Gomeler's Build Thread

[XC] gomeler · 01-05-2007, 07:55 PM

It seems like everyone is having a build thread so I decided to jump into the game and make my own. Here is my latest unit, I built this unit with the goal of benching with my current E6400 and future growth with quad-core processors. Sometime in the near future I'll be building a 300 watt dummyload with a potentiometer to help simulate various processors and then I can show more scientific results but for now I have to settle with my E6400.

Evap: I started with a 2"x2"x.5" block of copper and drilled multiple holes over the center in hopes to help increase surface area and facilitate heat distribution. Then 3 copper endcaps were cascaded inside of eachother to form the evap. This is a very cheap design (under $15) and is easy to make when you haven't got access to cheap machining services.

Compressor: This is another Rechi 1/2HP rotary. The previous Rechi that I had so many problems with was DOA so I got a replacement and it fired up just fine. It's a bit on the loud side but that is to be expected as it was both cheap and a rotary.

Cap tube: I used 2.7 meters of .028 captube, this was a rough guess and while I think I should have used a slightly shorter cap tube all has worked well.

Other odds and ends: Got a nice condensor from a junk yard, installed a sight glass as I had never seen the inside of a system before, and a liquidline filter/dryer was installed right before the captube.

Mount: Once again due to my lack of machining services I devised my own mounting mechanism. I took a .25" plastic cutting board and cut 3 squares that fit the bolt-pattern over the LGA775 socket. Then I drilled 4 bolt holes and 1 suction line hole in each plate, cut a notch from the outside to the center, and forced them onto the suction line above the evap. Another slightly larger plate with 4 matching bolt-holes was used as a backplate on the underside of the table.

Bottom of Brace

Top of Brace

The motherboard rests on a layer of neoprene and the backside of the board is entirely coated in a light layer of non-conductive grease. The socket was sealed with grease, the area below the processor was packed with grease, but the pins went ungreased. Grease was also applied to the SMT devices around the socket and a ring of grease roughly 1" wide from the socket was laid down to form a seal. A .5" neoprene gasket was made to cover the area around the socket and then a thick ring surrounding the evap formed another seal. Once compressed together there was an airtight seal around the evap.

Results
I just realized that the camera that had all my gauge shots while I was tuning the system is the same pos polaroid that refuses to work. From memory the system unloaded had between 1 and 3 psi and the highside ranged between 160 and 190 psi. The pressures varied with the fan speed as the temperatures shifted around, I just recorded my lowest temps with the fan @ fullspeed with -49 celsius registering on the Fluke 16. This temperature is taken from the probe that is squished against the outer side of the baseplate.

Processor: E6400 @ 4000MHz @ 1.55 Volts
Memory: Crucial 10th Anniversary @ DDR2-1000 @ 2.2 Volts 5-5-5
Motherboard: Gigabyte DS3 unmodded

Idle temperatures just sitting in windows are a flat -37 celsius on the evap probe. In the bios the temperature was reading -56 celsius which makes me suspect that the probe is either not making complete contact or the bios is reporting false data.
Load temperatures were recorded after ~33 minutes of dual 32M runs. After the 32M runs the evap was averaging -22 celsius, it would bounce between -22.3 and -22.2. I suspect if I had a better designed block I could pull much better temperatures but I'm happy. With this unit I was able to pull stable 4200MHz SPi runs back to back to my hearts content but anything past 4200 and the system crashes, my ram is suspect as no matter what multiplier it crashes. Been a blast building this unit and I love how easy it is to move it around on the casters. Next step is a 2 stage cascade with all the trimmings.

n00b 0f l337 · 01-05-2007, 09:03 PM

Very nice job Gom, I suggest modding a shroud to actually hold a fan properly though.

[XC] gomeler · 01-05-2007, 09:14 PM

Thanks noob =) I have a few things left to do to it. First off is a 90 second delay relay so that the compressor will have ample time to cool the system. Another is a shroud that'll seal up the fan around the condensor and also a simple duct to duct all this "hot" air around the compressor and help cool it. Lots of ideas, was eager to post some pictures before I broke it.

{.bLanK} GoD · 01-05-2007, 09:27 PM

Quote:

Originally Posted by gomeler

was eager to post some pictures before I broke it.

Haha, and fair enough too.
Nice job there Gom.

Quote:

Originally Posted by gomeler

I had never seen the inside of a system before.

Oooo I found this we treasure on Xs a few weeks ago you will want to see. Can't remember which thread it was in so I copied it, and put it up on my server.

[XC] gomeler · 07-31-2007, 08:26 PM

Thread revival, forgot I had this! This particular build is a Mach II GT gut/swap and I'd have to say it's been one of the most difficult of all my builds due to the space restraints. V2000's and custom cases are great with 20+ inches of depth to work with, even VapoLS' are nice with 19", but 17" is cutting it close. You'll see in the pictures, there is no wasted space.

Compressor: Danfoss NF9FX
Condenser: Chilly1 single-stage blue shroud, 120vac fan.
Filter/Dryer: Supco SUD111
Metering Device: 8'4" of 0.031" capillary tubing
Evaporator: Chilly1 2007 evaporator + mount
Misc: Suction Line Accumulator with ~4' of captube wrapped
Charge: R22
Target Range: Maximum of 300w but still capable of running in the 1XXw range, tuned for 14 degrees of superheat at 225w.

Observations: Power consumption per watt removed with R22 compared with R507 was astounding, the high discharge pressures of R507 coupled with hot ambients (33 celsius) stressed the compressor considerably. The sweetspot on this build was in the 225-250w range.

Performance: Measured utilizing the Mach II GT controller temperature probe attached to the evaporator. I had no Q6600 to test realworld performance with(waiting on a G0) but I dummyloaded it from 0w to 300w with a 6-wirebound resistor load with .5" x 1.5" x 1.5" coldplate. Dummyload was insulated on the bottom and 3 sides, un-insulated side has 4 bore-holes used to measure temps across the coldplate. Ceramique makes contact between the resistors and the coldplate and AS5 was used between the evap and coldplate (ran out of ceramique last week

)
Load/Temp
000w/-58C 065w/-48C
130w/-44C 180w/-40C
200w/-38C 250w/-34C
275w/-32C 300w/-27C

Conclusion: For high-load situations this system will run happy as a clam but as mentioned the sweetspot is in the lower 200w range. Fan noise was also an issue but there is only so much that can be done to avoid that. I believe the original Panasonic Panflo's were rated at 80CFM but the 120vac (Sunon?) fan included with the condenser easily pushes 110+ CFM which is extremely useful in keeping high-load pressures down but bothersome at lower loads. I think a potentiometer would be a nice addition to cut some of the fan noise.

Images: Images hosted on my website, let me know if they don't appear.
Initial Parts

Rightside

Leftside

[XC] gomeler · 08-06-2007, 05:17 PM

Compressor: Danfoss NF11FX
Condenser: Chilly1 single-stage blue shroud, 120vac fan.
Filter/Dryer: Supco SUD111
Metering Device: 8'4" of 0.031" capillary tubing
Evaporator: Chilly1 2007 evaporator + mount
Misc: Suction Line Accumulator with 5' of captube coiled inside.
Charge: R22
Target Range: 275-300w, capable of holding 350w before crashing. Tuned to 12 degrees of superheat at 275w.
Observations: Went with R22 again as I didn't fancy the 325psi highside it was pushing with 275w on evap. Lost 3 degrees on the evap but dropped discharge pressures by 60 psi.

Performance: Dummyloaded again with my 6 resistor dummy load, capable upwards of 375w. Temperatures were taken at the edge of the face of the evap with my Fluke probe. Found these temperature to be more precise than utilizing the VapoLS probe.
Load/Temp
000w/-54C 065w/-44C
130w/-41C 180w/-38C
200w/-36C 250w/-33C
275w/-32C 300w/-29C

Conclusion: System was tuned for 275w which gives it plenty of leeway handle heavier loads. When going from 300w to 0w the floodback was immense but the system will idle happily at 130w. Acoustics were much better than the Mach II GT above due to the additional 2 inches, it kept the backpressure on the condenser fan down and therefore lowering the amount of noise caused by the blades. Amazing what 2 inches can accomplish.

Media: Hosted on my website as I lost the attachments that were hosted above on the Mach II, not dealing with that again.

The Beginning

Midway

Final

Pentium D 805 idle, 1.54vcore, 4280MHz

Pentium D 805 loaded, 1.54vcore drooped to 1.46v, 4280MHz

For some reason this PD805 really sucks at SPi32M, you can see how badly the 1st instance is lagging behind the 2nd instance. Still was pulling 100% load and tallying up 380w on my Killawatt with a PCI VGA and 1 HDD, no fans. Also is terrible at 3DMark but that's another issue

runmc · 08-06-2007, 05:34 PM

Nice coolers gomeler

[XC] gomeler · 08-06-2007, 05:40 PM

Thanks especially to you, most of the parts came from your shop

[XC] 2long4u · 08-06-2007, 06:14 PM

Cob webs! Yuck!

[XC] gomeler · 08-07-2007, 09:16 AM

Lots of sneezing involved when I cracked open the case on that unit. Didn't expect I'd need a bio-hazard suit on a build

n00b 0f l337 · 08-07-2007, 09:52 AM

Real tight, how are compressor temps?

[XC] gomeler · 08-07-2007, 10:32 AM

Mach II ran a bit warm, around 70C at 300w (with a 34C ambient). VapoLS ran about 10 degrees cooler but ambients dropped around 4 degrees also. The close proximity of the condenser fan combined with the air being fresh from the condenser was responsible for the hot compressor temps on the Mach II. Wish I could have put an SC series compressor into the Mach II as they seem to be built for "tropical" environments so I'd guess they would deal with the heat better.

The_Beast · 08-07-2007, 11:23 AM

very nice coolers

[XC] 2long4u · 08-07-2007, 12:20 PM

What did you do with the old condenser?

[XC] gomeler · 08-07-2007, 12:23 PM

Bookends on my bookshelf

Thanks The_Beast

teyber · 08-07-2007, 09:25 PM

that old one was tinyyy!!!

How woould you find out if the condenser is to small or quality isn't high enough? im mean before you build the unit and see the temps lol.

[XC] gomeler · 10-24-2007, 09:32 PM

Build Name : The Twins
Description: A pair of custom single-stages for two different clients, built side by side. Each build will be tuned to remove 275w of heat, ideal for a hot 65nm Intel quadcore @ 4+GHz.
Components
Case: Ron's Icebox in black
Controller: Unseen's Freeze Pack
Compressor: Danfoss NF11FX
Condenser: Chilly1 Little Blue
Metering: 8'6" of 0.031" capillary tubing, wrapped around accumulator
Evaporator: Chilly1 2007 Evaporator Kit
Miscellaneous: 6' of 1/4" copper DSH/Dampner
SUD115 Filter/Dryer
4" long, 1" diameter accumulator
14" flexible suction line

Pictures

More pictures to come when I receive the accumulators and powersupplies for the controllers. Will update this post to keep everything together.

tim- · 10-25-2007, 02:24 AM

looks very nice, hopefully you will get some great results as well

regards
Tim

[XC] gomeler · 11-11-2007, 08:06 AM

Here's the final photoshoot for Twin 1. Twin 2 is still on the table and will be finished once I post this, just need to sleeve the suction-line and secure it. I'm just going to yank the text and images used in the blurb on my personal website but they get the job done.

Stats
At their core these builds are my standard bread and butter builds, not much variety here as consistency in quality and performance have been my ultimate goals. These units are part of my quad-core tuned line and are capable of holding 275w comfortably and will scale up to 350w albeit with positive evaporator and core temperatures. Here’s what went into these builds, not everything is listed due to the sheer number of little bits and bobs.

* Compressor: Danfoss NF11FX
* Condenser: Chilly1 Little Blue
* Suction: 16″ of flexible stainless steel
* Evaporator: Chilly1 2007 Spiral Evaporator with mount
* Controller: Freeze Pack with MeanWell 12v PSU
* Case: U-T-I.com IceBox with Freeze Pack LCD mounting
* Charge: R507

Performance
These particular builds, being of the quad-core tuned NF style, won’t be hitting crazy low temperatures at idle but they’ll hold loads that would make an Asetek VapoLS breakdown and cry. Under 275w load the system was holding stable at -16C on the evaporator face, -9C on the loadtester core, and a modest 9PSI suction and 265PSI discharge. Somewhere I have a photo of the unit idling around -40C but this isn’t safe in the slightest as the heavy charge causes near-immediate floodback when dropping from a load to 0w load. The minimum load necessary to prevent any floodback was 105w but if this unit is under a quad I can’t imagine it being loaded with anything less than 200w. Noise was rather low although the 110v condenser fan is still a bit loud and drowns out the soothing compressor hum. At 275w load the unit was pulling 402w from the wall, this included the condenser and exhaust fans, the controller, and the compressor.

Images
Here’s the final photoshoot, if you want pictures of the build process let me know but be warned, they are a bit messy. Giving birth to one of these is not easy process.

Excuse my rough appearance, had the flu this entire week so I was unshaven and looked like death. The jeans and hoodie though are my work clothes, haven't washed em in weeks

It was also a passive war with the roommates to get them to take out the trash, as you can see in the background

n00b 0f l337 · 11-11-2007, 08:12 AM

Very nice job, looks quite professional.

killermiller · 11-11-2007, 10:19 AM

Hey Gomeler, did you catch the hoolum stealing your cooler?

I really like those cases, I wish I had some cash.

Nice build there.

[XC] gomeler · 11-11-2007, 10:43 AM

Oh I beat him with an ugly stick.. wait a second

I absolutely love Ron's IceBox, like the thick steel, perfect for drilling into and mounting stuff.

Ashraf · 11-11-2007, 10:49 AM

I wish I knew how to build one already.

Heh, nice job, gomeler.

[XC] gomeler · 11-19-2007, 02:45 PM

Finished up the second twin last week but I just shipped it out today due to an emergency arising at my daytime job and keeping me tied up most of last week. This unit is identical to the first unit except for a minor change in how I wired up the controller. The charge and performance was identical within a degree and 1-3 psi, rather great when standardized stuff ends up being well.. standard. Glad to have this unit going to it's home out in California. This is the second to last unit to ship in this run. The following picture was taken with my camera phone as my roommates D80 is currently acting funky. Didn't feel I needed to take anymore pictures as the above images are an identical unit. This box isn't what I ship it in either, it gets put into a 24x24x24 box and surrounded in foam/peanuts =)

Glad this sucker is done

killermiller · 11-19-2007, 02:58 PM

Got any new builds planned?