Quad Core Phasing and Information : Read it here!

[hipro5]

I'll tell you something......

On the exact same cascade system we have tested two evaps with different cap tubes.....
One X (I don't want to say whose's) and one eleven's evap......
NOW:
X6800 at 5070MHz and 1.9VCore with X's evap and 5150MHz with the Eleven's evap.....
One evap could handle it at -82*C full load and Eleven's evap could handle it at -101*C full load..........HELLO!....-101*C full load?.....-109*C idle and -101*C full load?.....

Is it THE difference or not?......I couldn't believe it myshelf coz I've never used Eleven's evaps before.....As a matter of fact my idea was to build a rounded evap like Prommies' one with 5cm base BUT after I've tested eleven's evap, I drooped this idea.....

EDIT: My English are perfect....esdee's are crap......

[s7e9h3n]

For what it's worth, when Cryo-tek was load testing the Chilly1 evap on my autocascade, he came to the conclusion that the evap was reaching its limits @ ~200-225W. He had a series of graphs from testing with different charges that seemed to prove this point. I have a couple of the print-outs somewhere....I'll try to dig them out when I get a chance, but they're quite interesting. Maybe I should get him to somehow mount one of these evaps on a unit which he builds -> 10HP, -160C @ 3500W should do the trick

[kayl]

For what it's worth, when Cryo-tek was load testing the Chilly1 evap on my autocascade, he came to the conclusion that the evap was reaching its limits @ ~200-225W. He had a series of graphs from testing with different charges that seemed to prove this point. I have a couple of the print-outs somewhere....I'll try to dig them out when I get a chance, but they're quite interesting. Maybe I should get him to somehow mount one of these evaps on a unit which he builds -> 10HP, -160C @ 3500W should do the trick

that would be something to see hey

[epion2985]

No. Just my opinion. If you're trying to start an argument then you're an idiot...

Not at all. I was just point out that your comment was fluf as without providing anything proof or evidence you basically said so and so is true because one builder said so and he knows what he is doing, while another builder with no less experience who also knows what he is doing said the opposite. The last thing we need is people who don't know what they are talking about (relatively to the builders) posting in favor of so and so with no explanation, ie spam.

As to why you chose to resort to insults is beyond me.

[Gray Mole]

And here I was hoping you wouldn't justify him with a reply Tom

I've discussed the NEK series with Johann and his results are excellent in many ways.

Will they hold up? Who knows. Are they newer and likely more efficient and better designed? Very likely.

I know and love the NL Danfoss series, from the 6's all the way to the 11's and everyone in between. Power to noise has been excellent throughout the range.

However, I've seen some of Embraco's stuff that blows the NL series away.

I've tried the smaller NEK series and been very impressed.

So I have to agree with Tom. I believe that Hipro's opinion is right for him.

I also believe that Johann's results were fantastic, and from that I'm VERY much looking forward to seeing if his results are going to be the norm, as that would be me switching to Embraco NEK's from Danfoss' NL's as my primary compressor choice.

Time to get to work though :p I've got some coolers to build

Cheers

Gray

[esdee]

even tho im not aware of the technology used on NEK and NL11F i have to say i love NEK series too. they are powerfull and extremely quite!

NL11Fs have proved to be beasts of reliability!

[Ssilencer]

Hipro5

My idea is to build an evap like eleven's which "flows" ALL the gas on the base....and not from up to down and visa-versa.....ALL the gas ON the base.....

Already done for about 2 years in @itor evap

Esdee

1) i don't think base thickness must be more than 5mm. i have the feeling that with large bases we try to remove heat from a point too far from the cpu core itself!
2) mass mass mass... i think mass is important in a weird way.
an evap with more mass has more heat buffer to deal with the load until the systems low side responses

1) Already done for about 2 years in @itor evap base thickness of 3mm or less...

2)this is what all the guys were telling for the lasts years, that's why all the rest made evaps taller and wider and why people thought that @itor evap cannot handle high loads efficiently, NO mass mass mass but SURFACE SURFACE SURFACE .

Eleven's evap looks really really good, have enough surface

Ok, back to the topic, looks like to keep temperatures in the -40/-50c range we will have to use bigger compressors, more current, bigger bills, more noise
Anyway, quad cores are not near at all here in my country,nor the NEKS or even the NL11 or any other thisng from Danfoss more than the SC series, which I don't want to use for single stages, so I have plenty of time (I hope)

[n00b 0f l337]

Hey I'm back and its good to see some discussion guys, there hasnt been enough lately. As for Eleven making more evaps? I guess I just havent seen him around but I thought he had retired from the game. Is he still selling evaps? If so how can I contact him ASAP! :P

[Gray Mole]

Looking at Eleven's evap design closely I can see what you mean.

I've always liked that milled design, and it leaves a lot of room for surface area that it's difficult to produce on a stepper or spiral.

I'd like to see it with a 2-3mm base, and taller fins, that are 1.5mm thin. More of them too.

It's normally all down to the cutters available though, and the quality of the mill. I think that extending the fins, and thinning them out, would produce the best balance of mass and surface area. I've always thought that mass was more important, but as long as there is enough to keep a certain amount of heat storage (or the lack of it ) then the rest is up to a good surface area.

It's part of what makes @itor's evap so special.

I'm starting to build a collection of evaps again, and if I can get ahold of Eleven's (whether new or old rev.) and some others, I can take the data from previous builds using NL11F and similar tuning, and make up some graphs on the various heads.

I've been wanting to do that for ages, but it's finding time between builds to get the info properly charted. It's not easy when you're busy

But Hipro, if you'd be willing to help me find out how much and when I'd be able to buy one of Eleven's evaps for testing that would be really great

I've finally put together a nice fully variable power supply and load tester, the PSU can supply up to 1200 watts and I just need to build a 200 to 400w tester for it. I've got a good test block up to 300w, but if I push it too hard it'll just break so I want to make another for higher loads.


Cheers

Gray

[n00b 0f l337]

With Jinu getting his CNC mill up and going, maybe he'll get to test some designs. I've got a few waiting for him that are like elevens, channel high surface area but with a stackable amount. Cascades you could add 2 plates in the center, vs single stage 1 or 0.

[Ssilencer]

With Jinu getting his CNC mill up and going, maybe he'll get to test some designs. I've got a few waiting for him that are like elevens, channel high surface area but with a stackable amount. Cascades you could add 2 plates in the center, vs single stage 1 or 0.

Could you do a little sketch or diagram?


|Gray mole
I think it will be doable in a cnc, in a lathe keeps breaking toolsif you want to go deeper, as we already tried that design

I will see if I can show you the next evap insides that I will be mounting in a unit next week.

[n00b 0f l337]

Sure I'll do one now, but gota do it in autocad or paint since the cameras not working

EDIT: Hmmm autocads not installed on this comp, basically its an eleven evap in spiraling outward design, but on top of the first plate, another that receives flow from the first plate, reradiates out a circle, then out of the evaporator. But milled and such that theres a top peice on this 2nd layer, and that could be swapped for simply another middle peice to expand the evap upward.

[esdee]

Looking at Eleven's evap design closely I can see what you mean.

I've always liked that milled design, and it leaves a lot of room for surface area that it's difficult to produce on a stepper or spiral.

I'd like to see it with a 2-3mm base, and taller fins, that are 1.5mm thin. More of them too.

It's normally all down to the cutters available though, and the quality of the mill. I think that extending the fins, and thinning them out, would produce the best balance of mass and surface area. I've always thought that mass was more important, but as long as there is enough to keep a certain amount of heat storage (or the lack of it ) then the rest is up to a good surface area.

It's part of what makes @itor's evap so special.

I'm starting to build a collection of evaps again, and if I can get ahold of Eleven's (whether new or old rev.) and some others, I can take the data from previous builds using NL11F and similar tuning, and make up some graphs on the various heads.

I've been wanting to do that for ages, but it's finding time between builds to get the info properly charted. It's not easy when you're busy

But Hipro, if you'd be willing to help me find out how much and when I'd be able to buy one of Eleven's evaps for testing that would be really great

I've finally put together a nice fully variable power supply and load tester, the PSU can supply up to 1200 watts and I just need to build a 200 to 400w tester for it. I've got a good test block up to 300w, but if I push it too hard it'll just break so I want to make another for higher loads.


Cheers

Gray

eleven is always in the game, never retired!it's just that he no college student with all that free time (like me ) and he has a family
eleven is a very close friend to unseen and he sells his evaps (40mm base ones) on unseens shop http://www.xtremeshop.com.gr/catalog/index.php


now... i am thinking that i should modify a bit more elevens evap with 3-4mm base thickness and channels that are equal to 3/8" in surface and as much less width possible!, all that after we test the current 50mm mod!

[n00b 0f l337]

Hmmm I dont see the evaps on that link, but I'd definitly like to get a price or in contact with him about any problems with myself designing a new evap (not a clone) but based around a similar milling pattern, probably a full spiral though not a back and forth thing.

[epion2985]

curious to see how chillys new evap will perform on quads.

[kayl]

Ok ppl don’t laugh at me here, it’s a long shot, but I have been playing with some old CHEAP intel D chips (my new 925 does 5.1ghz orthos stable 5475 validation see sign), stock watage their 130w with low vcores. At first I didn’t think much of them load wise, but a cheaper Asus P5b-E Plus mobo I had vcore set to 1.7 droops to 1.44v load orthos, ouch, board only lasted a few session before I smoked an IC. Now that mobo was only phase 3 regulation I think. But I got another mobo now and its the Asus P5b Deluxe with phase 8 power regulation, and again at load with 1.7v bios that droops to 1.5v. Another member has a 820 which is also 130w stock and experiencing similar results. From what I have read some ppl suggest that the newer mobos don’t have the power regelation required to run the older P4 chips at high OC’s and extreme voltages.

So if the newer mobos can handle the quad cores when it comes to power regulation sure that means that they run cooler than the older P4 dual core chips. Whats the stock wattage for a Q6700?
Maybe these cheap cpus in the older mobos are the solution to load testers for the new quad cores.

[s7e9h3n]

Ok ppl don’t laugh at me here, it’s a long shot, but I have been playing with some old CHEAP intel D chips (my new 925 does 5.1ghz orthos stable 5475 validation see sign), stock watage their 130w with low vcores. At first I didn’t think much of them load wise, but a cheaper Asus P5b-E Plus mobo I had vcore set to 1.7 droops to 1.44v load orthos, ouch, board only lasted a few session before I smoked an IC. Now that mobo was only phase 3 regulation I think. But I got another mobo now and its the Asus P5b Deluxe with phase 8 power regulation, and again at load with 1.7v bios that droops to 1.5v. Another member has a 820 which is also 130w stock and experiencing similar results. From what I have read some ppl suggest that the newer mobos don’t have the power regelation required to run the older P4 chips at high OC’s and extreme voltages.

So if the newer mobos can handle the quad cores when it comes to power regulation sure that means that they run cooler than the older P4 dual core chips. Whats the stock wattage for a Q6700?
Maybe these cheap cpus in the older mobos are the solution to load testers for the new quad cores.

Asus boards are notorious for Vdroop @ high loads, that's why I have ALWAYS needed a vdroop mod when using their boards:



Dunno much about Intel, but I'm told that this particular cpu is supposed to be a nice little heater, but it wasn't so bad that it could kill a mobo @ 1.75Vcore loaded. I'm almost positive the newer quads put out more heat than the older cpu's. You may have had problems with compatibility with your cpu and that mobo. Now - once again I'm intel dumb - I'm guessing your first board was a 965 chipset? That chipset was designed for newer cpu's i.e. core2duo and quads....the short possibly could have something to do with running a pentium D cpu in it. They may state that the cpu will work in the board, but we all know that Intel sucks at making motherboards which support different generations of cpu's. Why else would they make a new chipset for every/any revision of their processors?

[epion2985]

I always thought that the power consumed was only one of many things that are a function of the heat output. Doesn't efficiency play in to it?

[kayl]

S79e9hn3 yes im also new to intel but have also heard a lot about the ANUS droop. Yes the other mobo was a 965 but was only 3 phase regulation and was missing a heap of ICS.
I normally only see 1.61v droop at most with E6600 and E6400 chips on my 965 mobo. My old intel D 920 was no near as hot as this cpu, stock I think they only 100w. this 925 is 130w stock and pumps the heat out.
Using the program in the link below seems that a Q6700 at stock is only 130w also, so overclock to say 5ghz or more ya would think it would generate more heat than a Q6700 at 4ghz???

http://www.extreme.outervision.com/psucalculator.jsp

[s7e9h3n]

S79e9hn3 yes im also new to intel but have also heard a lot about the ANUS droop.

LMAO.....Kayl, I think you deserve the award for the worst spelling mistake...EVER!

Getting back on topic - that cpu calculator seems surprisingly accurate for it's estimations of cpu wattage (that's a first)...BUT in terms of what you're trying to calculate, remember, you can't take two cpu's simply with the same default wattage input and expect them to scale similarly with voltage/speed in respect to heat output. Although the 9xx series of procs are built on 65nm tech as is the Q6700, that's about where the similarities end. The Q6xxx has more l2 cache which, generally speaking, adds to the heat output of a particular cpu. The q6xxxx is 4 cores compared to your 2 cores so it definitely should output more heat (as well as draw more power) as Vcore is increased. That being said, I have a pretty hard time believing that the calculations for the heat output on the Q6700 is correct using the calculator. I punched in a few numbers which my good friend OPB has used to achieve his record runs with his q6700 and the power draw is simply staggering and @ a number beyond comprehension:

[kayl]

Hehehe, that was a deliberate spelling mistake, little joke we have over at i4memory.com, but yes kayl cant spell as well, or make sense sometimes after late nights in the shed.
Yes point taken S79e9hn3, 508w hey, looks like ill have to change the load test to a 500w

[johann]

LOL Anus droop That made me laugh

[Drag]

thats single core and why do you need 1.75vcore and 5300mhz.

[epion2985]

why do you need 5300mhz.

Same reason most people here do, for fun, thats why they are here doing all this. Why else did you think people pore liquid nitrogen and dry ice on their possessors, build cascades, chillers and single stage phase change coolers, try to ever perfect their cooling and get colder etc

[boshuter]

Maybe these cheap cpus in the older mobos are the solution to load testers for the new quad cores.

Well you may be right. I never tried to find out how many watts it was putting out, but I have a PD 950 Presler that has overwhelmed 2 of my single stage systems The same single stage was very comfortable with my E6700; I've always told people that the 950 presler was the hottest cpu I've ever owned, but the numbers I get from your calculator (and I believe them) really blow me away with how many watts the presler uses.

These numbers are from my E6700es and 950es;

E6700es
vcore 1.68
cpu speed 4600 (4580 superpi stable)
watts 191

950es
vcore 1.68
cpu speed 5600 (5500 pi stable)
watts 358

That puts the 950 right up there with the quads... and explains my inability to cool one with a single stage that sure wasn't designed for that load.

BTW: 5.6ghz with the 950 was a world record, then "entsafter" took it with his 950 at 5.9ghz I decided to put my 950 on DI to retake the record, I stuck the 950 in an Asus P5B and figured it should have no problem doing better than the older P5WD2 that I always used...... well, the 950 would barely do 5.5ghz on dry ice, when it had no problem doing better than that on the single stage.... the new board clearly isn't good for the older dual cores.