2 Stage Cascade Build Log

[Gunslinger]

MIG welding will be much better in the long run.

[quintus]

MIG welding will be much better in the long run.

Becuse of rust?

[SteveRo]

ok so -98 at 300w - looking good

[Gunslinger]

Becuse of rust?

No, ease of use more than anything. MIG welding is pretty easy to do.

[wdrzal]

You guys who have 2 gas torches like acetylene/oxygen can buy flux coated brass rods to gas braze steel............no need for a mig welder.......But if you do buy a mig get one that uses gas as a shielding gas. fluxed core wire is expensive and doesn' t weld as well.

[n00b 0f l337]

Fluxed core is better for certain jobs, but normal is indeed considered better. Flux however can be cheaper, doesn't require a purging gas, and on small jobs like ours is more then good enough with proper application of skill.

[[XC] gomeler]

Looking good Scott Regarding the carts I'd estimate how long it would take for you to fabricate one, multiply that time by ~3, and then decide if it is cheaper to just buy a cart. I have a feeling it'll be cheaper to let the factories in China pump out the carts and then brace them with a little angle or continue using the wood as you have been.

No, ease of use more than anything. MIG welding is pretty easy to do.

A monkey can do it. By far the easiest of the options. If you can hot glue pieces of cardboard together you can MIG weld.

Becuse of rust?

It takes a bit more skill to braze steel together I have found. Way faster and cleaner to MIG and grind. I tried to gas braze steel before, it didn't end well.

You guys who have 2 gas torches like acetylene/oxygen can buy flux coated brass rods to gas braze steel............no need for a mig welder.......But if you do buy a mig get one that uses gas as a shielding gas. fluxed core wire is expensive and doesn' t weld as well.

Agreed on the gas shielding. Used an argon shielded MIG and it was awesome. Sometimes I wish I had a job that would warrant buying one.

Fluxed core is better for certain jobs, but normal is indeed considered better. Flux however can be cheaper, doesn't require a purging gas, and on small jobs like ours is more then good enough with proper application of skill.

Meh. Argon doesn't evaporate, for light welding it'll last long enough. I think I was using a 75/25 Ar/CO2 blend and it was silly easy.

[SteveRo]

The cascade is working great - evaporator head holds at -109 to -104 under load depending on the ambient temps (really heats up the room!).

The temp sensor next to the socket, touching the IHS is showing deltaT of ~-80 relative to ambient under load

Much thanks to Mr Dumper for what so far appears to be a very nice machine.
My only complaint - probably a common one - start up time - it appears I need to wait 25-30 minutes for stage 1 pressure to drop to 75-85psi prior to starting up stage 2.
This is after about 6 tests (Fri-Sun).

So that is the good news - the bad news is my 980 does not seem to be reacting well to the cold - not overclocking nearly as well as I was expecting. looks like I will be lucky to get 5.2 stable.

Still playing around with the evap to 980 mounting - since I am a newbie - that could be the problem.

Any advice on how to make sure I am getting good connectivity?
How tight should I crank down on the screws?
FYI - using ceramique

[sdumper]

just let it drop to -45c on the first before starting and use a decent bit of paste.

[SteveRo]

How/where do i measure the first?

[sdumper]

How/where do i measure the first?

The first has a k type probe wire. Hook it up to a a temp reader or just wait for the second stage pressure to drop to 105psi.

[SteveRo]

Very good - i will look for the probe wire when I am home tonight - funny i didn't notice it before.

[sdumper]

Very good - i will look for the probe wire when I am home tonight - funny i didn't notice it before.

it runs out of the Heat Exchanger which is the rectangular shaped object covered in black insulation.

[Jor3lBR]

the bad news is my 980 does not seem to be reacting well to the cold - not overclocking nearly as well as I was expecting. looks like I will be lucky to get 5.2 stable.

Still playing around with the evap to 980 mounting - since I am a newbie - that could be the problem.

Any advice on how to make sure I am getting good connectivity?
How tight should I crank down on the screws?
FYI - using ceramique

How are you spreading the paste? This is crucial. Ceramique is the way to go or you could try the Indigo Xtreme that is much better then anything else but requires a perfect cure time/mount.

Plus the evap head should be no where but the exact CENTER of the IHS...

I have somewhere a mod to ad between the IHS and the Cascade evap head to hold better loads (I can't find it right now but it's basically a pure high grade cooper square that sits on top of the IHS entire surface with thermal paste). I will try to find this info for ya.

Don't expect much OC running Linx 20 passes because that will "kill" your cascade when running above 5.2ghz and big volts <- it's normal Steve.

I will hook mine up again soon and we can compare OC's

[kitfit1]

I have somewhere a mod to ad between the IHS and the Cascade evap head to hold better loads (I can't find it right now but it's basically a pure high grade cooper square that sits on top of the IHS entire surface with thermal paste). I will try to find this info for ya.

Can you remember roughly how thick the copper was ?
I'm thinking of trying something similar with my SS. But not shure that it will help holding a load better, because we used to use them with de-lidded 939's to help get over the clod bug.

[SteveRo]

Thanks Jor - regarding applying the ceramique - i put down a good but thin coating, completely covering the top of the IHS with a very small brush.

Also yes - linx is what I was doing and it would run - for up to about 5 seconds and then just reboot.
watching realtemp - the max temp only got to about 28C and then reboot.
tried 5.5 down to 5.3 - all reboot
vcore anywhere from 1.58 up to 1.78v, cpuvtt 1.46-1.48v, cpupll 1.3 to 1.8v, dramv 1.66 to 1.86v, cpu & pcie clock drive both at 900mv, all other v's at stock or just one click up.
Any suggestions whould be greatly appreciated

[kitfit1]

Also yes - linx is what I was doing and it would run - for up to about 5 seconds and then just reboot.
watching realtemp - the max temp only got to about 28C and then reboot.
tried 5.5 down to 5.3 - all reboot

From a Cascade ?

My SS dosn't get +C running Linx at 5ghz. Wonder what's going on there then ?

[[XC] gomeler]

If those are IHS temps then you have a terrible mount. I don't understand why people spread the thermal paste over the IHS. Just put a blob the size of a small pea on the center of the IHS and mount your evap. THEN take it apart and see how the paste spread. This will give you a good idea of what your contact patch is. Remount a few times till you figure out what is obstructing the mount and then you should be good to go.

[SteveRo]

@kitfit - realtemp stays below 0 when you run linx @ 5ghz under your ss?
if i've got that right - sounds like i have a bad mount.
what cpuv where you running - i was runing up to 1.78v when at +28C.
I was at ~+18C even down to ~1.62v

[SteveRo]

@gomeler - ihs seems to be staying ~-80C under load but realtemp goes as high as +28C just before rebooting.

[kitfit1]

@kitfit - realtemp stays below 0 when you runlinx @5ghz under your ss?

Yep, have a look at the 975 link in my sig.

if i've got that right - sounds like i have a bad mount.
what cpuv where you running - i was runing up to 1.78v when at +28C.
I was at ~+18C even down to ~1.62v

My Vcore was set at 1.64v with Vdroop enabled for 5ghz, that of course is with HT on as well. With my chip, anything over 5ghz needs HT to be off. So getting to 5.3ghz/5.5ghz even with more Vcore, still keeps Real Temp at -C temps.

[sdumper]

bad mount

[[XC] gomeler]

Oh.. don't use software to measure temperatures. Get a probe on the evap and a probe on the IHS and use those for temperature measuring. I wedge my probe in the little dip in the IHS where the PCB numbers used to be written and I fill that area with ceramique and it is very responsive to temperature change.

[sdumper]

I would at least initially put a small blob of ceramique in the center of the CPU and screw the mounting down tight.

It just sounds to me like a bad mount.

[Buckeye]

I would at least initially put a small blob of ceramique in the center of the CPU and screw the mounting down tight.

It just sounds to me like a bad mount.

As SDumper said here.

I follow that up by turning on the PC and bringing it into the BIOS and let it sit there with the cooling unit turned off. I do the same thing with LN2 pots.

Watch your evap/pot temps with your probe and when it gets to about 60c shut it down.

This lets the heat really spread everything out and gives GREAT contact. Its so good that you may have to do the same thing to be able to remove the evap or pot