Christie Series 2 TPC issues

[Mete Tunca]

We are still having the same problems. Christie support mentioned that it could be the caps on the MB of the TPC. So your assumption is correct about capacitors.


We either let the TPC cool under air conditioning unit or blow air from dust blower to cool it down before replugging. Most of the times this works but takes time.

Or you can unplug-plug connector multiple times in a few seconds to force it to boot it up. This also works.

But eventually you'll smell the burned MB one day which will shortcircuit and power down the whole projector unless you unplug it.

[Christos Gartaganis]

I am curious to understand what makes the Christie TPC so special and why it cannot be replaced with a standard old school laptop running Windows XP. I believe there is some way to make a clone of the Compact Flash drive onto a hard drive.

[Marco Giustini]

A Compact Flash has an IDE connection in a different shape. You can get a cheap passive adaptor and it might just work on a random IDE motherboard.

[Caleb Williams]

I finally got one of these in the lab and did some testing. The TP650H model does have 3 electrolytic caps and a few smaller sealed organic capacitors on the board. I pulled all three electrolytic and tested them and they are all within spec. They are all between the 24v input and ground, suggesting they are only filter caps. If they short then that would immediately cause the unit to stop working, however any number of power cycles would not bring it back. I'm also not able to consistently get this unit to replicate the issue. It seems to only do it when the touchscreen connectors are plugged in, but otherwise boots normally 98% (except the 2% where I accidentally shorted something with solder and kept it from turning on lol). There are many power mosfets on different parts of the board, and there are a few different areas on the board where they cluster together. There are voltage rails for the CPU and GPU, and then a separate power rail for the dedicated flash chip (BIOS). Because the input is 23 volts and the rest of the board requires 5, 3, and 1.5v I suspect something weird is happening with power. The microcontroller ITE IT8528E should be doing some things that it's not, and I think one of the power mosfets or the signal to the power mosfet is to blame. Stuff is really small on this board. I'd need to see if an actual tech could tell us more. The datasheet for the IT8528E is here: https://www.alldatasheet.net/datashe...C2/IT8528.html

Christie has been no help on this, and the best Advantech could offer was an RMA. I'm not sure if our company wants to go that route, as it may just be more cost-effective to purchase a part from Christie. If a component level repair tech wants to hop in on this I'd be happy to share notes and pictures I've taken.

[Marcel Birgelen]

Diagnosing those boards without schematics is like trying to find a needle in a haystack, especially if nothing is obviously broken.

I've had a bad TPC in the part that was given up by the owner and replaced by a new one. I was able to revive it, by throwing the mainboard in a reflow oven.

Advantech is an OEM that works for many companies, they also produce at least parts of the ICMPs for Barco. If this Geode mainboard is a generic part, it might be cheaper to buy it from them than from Christie, especially if you need more than one.

I am curious to understand what makes the Christie TPC so special and why it cannot be replaced with a standard old school laptop running Windows XP. I believe there is some way to make a clone of the Compact Flash drive onto a hard drive.

The short version: Greed and the lack of good right to repair regulations.

It may be worthwhile to try to replicate those TPCs, maybe we should try to make a virtual image out of it, so we can run it forever in some environment like VirtualBox.

[Caleb Williams]

You can definitely take a disk image and run a virtual TPC on windows Hyper-V. The trick would be to figure out the networking to get the pib and tpc talking while it's virtualized. Re-building from scratch on an XP system might be a little trickier, because the installer may not completely cover all the required dependencies, but you've piqued my interest with this idea.

[Marcel Birgelen]

Your host should have two ethernet interfaces, if it doesn't, then you can easily add one using a cheap USB-to-Ethernet dongle. In your virtualization solution, you should bridge that port to the virtual ethernet port of your VM running the TPC.

What will be important is that the Windows XPe version recognizes your virtual Ethernet port. Most virtualization solutions don't have those old Intel "FXP" drivers.

[Caleb Williams]

Your host should have two ethernet interfaces, if it doesn't, then you can easily add one using a cheap USB-to-Ethernet dongle. In your virtualization solution, you should bridge that port to the virtual ethernet port of your VM running the TPC.

What will be important is that the Windows XPe version recognizes your virtual Ethernet port. Most virtualization solutions don't have those old Intel "FXP" drivers.

This is true. The other issue might be, what if you have more than one that you need to virtualize? The TPC is only set up with one network interface with a static IP that is actually identical for all series 2 cinema projectors. There is an internal router on the PIB board that takes care of routing traffic from "outside" (the theater management network) to the TPC. Unless there is a workaround, you would really only be able to host one virtual TPC per host computer. Perhaps a NUC or some other dedicated computer with display might be a better solution in this case.

[Ryan Gallagher]

A NUC, SBC, or thin workstation approach might be more readily implementable by less tech savy if there were instructions to follow.

As for touchscreen, there are cheap driver boards out there for all kinds of tablet touchscreens, but I think just getting a proof of concept going with a traditional touch monitor would be huge,

Hell in a jam who even needs the touchscreen or the monitor, as long as the VNC interface is functioning. You might need some kind of monitor for initial setup, but vnc can do everything after that. Tether an old tablet to the chassis.

[Ryan Gallagher]

Actually, now that I've said that out loud. And Caleb mentioned the issue was most prevalent when the panel was connected.... Can it boot headless without the panel connected? Just VNC into it after that?

Perhaps there is even a chance it IS booting... but no display. Can anyone VNC into their problem units after the pedestal is powered? From memory I don't even recall if there is a power on LED anywhere on it.

[Marco Giustini]

If those TPCs fail when hot, warm the PCB up with a hairdryer. Once they stop working you can cool down sections of the board or single components using cold spray to try to pinpoint what the problem is.
capacitors also can read fine even when they’re not. Though if they’re only on the 24v line, maybe they’re not so critical. But I’d swap them anyways.

[Caleb Williams]

If those TPCs fail when hot, warm the PCB up with a hairdryer. Once they stop working you can cool down sections of the board or single components using cold spray to try to pinpoint what the problem is.
capacitors also can read fine even when they’re not. Though if they’re only on the 24v line, maybe they’re not so critical. But I’d swap them anyways.

Our issue definitely isn't heat, though some of the cases mentioned in the thread that might be the problem. Some of the AMD cpu chips like to run really hot. The TPC will not boot without some power cycles after equipment has been off overnight. I did make sure to remove the electrolytic caps before testing and all were within 5uf of spec, which I wasn't expecting with such old components. Even if they did go out of spec, as long as they didn't pull the line input lower than 18 volts I'd expect they shouldn't cause any issues on the TPC side. If I had some better equipment I'd be looking at the signal generators powering the mosfets that drive power on the rails. Since it's powered by 24 volts there must be several step-down circuits and I'd bet that one of those isn't putting out the right voltage without some "fuzzing" delivered via power cycle. I wish I had an oscilloscope. Ah well.

Actually, now that I've said that out loud. And Caleb mentioned the issue was most prevalent when the panel was connected.... Can it boot headless without the panel connected? Just VNC into it after that?

Perhaps there is even a chance it IS booting... but no display. Can anyone VNC into their problem units after the pedestal is powered? From memory I don't even recall if there is a power on LED anywhere on it.

With some units it is possible that it is booting. As long as you hear the *beep* at the start you know something is happening even without a working display. Some of the TPC models have an HDMI port hidden behind all the Christie trim. I don't know if it actually works. The Christie applications running on the TPC are built with the specific touch screen resolution in mind, so I'm not sure what would happen if you tried to run it headless. I need to find a unit that's a little farther gone to do some testing with. I can't get the one I'm working on now to consistently reproduce the issue with or without the display connected. In all cases though, there is a tiny red led on the main board that lights when power is supplied to the board. Then if post is successfull you'll hear a beep and see display output on the touchscreen. Otherwise, the cpu and gpu just sit and get hot and nothing else happens. If the flash card is installed there is also an activity led that will light and blink after a successful post.

Christie probably will not be any help with this because I doubt they do component level repair on the TPC's. Advantech is the oem, and the TCP is sold as a full product to the Christie team, who add the software and then sell it as part of a projector. I think my next step is going to be to press Advantech to see if they will tell me what could cause post failures on those boards. There is also a bank of dip-switches I am very curious about, as well as several jumpers. They may not tell me anything, but it never hurts to ask.

[Ryan Gallagher]

Ah yeah, post beep would be the clue it had booted. I guess there are network LEDs on the rear too if the bezel is not fully hiding them.

Caps are cheap enough I’d still swap them if you don’t have a way to test ESR easily.

[Caleb Williams]

A NUC, SBC, or thin workstation approach might be more readily implementable by less tech savy if there were instructions to follow.

As for touchscreen, there are cheap driver boards out there for all kinds of tablet touchscreens, but I think just getting a proof of concept going with a traditional touch monitor would be huge,

Hell in a jam who even needs the touchscreen or the monitor, as long as the VNC interface is functioning. You might need some kind of monitor for initial setup, but vnc can do everything after that. Tether an old tablet to the chassis.

Imagine getting this to work on something like the RasPad. Its such a cheap part compared to the oem TPC which may not even be available much longer. This would be a really interesting project on it's own and I'm willing to take a deep-dive on it.

Is there anyone not running the latest 4.8.1(3) software for some reason?

[Marcel Birgelen]

This is true. The other issue might be, what if you have more than one that you need to virtualize? The TPC is only set up with one network interface with a static IP that is actually identical for all series 2 cinema projectors. There is an internal router on the PIB board that takes care of routing traffic from "outside" (the theater management network) to the TPC. Unless there is a workaround, you would really only be able to host one virtual TPC per host computer. Perhaps a NUC or some other dedicated computer with display might be a better solution in this case.

The problem with the NUC will be that the drivers won't match with the TPCs image. Maybe there is a workaround for this, but it would involve customizing the image and porting over the configuration to e.g. new network interfaces. Another problem is that the hardware inside a NUC changes regularly. The beauty of a VM is that the virtual hardware doesn't change.

But let's say you want to virtualize ten TPCs on one host, you'd run ten different VMs on said host. Instead of bridging ten separate physical ethernet connections, I would use a small managed switch with like 12 ports, the switch would fan out to the different projectors. Maybe you can even use the existing network, if it's managed and allows you to configure separate VLANs, that would avoid extra cabling. The uplink port to the VM host would be trunked and every TPC would get its own VLAN. On the hypervisor side, you just need to make sure that you bridge each virtual NIC onto the right VLAN. The way this is being accomplished varies between hypervisor, but most modern hypervisor products support VLANs.

The biggest problem will be to find a hypervisor that still offers 100BaseTX virtual NICs without hacks. VMWare, for example, only offers the "e1000" Intel cards, which are gigabit NICs.