ICP cert batteries depleting

[Carsten Kurz]

Taken from here: http://www.film-tech.com/ubb/f16/t003789.html

I am in discussion recently with a colleague about his series 2 Kinoton. It was installed in 2012. Last year I mentioned the upcoming mass-dying of ICP boards on the german cinema forum, and he contacted me recently about his options. For many smaller operations, loosing a mid four-figure board in these times because of a 2$ battery is not a pleasant experience. Especially if you can see it coming, but can't do anything about it. So, I made him take some pictures around the batteries and the top and bottom side of the PCB. As it turns out, the clock battery socket is smd mounted on his ICP, and the cert battery is soldered in, but through hole. So, first of all, it is very easy and safe to measure the remaining cert battery voltage on the solder contacts of the bottom side PCM. They are far from each other, and there is very little risk of a short.
This ICP had some isolation glue on these contacts, but that was easy to remove. So, he measured his cell voltage to be 2.98 volts. Leo mentioned 2.48 volts with that board that lost it's cert. After 9-10 years since assumed production date, there is more life left on that 2.98 volt cell than I thought.

Now, from what I see on these pictures, it should be easy to solder a wired bridge battery onto these bottom PCB contacts. I don't know how safe it is to leave the empty original battery in place - but as long as the bridge battery is connected on the bottom, it should be no problem to clip the top/+ contact off with a wire cutter. Then, the original battery is isolated, and you are free to leave it in place or desolder it completely.

I am wondering what that header next to the cert battery is. I can't see any traces going to the battery from there (also not on the bottom). So, it may be that this header is used for serializing the ICP at the factory, but not to attach a bridge battery.

Any thoughts on this?
From what I see, there should be a reasonably safe way to save ICPs from dying the cert battery death. More important, there is a safe and easy way to check wether such measure is already necessary. Even if we don't know the exact voltage level where it becomes critical. Until we learn about more dead ICPs, I would think it becomes critical when the cell voltage gets close to 2.6 volts.


ICP_bats.jpg

ICP_bats_bot.jpg

[Mark Gulbrandsen]

I have also contemplated soldering on another temporary battery, but I think it would be wise to isolate it with a diode. That diode is gonna give you about a .7 volt drop too. So you may want to use a variable bench supply or a regulated source set to about 3.68 +/- a few tenths of a volt. Also use an ISO-TIP soldering iron to heat up those points!!

[Carsten Kurz]

There are some cylindrical lithium batteries with 3.6volts, so, that would be perfect for a 0.7v diode drop, but there are also special low drop diodes. I would probably use a gas- or battery operated soldering device. These cert saving circuits/currents are very delicate.

Whatever, I think saving ICPs IS possible with not too much effort or risk. I would recommend checking the cert battery voltage during regular maintenance, at least whenever the ICP clock battery is changed. It doesn't make sense to take the slightest risk if the cell voltage is still high enough.

[Jim Cassedy]

You would think, that in these very expensive pieces of equipment, and given the critical nature of the ICP battery, they would have included a simple battery voltage reading parameter in the UI, or at the very least a "low battery alert" LED or error message in the design.

[Frank Cox]

Why would they do that, when they can just say "$4000 please" and you take it or leave it.

I guess they want to generate ongoing profits after selling you the equipment in the first place.

[Leo Enticknap]

The big problem is, if your replacement effort fails, in trying to explain to a customer why you took a component from their projector that worked, and then broke it. We all know that it would break by itself before too much longer if we did nothing, but if they suspect a scam, you could have a very sticky situation on your hands.

[Carsten Kurz]

It's very important of course to communicate this clearly. It helps if you measure the cell voltage early enough to give a rough prediction. The customer has to decide. What I describe is not yet a full solution, but it's important to know that it is doable, and it can probably be refined. I think it's also important to see more measurements of these cells of different ages/units to get a better idea about the dimension of the problem and how much lifetime to expect. This colleague has two more units of the same age and he will report back.

[Marcel Birgelen]

I think it wouldn't take much to develop a small circuit board that you could solder onto this board as a "daughter board". This board could then monitor both the voltage of the existing battery and switch over to an auxiliary battery on that same daughter board once the voltage drops below a certain reading. Maybe you can even put two batteries on there, in a replaceable socket and a green LED for every socket, indicating good contact and voltage... Even if such a board ends up costing like $50 in components, it's still a worthwhile exercise. I guess the most challenging part is to neatly mount it on there.

[Carsten Kurz]

I wouldn't mess more than necessary with that circuit, it may not be as insanely sensitive as the GORE board stuff, but, keep it simple, keep it safe. We see that the original battery life is up to expectations, with at least 10 years of realistic life time. Replacing it with a new battery of the same capacity should be sufficient. I don't see that discussion coming up in 10 years again. Well, maybe for those who still buy classic series 2 electronics now ;-)

- Carsten

[Marcel Birgelen]

It's hard to look into a crystal ball right now... The future is cloudy... But I don't think it's unreasonable to expect some of the Series 2 stuff to still be around in 10 years. The whole idea of the circuit board is to reduce the chance of blowing something up. The amount of electronics involved should be minimal. You want to solder a diode and battery to it. Just tacking it on will be pretty clunky. I'm trying to come up with a solution that could be field-installed by anybody with some basic soldering skills.

[Carsten Kurz]

That's what this discussion is about. I think there is not enough space below the ICP board to keep even a coin cell there. One probably needs to wire it to the top of the PCB. There's ample overhead. I think the safest way is as described - solder wires to the battery contacts on the bottom, feed the wires to the top as short as possible, attach a battery of roughly the same capacity (everything soldered, no clips/sockets), clip off the + contact of the old battery, which is easily accessible. Fix the new battery somehow, even with a nasty blob of hot glue or something similar. It really doesn't need to be pretty. Dolby does it on their CAT745 more or less the same way.

I think that is very safe to do with an isolated soldering tool. One would probably need a dead ICP to check wether there may be other ways to get there. It is probably possible to get ground/- somewhere else on the top PCB side. It is certainly not wise to measure more than the cell voltage on a working ICP, as it is too easy to mess with these minute cert keeping currents.

[Marcel Birgelen]

I agree that soldering or mounting anything underneath the ICP board is problematic, I guess you have just a few millimeters of clearance and you don't want your battery to become stuck behind another card or whatnot when pulling the card out.

As long as the battery is in healthy teritory, I guess a simple voltage measurement with some quality voltage metering instrument should be safe. What you need to avoid is creating any kind of short circuit though.

I think that getting wires soldered on the back to the top of the card also isn't the most desirable thing to do. You can't run them along the sides, but there is a bit of room between the face-plate and the sides of the card. Since the ICP is a multi-layer PCB, drilling a hole in it would be close to suicide. I think the best solution is to find some pins on the +3,x V bus of that battery on the top of the ICP. Also, you'd need to make sure that the ground of that cell is connected to the common ground. I guess it will be, but the battery circuit could also be an opto-isolated circuit, for example.

I think the safest tool for the "laymen", which can also be easily field-transported is a gas-based soldering iron.

[Mark Gulbrandsen]

To measure the battery, just use a meter with very high input impedance. >/= to 100K.

[Carsten Kurz]

On the pictures I have of different ICPs, it looks as if the PCB is very tight on the face plate, even if these wires may be very thin, I don't think you could feed them between PCB and face plate. The board has some break outs on the sides and some mounting holes. But I don't like the idea to have something like 30cm or more wire across all the board to use these.

I guess you could use the bottom through-pins of the existing cell just to attach a wired temporary bridge battery, then clip and wiggle of the old battery + and -, finally solder the new battery to the remains of the old battery pins. Secure the new cell with hot glue, remove the bridge battery. You need some soldering experience and determination, but it's doable.

It's a CR2032, you can get them in a variety of types with solder leads, even ready in a shrink tube with wires attached. I can't see any trace going from the - of the cell on bottom or top side, so I assume it's connected to the inner GND layer of the PCB. Again, having access to a dead ICP would make it possible to perform some more measurements than you would do on a living species.

I once thought to simply use the clock battery for the cert memory as well. Probably GND is the same, then we would only need to use a short piece of wire from it's + over to the cert cell lead.
However, I don't think it's safe to connect the clock battery and cert memory circuit that way. Then if the clock battery needs to be changed, we're back where we started...or if the tech changing the clock battery doesn't know about the mod...

[Steve Guttag]

I'd be interested in knowing where the cert battery traces end up and see if a temp battery could be attached there for the duration of the battery swap. That is sort of how GDC did their box-server type mediablocks with four header pins sticking up (a bit too close to the batteries) so you could put a temp battery in place. USL/Dolby also has tried that scheme on the CAT745 though abandoned it when it proved that there was better than a 50/50 chance of bricking the mediablock. the problem with working with such low currents is that it doesn't take much of anything to cause them to fluctuate. I still say that these people should have had a super-cap in parallel with the battery, with a protection diode, and that way one could safely change the battery as the super-cap would always have the ability to hold the voltage for the duration and would also be soldered in.