CurtPalme.com Home Theater Forum

ronaldus · Joined: 25 Dec 2010 Posts: 183 Location: france

Hi everybody,

I order to get a better thermal coupling between the transistor on the rear heatsink of the marquee i was thinking of using the following strips from panasonic to replace the grey strip that's there now:

http://uk.farnell.com/panasonic-electronic-components/eyga121803rv/pyrolyt-graphite-interface-sheet/dp/1800098

It has a thermal conductivity of 2 times that of copper.

It's very expensive but i have a few sheets that i want to try out the next time i open my projector

Any opinions are welcome.

Regards,

Ron.

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

Ask yourself if the existing thermally coupled insulation is adequate to the job at hand.

Is there a problem with those transistors failing due to overheating?

If not, there's no benefit for the additional cost.

Not everything NEEDS to be modified.

deronmoped · Joined: 03 Nov 2006 Posts: 1154 Location: San Diego

If I though I was having a problem with heat in that area, I would try to add a fan. Maybe just add air conditioning to the theater and aim a duct at the PJ.

ronaldus · Joined: 25 Dec 2010 Posts: 183 Location: france

Hi everybody,

Sorry I was not very clear,

I don't want to do this to repair anything. And it has nothiong to do with the shrinking Vertical height i had a post about a few weeks ago.

I thought it would be an idea to fight against some convergence drift by getting a better thermal coupling between the transitors. I think it was mentioned that the red channel of the convergence always drifts different because it is on the opposite side of the PCB side In some posts i've read also that the transistors are not very well thermally coupled.

cmjonson I also posted this for you since you are trying to make a "super marquee' and I though all small bits could help.

Any way it was just an idea. I have the stuff and i'll try it the next time I'll service my projector.

Regards,

Ron

barclay66 · Posted: Thu Oct 27, 2016 9:32 am Post subject:

Hi,

The heat transfer between the transistors and the heat sink is good enough with the insulation strip on a stock machine. This is evident when verifying the heat sinks temperature during operation: It usually gets too hot to touch where the vertical board and the convergence board are mounted. But: This heat isn't dissipated well enough.

If You want a lukewarm heat sink, then You will need to move that heat away from it.
I use a thin piece of wood, bolted to the outside of the heat sink (see pictures). In the middle there's a hole where a 12cm (4.7'') fan will blow against it. The air passes horizontally through the grooves and exits to the sides, taking a large amount of heat with it. The foam (taped to the fan using aluminium tape) forces the fan to suck fresh air form the outside of the cabinet only and prevents any air from the inside being sucked in again.

Convergence and vertical size is very stable now. I only notice a small drift if the room temperature is much lower (~10 degrees) at start up. This drift is lower than 2mm (0.07") on a 114'' screen and disappears after a few minutes.

Regards,
barclay66

P.S.: Please ignore the additional board on the wooden piece. That's my first version of a fan controller which has been replaced since.

ronaldus · Joined: 25 Dec 2010 Posts: 183 Location: france

Hi Barclay66,

That looks like an intersesting addition to the heatsink. Is the fan temperature controlled also?

Isn't the marquee cooling sytem designed to blow from the sides of the heatsing towards the middle?

You are blowing from the middle towards the edges. Doesn't this counteract the heat flow?

According to your results no.

Regards

Ron.

barclay66 · Posted: Fri Oct 28, 2016 9:21 am Post subject:

Hi,

There are two more or less separate air flows on a stock machine:

- LVPS fans: Sucking through the HVPS and blowing through the LVPS

- Belly fans: Sucking from the base plate into the tube housings and the HDM/FGM cages. From inside the tube housings equally into the neck board cages and the metal tube housings. Only a small fraction of the air flow reaches the component side of the boards on the rear heat sink.

The grooved side of the rear heat sink doesn't get any active cooling. It is supposed to be cooled by convection only. For this, the grooves would ideally be oriented vertically, but that would have required a much more expensive production process (milling instead of extrusion).

On my machine I have reversed the direction of the belly fans. Why? Because my Marquee is hanging on the ceiling. Trying to push heat downwards didn't seem to be the best idea and so I simply tried it and it proved to be a good solution. The air coming from the rear heat sink fan becomes part of the flow that gets sucked to the outside via the belly fans. And I have two additional fans that take care of the HDM and the FGM as their cages wouldn't get enough air flow after the reversal of the belly fans.

All of the fans are temperature controlled:

- Channel 1, Overall temperature -> Belly fans
- Channel 2, Rear heat sink temperature -> Rear heat sink fan
- Channel 3, HDM temperature -> HDM fan
- Channel 4, FGM temperature -> FGM fan

The only fans I left alone are those at the LVPS. I just replaced them with low noise parts that at the same time would ensure the same amount of air flow.

Check on tschaeikaei's fan controller project (http://www.curtpalme.com/forum/viewtopic.php?p=456758). He did something quite similar...

Regards,
barclay66