Where can I get custom H20 blocks?

[Top Nurse]

As the title said I am looking for someone who can make me a custom H20 block to substitute for the MOSFET heatsink on my new ASUS A8N-SLI board. Any leads or suggestions appreciated

 

[diredesire]

how complicated of a block are you looking for? just curious

 

[thewhiteguy]

Someone has to say it: Don't waste your time and money, you don't need watercooling on mosfets. Slap some heatsinks on them and make sure you've got at least a little bit of airflow and it will be fine.

 

[Whatsisname]

Seeing that the barb would be as big if not bigger than a TO-220 package, making a waterblock for them would not be possible. Unless there is a large strip of them in a row or something, it's a lost cause.

 

[HeThatKnows]

Do you have access to basic tools like hacksaw, drill (drillpress would be better), and torch? If wouldn't be tough to make your own.

 

[Top Nurse]

The present heatsink is .625 x 3.00 x 1.00 (WxLxD inches). I am using a Koolance EXOS so the 1/4 nipple could be sweat soldered into place.

As far as the heat question goes I sawed up heatsinks for my A7N8X-Dlx board and put them down with Artic epoxy. And they do get quite warm if not hot even with an 80mm fan blowing directly on them...

So I want to cool these down even better and since I am already H20 cooling whats another cooler block?

 

[nweibley]

So I want to cool these down even better and since I am already H20 cooling whats another cooler block?

Unnecessary excess heat and restriction... if it aint broke, don't fix it!

 

[plywood99]

Unnecessary excess heat and restriction... if it aint broke, don't fix it!

I disagree. If we didn't "fix" things, there would be no need for water cooling, overclocking and etc. Just because it isn't broken doesn't mean it can't be improved. Besides, I enjoy looking at all the cool mods people do to their systems.

 

[nweibley]

I disagree. If we didn't "fix" things, there would be no need for water cooling, overclocking and etc. Just because it isn't broken doesn't mean it can't be improved. Besides, I enjoy looking at all the cool mods people do to their systems.

I agree when there is some tangible gain. He isnt going to see any improvement from doing this, besides ambient case temps which really dont matter since he has water and airflow anyways. This is a lost cause IMHO.

 

[mwarps]

Your overclock isn't limited by your MOSFETs, especially not at 1.8V.

But hey, it's your time and money, feel free to waste them as you see fit.

And your heatsink getting hot or warm means it's doing its job. Without it, those MOSFETs would be probably 75-80C. Your heatsink isn't 80C, which means your MOSFETs aren't that hot either.

 

[Erasmus354]

I agree with mwarps, hotsinks being warm or hot to the touch is a good thing. It means that they are cooling and doing their job just as mwarps said. Dont be put off by the fact that your heatsink is hot. Those passive Zalman GPU coolers get quite warm too, but the cards are fine because the heatsink is supposed to get warm.

 

[Top Nurse]

What can I say...I like fixing things that ain't broke, because when it comes to computers one never really knows whether it is broke or not. That's how VapoChills got started...by someone putting their mb into the freezer and turning on the juice

I do find it interesting that everyone wants all this network grade power, a Smart-UPS, a high grade power supply, and then doesn't care what happens to the juice after it gets to the most important part...Your Computer!

Added:

I understand that HS's get warm is as good thing. I ditched my HSF off my 9800 Pro and replaced it with just a custom made heatsink...and it worked better without the fan

BTW, the A7N8X-Dlx board didn't have any heatsinks on the mosfets. Now the A8N-SLI comes with a big old heatsink on them. I'm pretty sure they didn't do it for looks...

 

[v3rt1g0]

BTW, the A7N8X-Dlx board didn't have any heatsinks on the mosfets. Now the A8N-SLI comes with a big old heatsink on them. I'm pretty sure they didn't do it for looks...

They did it because it's a popular enthusiast mod. Nobody but the most hardcore gamers are going to be buying this board, so it makes the board more appealing to them. I've never seen any of the super knowledgeable guys anywhere say that mosfet heatsinks are ever truly needed. Cooling dorks like us just like to keep stuff cooler than the norm.

 

[HeThatKnows]

I've never seen any of the super knowledgeable guys anywhere say that mosfet heatsinks are ever truly needed.

Not that I qualify as super knowledgeable...

A cool MOSFET has a lower resistance than a hot one (lower Rds-on, to be exact), so a cool one will waste less power and create less heat. Additionally, a cool one will switch from on to off faster; again decreasing waste power/heat and improving output voltage stability.

Needed? Hell no. Can it get your overclock another couple MHz? Sometimes, yes. Will water-cooled MOSFETs get you attention at a LAN and increase the size of your e-penis? Oh Yeah!!!

Can you tell that I've done mine? It halved the amount of fluctuation on Vcore, from a 0.06V difference between MBM's reported high and low, to a better 0.03 V variation. Didn't increase my overclock, but it was stable at a lower voltage.

 

[mwarps]

Not that I qualify as super knowledgeable...

A cool MOSFET has a lower resistance than a hot one (lower Rds-on, to be exact), so a cool one will waste less power and create less heat. Additionally, a cool one will switch from on to off faster; again decreasing waste power/heat and improving output voltage stability.

Needed? Hell no. Can it get your overclock another couple MHz? Sometimes, yes. Will water-cooled MOSFETs get you attention at a LAN and increase the size of your e-penis? Oh Yeah!!!

Can you tell that I've done mine? It halved the amount of fluctuation on Vcore, from a 0.06V difference between MBM's reported high and low, to a better 0.03 V variation. Didn't increase my overclock, but it was stable at a lower voltage.

They will switch on faster, yes, but these MOSFETs do not ever turn on of off, at least not in the classical sense. Their gates, at least on the boards I can look at currently, are cut. And as far as being exact.. Rds changes due to increased band gap, increased Recombination-Generation and increased ni [intrinsic carrier concentration]. The increased Ni has the most effect on the transistor, as ni^2 doubles or halves for every 10C you go up or down.

Even still, with a 50C difference, it's still small, and it won't do hardly anything.

Normalizing the operating environment will, however, cause more stable voltages. There is no doubt about that.

 

[HeThatKnows]

They will switch on faster, yes, but these MOSFETs do not ever turn on of off, at least not in the classical sense. Their gates, at least on the boards I can look at currently, are cut..

They wouldn't be much use if they didn't turn on and off, in a classical sense or otherwise. Are you looking at MOSFETs that have the middle leg snipped? That's the drain, which is also connected to the large mounting tab.

Looking at a couple datasheets that I had handy, it looks like Rds-on increases anywhere from 3 to 10 percent with a 25°C warming. Not a huge difference, but it counts (especially with the lower MOSFETs).

Mmm, intrinsic carrier concentration...

 

[Top Nurse]

They did it because it's a popular enthusiast mod. Nobody but the most hardcore gamers are going to be buying this board, so it makes the board more appealing to them. I've never seen any of the super knowledgeable guys anywhere say that mosfet heatsinks are ever truly needed. Cooling dorks like us just like to keep stuff cooler than the norm.

I suppose I could go along with that. Especially when the heatsink isn't even down very tight as you can move it with a Q-tip! However, I can see a lot more reason to have mosfet heatsinks than video card ram sinks.

BTW, I just took off the HSF on an ASUS N6600 GT as I 'm getting it prepped for a H20 cooler. That gigantic thing is all show and no go! From the imprint on the chip and the HS it appears that maybe only 25-30% of the thing was even touching the chip. Also the finish was reminiscent of a saw cut...I'd estimate about a 250 RMS finish. In fact the part that you can see where the fan is has a better finish than the important part that you can't see Though I suppose a big fancy fan helps sell a lot more cards than a 486 HS with a little fan on it

The ram heatsinks were put down with plastic sticky tape...can't see that transferring too much heat.

Yeah that's me...the cooling dork/nerd/tweaker

Added: Just got the chip cleaned off and the top of the die is finished better than most mirrors I've seen in a long time! Putting that piece of %$@! cooler on top of it was a crying shame!

 

[diredesire]

why do you have your AGP set at 1.7v?

 

[Little Grabbi]

I'm thinking of doing something like that myself, with a simple channel design and 6mm ID tubing. With other people it works amazingly well (although it does not improve OC), and it gives you that extra bit of security in zero-airflow systems.

 

[mwarps]

They wouldn't be much use if they didn't turn on and off, in a classical sense or otherwise. Are you looking at MOSFETs that have the middle leg snipped? That's the drain, which is also connected to the large mounting tab.

Looking at a couple datasheets that I had handy, it looks like Rds-on increases anywhere from 3 to 10 percent with a 25°C warming. Not a huge difference, but it counts (especially with the lower MOSFETs).

Mmm, intrinsic carrier concentration...

Indeed. Checked TO-263AB and they are indeed GDS. Strange, but that's the way it is. As far as on and off goes, these MOSFETs, as I said, are always on. All they do is regulate current. The on-off speed you were referring to is only relevant when the MOSFETs have to switch.

 

[Top Nurse]

why do you have your AGP set at 1.7v?

As opposed to what? It ran better when I started OC the video card. I also increased the voltage going to the mb chipset as well.

 

[HeThatKnows]

As far as on and off goes, these MOSFETs, as I said, are always on. All they do is regulate current. The on-off speed you were referring to is only relevant when the MOSFETs have to switch.

Those MOSFETs are switching on and off, at around 100 kHz. The "voltage regulator" on a mother board is full blown switch-mode supply, a DC-DC buck converter. Pulses of voltage/current fed via the upper MOSFETs into inductors for smoothing/storage, with free-wheel current handled by the lower MOSFETs.

If they were use as linear regulators like you're implying, the waste heat would be enormous. Not to mention the huge currents the PSU would need to supply.

 

[diredesire]

As opposed to what? It ran better when I started OC the video card. I also increased the voltage going to the mb chipset as well.

upping the voltage to the video card doesn't do anything as far as pumping voltage... the voltage regulation is all handled on the video card. The ONLY reason that it would "run better" is because the voltage to the chipset is generally closely related to vAGP.

 

[Top Nurse]

I though that I was increasing the AGP bus voltage, no?

 

[Top Nurse]

Okay dudes put your thinking caps on...

I got a tool and die maker friend who is willing to make me a water block for free

So what is the best material? Copper, aluminum, ?

Also should I consider gold plating if I go with copper?

I thought about perhaps a modified "S" channel utilizing a roughing endmill to roughen up the channel. Will use a 1/4" hose barb.

Should I use plexiglass as a cover or is there a better material?

 

[zer0signal667]

Okay dudes put your thinking caps on...

I got a tool and die maker friend who is willing to make me a water block for free

So what is the best material? Copper, aluminum, ?

Also should I consider gold plating if I go with copper?

I thought about perhaps a modified "S" channel utilizing a roughing endmill to roughen up the channel. Will use a 1/4" hose barb.

Should I use plexiglass as a cover or is there a better material?

Copper is the best material, gold plating won't hurt anything. Best material for a lid, in my opinion, is again copper. Plastic lids are much more prone to cracking or melting.

 

[diredesire]

you do increase the AGP bus voltage... but it has no effect on the video card... all the voltages on the video card are regulated on the video card itself... You're probably just putting more stress on your motherboard than anything else...


As far as a block copper is where it's at... Avoid aluminum if you have copper everything else.