You know how on the old like Penny, um, systems and stuff like you have the CPU and it had a cooling tunnel that would, like mount up with a hole in the side of the case so that the blower only pulled cold air from outside.

We're gonna make one for 5700 blower card.

A lot of people about the early 575,700 sixty's which had these blowers on them.

And then the custom cards came out later.

So we're gonna head the lows, and we're gonna get some stuff and see if we can't make a thing.

Tired of all the extra cables in your system, design has it covered with their new eye on serious power supplies.

Available in a variety of sizes and wattage is the new fully modular ion plus syriza power supplies have what you need for clutter free reliability.

You learn more quickly below, sir, that we got better pipe cutter that I said I needed for cutting the tube because I had that little bitty one.

This is unrelated, but now I can easily cut my two bitter thes air fittings for my dripper system in the backyard because my husky keeps eating my drifters so you keep going to replace him.

And then we got a whole soft and some different PBC connections.

Now we were a little concerned because here's the problem.

Given the size of this opening here, two inch.

As you can see, it's too small.

I think most people would agree.

Next.

Video.

We just blew it to it and turn it on anyway, so we've got this reducer.

Technically, it's a reduced because it's a three to a two inch.

So as you can see, this is the perfect size.

It's a fit like over the opening, but the problem is, if you look in there were actually choking down a little bit.

I think a two and 1/2 would have been perfect, because this is stepping down to a to its a three inch to a two inch reducer.

The problem is, apparently, there's no such thing as two and 1/2 inch PVC pipe at, like, clothes or Home Depot or whatever.

So this is what we're stuck with, but it still fits within the edges of the GPU.

So then we've got this guy here to make the turn, and then I'm not sure if that's gonna be enough to make it to the edge.

If not, I got this piece so we can extend it out to make a major case.

Then we've got a whole saw to cut the opening in our case, which probably won't line up perfectly cause I really suck it, eyeballing things, and I would like to measure, so that's our idea.

Now I'm I would have loved to have just used a riser card and then had a straight piece because any time you introduce, introduce any sort of curve or bend, you do increase the resistance.

Now we're hoping that the radius of this is smooth enough to where it won't really do that.

And I'm also hoping that the choke point or it being choked like this down to a two inch is it gonna be an issue?

But before we do all this, we need a baseline.

We need to know exactly in this chassis in this current ambient temperature, blah blah, blah, blah ba What is our control?

Our baseline?

Because we don't have that.

We have no idea if there was an improvement or we did ourselves a disservice.

Huddles and stole graphics cards.

Okay, so here's our control.

We're gonna use super position, Actually, not heaven.

Superposition is harder on Jeep use.

We've got our chassis fans set to a static 40% CPU fan curve is set to normal.

Although that won't matter to do the Senate bench run.

We'll talk about sin invention a second.

Why we're gonna do that?

And we've got the GPU set to a static 43% now.

The reason why I chose 43% is we tested this already to see where the Max fan curb Waas, um, out of the box and 43 is where it stopped.

So we went ahead and just said it there manually because we want to do is kind of reduce the chance of our test being skewed because of the fan curve moving around, because it's only going so far.

So we want to lock it down, and then we can kind of take a look at the GPS.

Um, clock speed.

Take a look at the A commoner, a tackle meter and all that sort of stuff.

But basically what we care most about is at the end of a 10 minute run Right now, we're gonna record what are Max.

Temperatures are on both edge temp and junction temperature, and then Phil wants to add Attest to this where we're going to test a heat soaked chest because the GPU being a blower style cooler we've said this a 1,000,000 times will not affect the chassis temp nearly as much as something that has multiple fans on it blowing air into the chassis where the chassis fans have to remove that air.

It's removing its own air through its own blower.

So I don't think this test right here is gonna be really affected by this because of the fact that it's already getting extremely cool air because there's nothing else heating up the air inside the chassis.

So what Phil wants me to do is after this 10 minute run and we record these best case scenario temperatures we're gonna run Cinna bench are 20 and so we're gonna get the air inside the volume of air inside of here, nice and warm and then run this test again.

I tend to think that I should set a few course that toe run this bench or this test bench going while or this benchmark going well, Senate bench is running, keeping the air consistently warm.

My concern here is that the chassis air is gonna cool down the chassis faster than the 10 minutes it's gonna take for us to record temperatures with our cold air intake.

So I know we'll see what we decide to do in this 10 minutes is that.

But I kind of feel like we need to keep the air temperature.

And they're consistent to see if this works kind of simulating a game that is very seep you bound as well a cp about yeah, CPU bound as well GP bone.

So that's kind of my methodology there, lots of different ways we could have tested this, but that's the way we're going with.

And there's no vents anywhere on the sides of these chassis is the top is actually, uh, got the filters on their which blocking up a lot of that air but there's no fans up here on top.

So the only exhaust fan is the one right here in the back, which, as I can feel it, is cool air coming out.

So that's why so far a few minutes into this test, we've kind of normalized already with our edge Tempus 60 60 and our junction Tep bouncing between 77 78 our core clock bouncing between 16 90 17.

20 or so that you can see our RPM and our speed.

But they're locked.

So all right, we will go ahead and start.

The bench will start that.

See, if you think I came in, drag a window.

So I decided I'm gonna head and block off the exhaust fan.

So all we have been it's positive pressure from the front because with the exhaust fan right next to that blower cooler, it feels like it's bringing the heat out before we could heat up the chassis.

And for the sake of our test, we need the chassis to get warmer.

But, Jay, it looks like you're causing all these problems yourself.

My wife said, Stop, It is the science, sort of whatever.

So channeling my inner Kyle.

I've decided I'm gonna go ahead and take up my case with Junior.

Did European Kyle, you know, duct tape.

He admitted it, but I'm just what I'm trying to do here is block off the vent right above the CPU cooler.

Once again.

Remember, this is a down fire CPU, so I'm trying.

That was too short.

I'm trying to make it just trying to saturate the temperatures in there.

And then at the end of the day, this is a pretty pointless test.

But I was bored, and I wanted to do stuff science wise.

Well, the tape and stuff, it's certainly warm.

We're sitting at 92 on the junction temp and right around 79 80 on the edge, which is funny because it's still not meekly as hard as we would think it would go if it was the next tea, then nothing for sure.

The irony is, we were like, It's Nora, near as hot as if it was gonna be an intel and invidious set up just kind of funny.

It's like the tables have certainly turned.

So now that we've done that, I guess it's now timeto do the completely useless and unnecessary mod because, well, like commode around.

This is your captain speaking.

If you look at the start of the year Kraft, you could see Uranus, the planet, not your buddy.

So I just realized we're both fortunate and unfortunate that this clears the power supply because it's really long.

This is where a bigger tower would come in handy.

But all my other big towers here tempered glass, which isn't gonna work, Which means that I'm gonna be drilling into huh?

The curved part down, uh, is gonna suck.

Did it get immunity?

Defeated it.

D'oh!

It looks like a super Mario.

I know, right?

That was perfect.

That works too well like that.

Black was fitting.

Right?

Here was an accident.

I'm making a purple man.

A perfect circle.

Got me by script.

Garage fans out there.

She'll run.

Give her a little fire starter and a little in the guy might think that this is unnecessary, but, uh, guy also doesn't have any money.

So I love my script around so much.

I hope someone else watches him.

All right.

Moment of truth doesn't line up a a a factory.

That's how you extend o meter extend.

That is clearly the cleaners cuts of any cut that they have been cut in the history of cuts.

And here's the thing.

It's threatens.

You could screw on a cap when you're not using it to get I'm gonna start with just a Jeep.

You this what we're going to see right now?

I should be the same as when the CPS running right.

So when the CPU is in there making the chassis warm, the GP now shouldn't care.

All right, so our temperatures right now are 65 76 they're bouncing between 76 78 on the hot side and 65 to 67 uh, saying usually more around 65 on the cold side.

So we lost about two degrees versus our control, Which is kind of funny because we have all the tape on here which we didn't have on the control, however, were isolated from the chassis, so that really doesn't make any difference whatsoever.

Now, one thing to point out is our A C just turned on right now to, um we did our first control.

It was 71 degrees Fahrenheit here the a C turns on at 73.

So we were actually two degrees Fahrenheit harder on this particular test with all the tape on the chassis, which means that we truly did isolate the GPU from the chassis, which gives me good hopes for this next test.

I'm expecting to see significantly less temperatures in what we had in our last hot test.

Now, the last test with all the tape on here with the CPU running with sediment are 20 looping.

We saw a temperature of 79 on the cold side and 92 on the hot side.

So we see anything less than that.

Then this stupid stupidity get something.

It's like exactly the same temperatures as it was without the CPU test going and send a bench running back there in the background.

It works who were freaking out.

There's one more part of this test.

I kind of want coming down 10 degrees Theo so clearly that work.

All right, look, here's the bottom line.

This is unnecessary.

Unless you just want to come up with some cool mod theme, then that is kind of neat.

But you know what people were doing when these first came out and people realized they could get hot, especially 5760.

They would find an old cooler off like a 2 90 X or something, because the bolt pattern around the die was the same.

And so you could get yourself a nice heat sink with heat pipes and multiple fans and just mounted.

And if a little sticky heat sinks on your memory, if you want your V R EMS and call it a day, this is a necessary.

However it worked, and I'm happy to report that because I spent about $30 on materials.

But not only did I spend $30 on this, some of you might argue that this isn't science, but, you know, in the wise words of Adam Savage, the only difference between messing around and science is writing it down.

I'm Jay's two cents.