Video cards hung upside down

Looks fine to me, max cooling capacity is 200W and your card draws 110W at absolute max (even less of will actually be used/cause-heat-in the main chip that you're cooling with this). Temps aren't bad to begin with using stock cooling, so this should handle it with ease.
 
I had accelero fanless gpu cooler and zalman copper fanless cpu cooler
on my desktop
with 1 casefan and PSU fan it was always cool even with games on full options etc.

wasn't a gaming rig but I likes when it was quiet. still

problem with controlled airflow and sealed case is dust.
then theres filter but it decreases airflow a lot...
now I just have as many holes in the case and one fan pretty much it allows high volume of air to enter the case without creating high velocity anywhere except the outlet fan
it accumulates a lot less dust this way I only blow it out once a year instead of every 6 months
 
On hold due to some twit cleaning out my card. It's being handled but will take some time to recover my money. Got a new card coming.
 
In the mean time let's discuss how the card controls the RPM of the stock fan.

With a 4 wire fan using a PWM speed controller signal, the microprocessor in the fan, determines which cluster of windings on the armature get pulsed and in what polarity, while maintaining a constant voltage and current on the motor. More efficient, quieter, and far less prone to stalling.

The tach lead is attached to a Hall Effect transistor. Every time a pole set goes by it sends a pulse to the controller chip on your device such as a computer mother board. The controller chip is programmed to listen for a set amount of time, add up the pulses it hears and divide by 2. Regardless of the blade hub diameter you get an accurate rotation count. Case and cooler fans only have two magnets. They are mounted on the inside of the fan hub.

The difference in the 2 and 3 pin versus the four pin, is a simple driver transistor that pulses the input power for the fan. The transistor driver signal is the actual PWM logic level signal that would go to the PWM lead on the fan. On a 2 pin fan, the tach pin on the controller is strapped out with a pull up resistor. Same chip just different usage set ups.

I can remove the GPU heat sink and look for the controller chip and see how it's configured. If it's using a power transistor to regulate the fan power, it can be jumped out making the fan plug on the board a constant on power source. I then can cut the trace going to the driver signal on the transistor and solder a wire to it to get the logic level PWM driver signal. On the other side of the chip second pin from the bottom should be the tach pin. I'll have to cut out the pull up resistor and solder a lead to the trace coming from the tach pin on the controller chip.

I can make my own harness. I have plenty of 3 and 4 pin plug housings and pins. I have a crimping tool around here some where.
It's hit or miss if the HWM and PrecisionX programs will show True RPMs after the mods. But I'm willing to take the risk. If it doesn't by unstrapping the tach pin and putting it to use, and putting the PWM system in to play in the fan, it will less likely stall, run quieter, and provide more efficient cooling at lower RPMs. It other words it will be real time fan control instead of letting the temps rise up to a set point, then rev up the fan.

I've seen the specs on the TI and NS brands of the same chip. LM64 It's a real straight forward sort of chip. They're just $1.71 apiece at Digikey but they want you to buy a 1000 of them at a time. I just need a couple. Crap...
 
Last edited:
I'm not real sure why this is an issue? Those cards never got too hot to begin with. I guess "for science" would be a great response to tearing up a card but just keep it clean and the stock solution is fine.
 
Back
Top Bottom