Film-Tech Forum ARCHIVE: Barco DP4K32B Overtemp on red chip

This is topic Barco DP4K32B Overtemp on red chip in forum Digital Cinema Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=16;t=003474

Posted by Jesper Meng (Member # 7575) on 10-18-2018, 02:26 PM:

I have this random failure where the red chip overheates. The other chips is 45 - 47 but the red suddenly rises to 61 - 64.
I have had the failure once or twice before, but it has stopped again.

I vacuumed the filter before leaving the workshop.

Posted by Stephan Shelley (Member # 8587) on 10-18-2018, 06:29 PM:

Your filters may need more than just vacuuming. They can collect stuff that won't vacuum off. You may need to soak them in a solution of washing soda. This is not the same thing as baking soda.

Also check your liquid cooling system as well.

Posted by Marcel Birgelen (Member # 6801) on 10-18-2018, 06:55 PM:

To me this sounds more like an erratic temperature sensor than a real overheating problem.

I've seen a single DMD overheating after a heatsink came lose on one of them, but in this case, the overheating was pretty consistent.

Posted by Dave Macaulay (Member # 813) on 10-18-2018, 10:21 PM:

I agree with Marcel. The liquid cooling is a single loop from one DMD heatsink to the next, so they all get roughly the same cooling. If flow is restricted all three DMD temps would rise... not the same rise as each color has a different heat load, but all three would be higher.
I don't remember what happens if a temperature sensor is disconnected - whether it reads unusually high or low or if it triggers a "sensor fail" error when the signal is outside its normal range.
A defective sensor or dodgy connection somewhere the wiring could give a believably high reading though.
I might swap the sensor cable plugs between red and blue (or green) and see how the readings change - red staying hot means a "real" reading, if the high reading is on the other color then it's a sensor issue. These wires plug in at the connector array on the backplane, behind the button panel. The connectors are labelled although it's a bit hard to see the text.
I do not think a sensor can be replaced without partial disassembly of the light engine - if you could get a replacement sensor (unlikely?). This is not a job I would try in the field anyway.
If the cold mirror has been damaged and reflects too much IR light this would mostly go to the red DMD and cause high temps... but that would not be intermittent.

Posted by Marco Giustini (Member # 4544) on 10-20-2018, 06:02 PM:

air bubble in red heatsink? Unlikely but worth draining the coolant to check (when was last time coolant was replaced?). Also check the pump is working fine - it may just not be pumping enough. My rule of thumb was that if you open the reservoir, you should see the flow of water making a little bump on the water surface.
Finally, if your filters are cleaned then you should be able to see through them. If not, wash them with hot water and soap/soda, depending on what's on them.

Posted by Juan Jose Vazquez (Member # 10496) on 10-20-2018, 06:28 PM:

I had the same fault but in a 20C projector, the fault was caused in an antifreeze duct, which was blocked and the circulation of antifreeze was not allowed. There was no "bubbling" in the antifreeze tank. The solution was the replacement of Light Engine

Posted by Stephan Shelley (Member # 8587) on 10-20-2018, 06:54 PM:

You should double check whether it is the red dmd or red tech over temp.If it is the red tech then the peltier cooling system is overheating and is an airflow problem. Try pulling out the main air filter and see if the temp goes back down.

Posted by Leo Enticknap (Member # 534) on 10-20-2018, 09:46 PM:

quote: Juan Jose Vasquez
The solution was the replacement of Light Engine

$25k. Yikes.

I tend to agree with Dave, that a failure of the cooling system is unlikely, because such a failure would be very unlikely to cause overheating in one DMD but not the other two. However, given the cost of a light engine failure, I also agree with Marco, that it would be a good idea to check the cooling system for leaks, then flush out and refill it, out of an abundance of caution.

That having been said, all but one of the Barco cooling system failures I've had to deal with have been in Series 1 models with the pressurized system. The only Series 2 cooling-related breakdown I've encountered was in a 20C in which the pump failed - not surprising given that it had run for 12,000 hours. This was a Bel-Air Circuit installation, in which the projector had gone for many years without any maintenance whatsoever (the air filters were beyond cleaning). On that occasion, the system sensed a pump failure, and shut of the lamp before any harm could be done to the light engine.

Posted by Leo Enticknap (Member # 534) on 10-20-2018, 11:30 PM:

Afterthought (it won't let me edit the post above - timed out): 45-47 seems pretty hot to me as a regular operating temperature. High 20s to mid-30s is what I'd expect to see for DMD temperatures on a B series. If memory serves me correctly, the upper limit given in diagnostics > actual > temperatures is 60, but mid-40s is still warmer than I've ever encountered one of these operating without any fault.

Posted by Dave Macaulay (Member # 813) on 10-21-2018, 08:24 AM:

In the maintenance test, after 1H of black pattern at high lamp power (6kW lamp) I usually see low 40s DMD temps on all three chips.
Another note - the DMD and TEC heat is entirely dumped to the coolant... there is some air cooling of the formatter chips and the general prism assembly. Clogging the front filter won't noticeably increase the DMD temperature. A badly blocked radiator filter will increase the coolant temperature a bit, but I think the cooling system has a capacity well above the actual heat load and tolerates that rather well.

Posted by Randy Stankey (Member # 64) on 10-21-2018, 05:00 PM:

Many (most) electronic temperature sensors use a fine piece of nickel-chromium wire as the sensing element. Ni-chrome wire has a positive temperature coeffieient. That means that its electrical resitance increases as the temperature rises.

The sensor works by measuring the electrical resistance and calculating the temperature based on the resistance. If the resistance increases, the temperature sensing device infers that the temperature has increased.

But...if something interferes with the electrical connections between the temperature sensor and the device and increases the resistance, the device will incorrectly infer that the temperature has increased when, in fact, it hasn't.

I have seen a similar thing at my job.

I (we) have to maintain several baths of chemicals at a particular temperature and we use sensors/thermostats to regulate the heating elements.

Most of these chemical baths are pretty corrosive... acids, caustics or cyanide... and the fumes from these chemicals boil up and rust/corrode just about every piece of metal in the place.

I recently saw the temperature readout on one tank starting to creep up when I knew that the actual temperature couldn't be rising.

Come to find out, the connections between the temperature sensor were corroded, causing the thermostat to give false readings.

I guarantee that the atmosphere in your projection booth is nowhere near as corrosive as the place I work but I think it would pay for you to look at the connections inside your projector to see if there are any loose or corroded connections.

Posted by Stephan Shelley (Member # 8587) on 10-24-2018, 07:06 PM:

I worked on a DP-3000 that had a high tech temp and it was airflow. They had updated it to the wire mesh filters and while the vacuumed it regularly it got clogged with black goo and soaking it in washing soda removed it and restored airflow.

Posted by Leo Enticknap (Member # 534) on 10-26-2018, 11:44 PM:

I once saw what I suspect was a clogged (with popcorn oil) mesh filter in a DP3000 FUBAR a cold mirror bottom fan, too. Couldn't find anything else amiss that might have caused the fan to die.

While I applaud Barco's thinking with the mesh filters (preventing end users from having to buy a set of disposable ones every few months), the problem with them is that if they are used in even a slightly oily or greasy environment, getting them clean is a task that would make being dragged into a cellar and introduced to The Gimp seem positively pleasant. At least with the NEC paper ones, it doesn't matter how much slime, goo, or grease they've attracted: in the trash, job done.

Posted by Chris Markiewicz (Member # 1415) on 10-31-2018, 09:51 PM:

Working with Barcos in drive-ins, we found that Zep 505 degreaser in a spray bottle worked well to clean the filters. I set up our booths with cheap laundry sinks to spray down, wash and rinse the filters. After air drying they looked and worked like new. The key was rinsing them immediately after using the degreaser and getting them absolutely dry before reinstalling.

Posted by Sean McKinnon (Member # 612) on 11-02-2018, 05:50 PM:

I have found the red dmd to cause over heating alarms first in the event of inadequate cooling. I would check the liquid cooling and the fans behind the sheet metal plate tha is directly behind the light engine on the non operator side.

Posted by Matthias Grausgruber (Member # 6884) on 11-03-2018, 04:56 AM:

had the exact same problem on one of our DP2K-19B. something blocked the liquid cooling.
we checked the flow from clamp to clamp until we found the faulty part. i think it was the light engine part of the cooling curcuit... we got rid of it with a hand pump. it's definitly a high risk operation, but much cheaper than a new light engine (if succesful)...