CurtPalme.com Home Theater Forum

the dude · Joined: 11 Sep 2013 Posts: 179

Hi folks,

im back with new insane stuff! While others would say its not worth to spend a penny on crt and shy away even from setting up one of those old monsters i spend money and a looooooot of time to find the final answer to unwanted noise. And with a lot i mean a lot. Not hours or days...

So this project is a water cooling for a barco 909! Some years ago i saw a G70 that was modified for water cooling. It was not so hard, only convergence and RGB output stages needed cooling. But a BR909 is a monster that is loud and has many boards that run quite hot.

For sure a insane, time and money demanding and difficult project. I plan to improve (passive) coolers and to install water coolers. The plan is to have the fans only spinnig at minimal speed (or not spinnig at all) so they get absolutly inauditable.

Also, i plan i install many temperature sensors. This is some kind of luxery because once well designed individual temperatures of heatsinks may not be of interest, but on the other hand this information can be interesting and sensors are cheap. Beside of that i will use a thermal imager for finding hot components.

This is a long term project, and i will work on it without a hurry. I will try to document as much as a can.

Enjoy!

the dude · Joined: 11 Sep 2013 Posts: 179

1 Wire interface

This is the 1wire interface i build. It uses the ADuM 1251 chip from analog devices, wich is a remarkable IC. It provides full isolation up to 2.5 kV with almost no external components! It uses a quite new technology wich utilizes small on-chip transformers to transfer the bus signals instead of photo couplers. It also provides all the logic required to implement bidirectional bus communication. It is designed for I2C, but works with 1wire as well. The isolation will protect the busmaster from arcing and other disasters.

Why 1wire? Because its easy. Many sensors can be atached to the bus, and with the DS18B20 a cheap precision temperature sensor (2 EURO per chip) is available that provides a build in unique serial number and comes in a TO92 case. Also i can use my home automation system (Loxone) to read the sensors and show the values in the app. PC Interfaces are available as well.

I modified the mainboard to distribute the bus to all boards (no photos of that).

(to be continued)

the dude · Joined: 11 Sep 2013 Posts: 179

Cooling the PFC

The PFC has a small and noisy fan installed. But unfortunally the heatsink is at a position wich makes a cooler difficult. So i decided to install a passive heatsink on top of that. A small water cooler may be installed later in case the PFC will run to hot.

The heatsink is connected to GND_MAIN. This is a bad thing because when it touches anything this will give a big firework. It is also a bad thing to connect hot-to-mains heatsinks to ground, because a isolation failure can damage the set when massive currents run through the groudplane due to the voltage drop. For the PFC i connected the heatsink to the upper metal sheet wich is connected to the chasis. For the PSU and the G2 board i installed a secondary (protective) ground on the mainboard using unused pins of the connectors (no photos of that, sorry).

Then i used my milling machine to remove some part of the heatsink and made a isolator. One of the voltage regulators must be isolated with a isolation pad as well.

Isolation:

the dude · Joined: 11 Sep 2013 Posts: 179

A part of the original heatsink was milled away to make the connection to the new heatsink...

the dude · Joined: 11 Sep 2013 Posts: 179

Then a new custom heatsink was made and attached

**Curt Palme** · Posted: Mon May 23, 2016 8:13 pm Post subject:

This is great, nice work! Smile

the dude · Joined: 11 Sep 2013 Posts: 179

Thanks Curt! Very Happy

the dude · Joined: 11 Sep 2013 Posts: 179

Here you can see a small PCB for the temp sensor. It contains a capacitor, a resistor and a zener diode to provide power to the sensor. Also you can see how the cooler was milled on the backside to provide proper space for the choke and the resistor.

the dude · Joined: 11 Sep 2013 Posts: 179

The next bad source of noise is the tray for sure...

And the most demanding part for cooling are the convergence endstages for sure. So this was my next target. I used all the space available to make a cooler. This is much oversized i think, but my measurements showed that the cooler was only 1,5 ° celsius warmer than the room after 20 minutes free run (no video source). So temperature drift should be no issue any longer Very Happy

The cooler was milled of 2 sheets of aluminium, 10mm strong each. The black coolers are chip coolers (36x36x12 mm), they are glued with termal conductive epoxy resin.

Also the cooler has many screws, much more than needed for sure. But this will cause great preasure and give good termal conductivity beetween the parts wich are sealed with hylomar jointing compound. A few srews cost only time to drill holes and cut threads, i like it much oversized because i don't want the coolant to kill my set.

the dude · Joined: 11 Sep 2013 Posts: 179

The DS18B20 temperature sensor was glued directly to the original cooler

the dude · Joined: 11 Sep 2013 Posts: 179

The water terminal are made of brass, 15x15 mm. The tubes are soldered into the brass blocks. The different materials are not a problem, because i use a mixture of automobile cooleant fluid and deionizised water. The glycol inhibits corossion and the organic salts in the fluid will passivate the aluminium and protect it from a galvanic reaction. Such a protection is needed anyway, because the other components (especially the radiator) are made of cooper.

cmjohnson · Joined: 03 Apr 2006 Posts: 5180 Location: Buried under G90s

You, sir, are a madman. And I really like that in a man! You almost (Almost!) make me wish I had a 909 just so I could appreciate your work all the more.

You are one of a few people who think like I do. If something can be made better, try it!
The Titanic (we call it the CRT) is sinking, holed by a cheap LCD iceberg that won't last til summer,
but we're still trying to figure out a way how to make the engine create more horsepower with less
coal shoveling needed.

the dude · Joined: 11 Sep 2013 Posts: 179

Nice words, i like to be called mad sometimes... thanks!

the dude · Joined: 11 Sep 2013 Posts: 179

The final mounted cooler!

the dude · Joined: 11 Sep 2013 Posts: 179

And now some results..

The PFC was 47° Celsius in freerun mode after 1.5 hours. Not sure what happens when room temperature raises and a video signal is applied. Maybe some small water cooler is required, i will see later.

The fan monitoring circuit was defeated and the conv fan was disconnected. I run the set for about 20 minutes. The conv cooler was only 1.5° over room temperature after that. But the Controller cooler was 52°. Hmmm. With the metal cover installed on the conv board the controller is not cooled by the fan, so io think it gets that hot even in a original set. When the 2. fan gets disconnected later on it may be a good idea to improve the heatsink or install a small water cooler. But anyway, 52° is Ok right now.

I made some pictures with a FLIR E5 thermal imager. The temperatues of the aluminium heatsinks are not very presicse, because they are reflective. Also the driver transistors under the water cooler are colder than they appear, the cooler is reflecting the IR light that comes from the D/A converter section on the controller board.

The coolest part is the convergence output now Cool

the dude · Joined: 11 Sep 2013 Posts: 179

So whats next?

Next part will be the focus board. Space is more limited there. Then i will see if further cooling is required or not. Also i will make a monitoring board to monitor water temperature twice, once for external readout and once for shutdown. Also a flowmeter will be attached. A Atmel controller will be used for that. The connector i attached to the controller board has the FAN_FAIL signal connected. So if the pump is failing for example the set would generate a nice fan fail error on then screen and shut down.

After that the cards in the rear part will get modified.

For sure it will take some weeks until i find the time to complete the next steps. Stay tuned!

**kal** · Posted: Tue May 24, 2016 1:09 am Post subject:

Nice! I love to see well documented stuff like this that's over the top. Well done.

Ao you mention wanting something 100% silent, where are the radiator(s) located? In a different room? Usually the radiator has fans on it. Any pictures?

Kal

the dude · Joined: 11 Sep 2013 Posts: 179

Not sure about the radiator(s). I guess one big with very slow spinning fans will do. Right now i have a big (420x140x80mm) PC radiator attached for testing with no fans, this is enough for the single cooler. But i expect to require fans. Its no problem to but it in the next room, i will decide this at last when i know how much cooling is required.

km987654 · Posted: Tue May 24, 2016 8:37 am Post subject:

What can anyone say but WOW Thumbs Up

Its really great to see custom work especially so professionally done.

CRT definitely worth the time.

cmjohnson · Joined: 03 Apr 2006 Posts: 5180 Location: Buried under G90s

I don't suppose you have given any consideration to immersion cooling in a big tank full of 3M Fluorinert?

That's solve those cooling problems!

But you'd have to create a leakproof method of sticking the lenses out the side of the tank.