I haven't had big issues with the SX-2000AR or SX-3000 certs being lost due to dead batteries. Like Steve, I've noticed the SX-3000s (and SX-2000ARs w/ SOMs) seem to drain batteries faster than the regular SX-2000AR--there may be a slight increase in load on the batteries from the daughterboard's presence on the IMB.