CurtPalme.com Home Theater Forum

[redfox001]

If there is 24V could you not just set two 12V fans in series?

[Nashou66]

Yes but if one fan fails they both stop.

Nashou

[redfox001]

[redfox001]

But perhaps a way to solve that problem would be to use 6 x 12V fans. 3 in the belly of the Marquee below and three to place on different locations like on top of the HD and FM. One could connect the fans crossed so that if one fails two fans will stop but not one board. Just an idea maybe to complex?

Or the extra three behind the neckboards?

[barclay66]

Hi,

From what I've seen, it seems that Julian is going about the same route as I did: Dynamic control of the fans!
Trying to control the fans individually doesn't allow for series or parallel connections as You need to feed each of them individually.
On my machine I have the three body fans, one additional fan that blows onto the rear heat sink, one more for the HDM and the one more for the FGM. I left the fans in the LVPS without control (only replaced them with quieter 12V Fans connected in series -> there it can be done!). The air flow inside the Marquee is reversed. The heat sink fan and the ones above the HDM and FGM suck air from the outside and push it directly onto the heat sinks. The belly fans suck the air from the inside and push it out to the ceiling (the Marquee is mounted upside-down on the ceiling). Then I have temperature sensors in four places (HDM, FGM, rear heat sink and body). For them I've built a board which controls the fan speed depending on the temperature measured.
So when switching on the Marquee, only the LVPS fans will work. When any of the temperature sensors reads more than 30°C, the corresponding fan will start spinning and will spin up as long as the temperature rises. With this setup I get average temperatures of 35-40°C which stay quite stable. Especially on the rear heat sink this is beneficial as it reduces convergence drift to almost nothing...

Regards,
barclay66

[tschaeikaei]

Yes, barclay66,
i am planning to do almost the same you did.
A 5-channel Atmel based fan control, 5 temperature sensors, 5 fan outputs, each independent controlled.
One sensor for the HDM, FGM, rear heatsink, one for the green neckboard. One left for ... make suggestions.
I suppose all 3 neckboards and tubes to be equally hot.
So one 120 or 140mm fan for the big backside heatsink,
3 for the HDM/FGM (at the original location) sucking air from the HDM/FGM and blowing outside the PJ,
one fan for each of the neckboard/tube assembly, one channel left there, too.


My LVPS already has temperature-controlled fans installed.
Arctic F9 pro TC
They are connected at one of the original 24V fan connectors in the LVPS via a
7812 voltage regulator. After installing the fans, i realized them always spinning
at full speed. So if you install these fans (wich i strongly recommend) in your LVPS,
be sure to do something against electromagnetic interference effects.
Those temperature sensors are about 12kOhm at room temperature,
so the floating current through them is very small. I added an 68nF at the end of the cable
and 1,6kOhms in line to rise the temperature curve a bit. They do now spin up at about 38°C
and are at full speed at about 45°C. The hottest parts of the LVPS are not the big heatsinks,
but the small single transistor heatsinks shown in the photo i stole and edited.

Regards, Julian

[barclay66]

Hi,

How do You deal with the LVPS fans and the Standby Mode?
On a normal Marquee, the 24V inside the LVPS are reduced to about 10-14V during standby. This is too low for an 7812 as its input voltage, so it won't produce enough output for the fans. But the fans need to stay active during Standby because some parts of the LVPS are still powered and need cooling.
This is why I don't use temperature control inside the LVPS...

Regards,
barclay66

[tschaeikaei]

Hey, they do work well in standby. I didn't measure the actual voltage they receive, but the do spin at low speed in standby.
I'll go down to the living room and try to capture that on video.
Regards, Julian

[barclay66]

[tschaeikaei]

I did capture it, but the F***ing mobile phone camera is just sh*t.
The swoosh in the video comes from the camera, not the projector.
https://www.youtube.com/watch?v=qDWN_uPq6SM

There is nothing set up on the screen. The projector was aligned to the wall behind the screen
and i just put the table with the projector back after hanging the screen.

Regards, Julian

[barclay66]

Hi,

I have to check out later with sound (can't use it at the office).
At least I can see that You finished Your curved screen. Nice!

Regards,
barclay66

[tschaeikaei]

It is finished, yes. But i am not happy with it. If you stand close, everything looks fine.
But if you go further away, higher and deeper spots get visible. So a moving image
uncovers darker and brighter spots.
I don't know a solution for this, but sand the higher spots. And this annoys me.
Terrible thing, next time i will just build a curved wooden frame and stretch a cloth over it.
Should be done in one weekend. This took me 7 months and wasn't cheaper at all.

Yes, you all were right. Curved in one direction would have been enough.

The video isn't more interesting with sound. All you hear is the sound that comes from the camera and
the relais switching when i turn the PJ on.

Regards, Julian

[Nashou66]

Be Careful with the LVPS fans not moving enough air all the time while powered up. It is there to also pull air first through the HVPS. I would measure the temps in the HVPS section as well. Ind if that gets hotter first I would move the temp probe to the HVPS to kick those fans on.

Athanasios

[tschaeikaei]

Hey Nashou, because of the air sucked through the HVPS and through the front cover holes, i think that there is not much airflow needed through the HVPS.
If the HVPS should be cooled by the LVPS fans, they would not have made those holes in the front cover.
That's how i understand the concept, VDC made.
If the fans should cool the HVPS, the easiest way to go would be closing those holes in the front by simply tape them.
So all air would be sucked through the HVPS.
I also opend the HVPS enclosure and found that the only semiconductors are mounted directly to the big thick metal case.
There also are two bigger coils, i do not know if they get hot at all.
Did anyone measure temperatures at the HVPS yet?
I will watch the temperature of the HVPS case when running and tell you.

Regards, Julian

[Nashou66]

measure near the back not the front, the front it the potted section.

Id be interested too.

The newer HVPS did add a larger heat sink to those semiconductors , the older ones just mounted to the case , so there must have been some failure to those chips top make them do that.

Athanasios

[tschaeikaei]

My projectors are from 2001, i thought i already have a new version.
Part No. X2532 Rev.B4
Model PTV34.9P143X2532
Electrohome # 10-000310-02P
Serial No 80036-2482

Sure i will check temperatures where the semiconductors are mounted, but i think the heat will spread over the whole case.
Regards, Julian

[tschaeikaei]

please delete, double post

[tschaeikaei]

Hello, fast answer wanted
If i do remove the heat conducting stripes which are between the transistors on the big backplane heatsink (astig, convergence and vertical board),
do i have to electrically isolate the metal plates of the transistors from the heat sink?
I have 20g of Arctic MX-4 heat conducting compound, hope that's enough for two marquees.
Regards, Julian

[gjaky]

[barclay66]