Quad Core Phasing and Information : Read it here!

Quote:

---

Originally Posted by s7e9h3n

TDP = Total amount of heat DISSIPATED by the cpu = It's the maximum amount of heat expelled by a particular cpu in the worst possible conditions, i.e. closed case with no airflow, hot ambient, etc.....

---

Hmm...Why would the power dissipation depend on conditions ? Temperature will definitively, OK....

Quote:

---

Originally Posted by SoddemFX

Hi NoL,

What i meant was that we can simulate, guestimate until the end of time but until someone with a 4.5GHz phase change cooled Quad Core at 1.6v takes the two yellow wires going into the 4-Pin ATX connector - and puts them between a high quality (Fluke etc) clamp ammeter whilst running four instances of Prime95 etc, we're still just guessing :)

And i know someone's going to say "not all of the CPU's power comes from there, blah blah blah" - but it's good enough for me.

So clamp meter, Quad Core - GO! Someone round here must have the two :D

Tom

---

Sorry, and I don't mean to be rude, but you CANNOT measure low DC currents accurately with a clamp on ammeter. I sold them for many years from Fluke and AEMC. They rely on an IC in the clamp to try to offset magnetic influences from other sources, and are not that accurate.

Quote:

---

Originally Posted by harvshark

Sorry, and I don't mean to be rude, but you CANNOT measure low DC currents accurately with a clamp on ammeter. I sold them for many years from Fluke and AEMC. They rely on an IC in the clamp to try to offset magnetic influences from other sources, and are not that accurate.

---

And, upon further reflection, WHERE are you going to intercept the current going ONLY to the processor with a clamp-on or shunt to measure only the current being used by the processor? It's better, I think, to extrapolate the processor power dissipation by emperic means, such as we propose by comparing test temps with processor induced temps.
Anyway, it may all be moot. If we tune for 300w and only dissipate 250w, it would only lower our temps, correct?
And I have no doubt that a quad could generate 300w if pushed to the limits (that's why they call this Extreme Systems, ya)?

Quote:

---

Originally Posted by {.bLanK} GoD

I know there is a way to calculate the capacity of a coil, but I'm dammed if I know how :p:
A wild stab in the dark guesstimation just by looking at the sizes of a Vapo condensor, my guess is they are good for 150-200Watts max.
My home made jobby, (as long as it does actually work properly) my guess would be around the vicinity of 1kw. probably more.

---

photo of it pls!!!

i got some condens but dunno theyr capacity, is there any way of calculating?
i was thinking of using 2 condens with an mp14fb...

gr,
q

Quote:

---

Originally Posted by n00b 0f l337

The concensus is NO, unless you plan to get it modded. Please read the entire thread.

---

WHY? why shoulden't i risk it i don't seem how i could destroy it please explain :slapass:

Drag, please read the thread. The point of the thread is that Vapo's and Mach's (more so vapo's) are dying under the load of overclocked quads. The units cannot take the load.

thanks, in that case all buy a blue orb 2 :D

Quote:

---

Originally Posted by SoddemFX

Hi Harv,

Don't worry, it's not rude at all :)

Would you call ~20A low? :confused:

I thought the newer Hall effect probes were fairly accurate, i could be wrong.

The 4-Pin ATX connector, just Vcore is regulated from this. As far as i know the vast majority of the power drawn by the CPU comes from this...

Unless the user measures superheat as well as evap temperature and also has the same ambient as when the unit was under load sim, it's not a true comparison. I say this because i did the same thing being proposed about a year ago with an Opteron and came up with a whacked power dissipation.

Tom

---

In my experience, clamp on DC ammeters have never been real accurate, regardless of their claimed specs. Breaking the circuit and taking the current through an accurate dmm is the best way to get accurate readings. But I don't know how any readings at the atx power connector would be valid anyway, because I don't know how you could determine how much of the 5v and 12v is used by the cpu, how much is being used by the NB, SB, fans, onboard lan, onboard audio, etc.. Even the vcore voltage regulators will dissipate some current in the form of heat just doing their job, and that would not actually count as power dissipation of the cpu.
I think NoL and I are erring on the high side trying to tune for 300w, but you start kicking the fsb and vcore on a quad, and the little beasties do generate some heat!

Quote:

---

Originally Posted by quintus

photo of it pls!!!

i got some condens but dunno theyr capacity, is there any way of calculating?
i was thinking of using 2 condens with an mp14fb...

gr,
q

---

http://xtremesystems.org/forums/show...=129043&page=2
Post #35

There is a way to calculate capacity using pipe length, diameter, effective fin area, fins per inch. and probably other variables. Dammed if I know though.

Quote:

---

Originally Posted by SoddemFX

Hi NoL,

What i meant was that we can simulate, guestimate until the end of time but until someone with a 4.5GHz phase change cooled Quad Core at 1.6v takes the two yellow wires going into the 4-Pin ATX connector - and puts them between a high quality (Fluke etc) clamp ammeter whilst running four instances of Prime95 etc, we're still just guessing :)

And i know someone's going to say "not all of the CPU's power comes from there, blah blah blah" - but it's good enough for me.

So clamp meter, Quad Core - GO! Someone round here must have the two :D

Tom

---

I got this clamp meter from the local "ebay" at U$S 10 (retail cost in the states is (U$S275)
Prova CM-07 True RMS Clamp Meter specs

http://www.isre.com/images/newproducts/clamp8.jpg

And I have access to a quad core 6600 under water, if you think this clamp is good enough I can try to measure at a determined V.

Quote:

---

Originally Posted by SoddemFX

I appreciate your input, thanks :)

Switchmodes are very efficient, i know you're right about "breaking the circuit" to measure - but it's something which can easilly go wrong which is why i dont realy like the idea, also there's the electrical aspect as the leads would add inductance and voltage drop etc and i don't know what would happen...



Seeing as you have it, it won't hurt to take a measurement as long as we take it "with a pich of salt" due to the concerns raised by Harv and {Blank} God i think it would be usefull. On the written specs it does seem to be very accurate :)

If you measure the current on the two yellow wires under idle and load at high clocks it would be great :toast:

Tom

---

Give me some days and I will be back with the info, the quad is not mine, but from a friend.

I have the QX6700 and I allready ordered a Vapochill LS to cool it, before reading this thread. My bad.
I'm a total noob at phase change cooling and any explanation of 0.36 12ft capilary tubes with xxxxx condensers gets me very confused.
All I want to know is how far will I be able to overclock without damaging the LS? Will I be able to overclock at all? If needed be, who/where can I mod the LS? I live in Romania btw.
If I can cool the compressor's shell, will it be of any use? I'm sure we could use a couple of these for that. A couple of sets that is.
I could also fit some extra fans inside the LS unit. Make a sandwich out of 2 fans and the condenser. Would that help?

Thanks

Don't worry terente, you'll still be able to overclock with the LS.

Under high overclocks and vcore though, you should keep a close watch over it initially though. If you find that it's unable to keep up with the heatload created (and you will find that the temperature of the display will slowly creep upwards) then you may need to back off a little so that the unit can deal with the heatload.

There is a safety system built into the Vapo's control system which you can manage in the software. Setting the 'safety temp' to something like -10 will ensure that your processor get much higher than about 20c or so, and the system would shut down completely. Then you can watch over it later on to see where and how it's shutting down.

The heat being generated by Quad core cpu's under load and high clocks IS high, and higher than than these systems have been set up for. However I've already expressed my belief that it's not nearly so high as 300w, but likely closer to 230w at most, at least for single stages and their ability to allow for overclocks and vcore.

I do recommend running the fans as high as possible, but no other 'mods' that you can do will make a great deal of difference. The highest fans will give the greatest condensor capacity and that's about all you can do.

If you find that it's overall capacity is close to the max heatload that your cpu put's out, you can decide if that overclock is enough for you.

If you do decide to have it modified, there will be modders around you to help, both in Europe and in the UK. I'm in the UK myself, so if you can locate one that's closer it could save you some postage costs if it came to that.

Anyways, running a unit beyond it's capabilities, or right on the edge, can lead to problems with it's lifespan over time, but if you're careful about how you overclock your cpu and ensure that the cpu is generating less than it's capable of cooling then it should work well enough.

Cheers

Gray

Quote:

---

Originally Posted by quintus

photo of it pls!!!

i got some condens but dunno theyr capacity, is there any way of calculating?
i was thinking of using 2 condens with an mp14fb...

gr,
q

---

bump!!! maybe should i post some pics?
thx

Quote:

---

Originally Posted by harvshark

I don't mean to speak for him, I don't think NoL has a quad, but he is working with me on a custom build for my QX6700, using a SC12MLX, with .036 cap and large condensor. Shooting for tuning at 300w with his load tester, which will provide him with a 300w power load to simulate the quad at very high OC. It is a work in progress.

---

Is there a build thread for this unit you all are working on? I'd be interested in seeing it from start to load testing... it would compliment this thread very well :)

Sure I'm waiting on the .036 capillary at the moment and I guess I'll keep a picture log posted somewhere.

Hate to double post, but Boshuter here's some opening pictures for Harvsharks unit. This is just the SC12MLX and the condenser with the 6.5" axial fan attached, Accumulator as well. I should get the .036 by the end of the week from Runmc.
http://img201.imageshack.us/img201/9702/dsc01483fm5.jpg
http://img408.imageshack.us/img408/1383/dsc01484pp3.jpg
http://img201.imageshack.us/img201/8880/dsc01485my0.jpg
http://img244.imageshack.us/img244/9262/dsc01486xp8.jpg

Quote:

---

Originally Posted by {.bLanK} GoD

Have to remember there soddem, clamp meters are not accurate enough to measure current. even a fluke. For an accurate reading you need to wire the ammeter in series with the load. Most clamp meters have a guaranteed accuracy of +/- 30%

---

well get a 30 % clampmeter to be with in 3% is easy... get a length of wire long enough to wrap around the clamp 10 times 1 amp actual will read as 100 amps.

Quote:

---

Originally Posted by harvshark

Sorry, and I don't mean to be rude, but you CANNOT measure low DC currents accurately with a clamp on ammeter. I sold them for many years from Fluke and AEMC. They rely on an IC in the clamp to try to offset magnetic influences from other sources, and are not that accurate.

---

FOr extremely accurate dc ampmeasurments at low currentsuse a millivolt meter and known resistance inseries say a .01 ohmresistor I=E/R

Quote:

---

Originally Posted by quintus

photo of it pls!!!

i got some condens but dunno theyr capacity, is there any way of calculating?
i was thinking of using 2 condens with an mp14fb...

gr,
q

---

Capacity is a function of the mass flow through the block and weither it all boils off in teh block so goal#1 move as much refrigerant through the block as you can and make sure it all boils off using only the heat from the load

Quote:

---

Originally Posted by n00b 0f l337

Hate to double post, but Boshuter here's some opening pictures for Harvsharks unit. This is just the SC12MLX and the condenser with the 6.5" axial fan attached, Accumulator as well. I should get the .036 by the end of the week from Runmc.
http://img201.imageshack.us/img201/9702/dsc01483fm5.jpg
http://img408.imageshack.us/img408/1383/dsc01484pp3.jpg
http://img201.imageshack.us/img201/8880/dsc01485my0.jpg
http://img244.imageshack.us/img244/9262/dsc01486xp8.jpg

---

SC12 MLX use a 0.031 cap tube at 110 inches and set the system to have a 25 deg F suction line 1 inch from the compressor and a little frost on teh compressor with a suction pressure of R507 @ 0 to 2 psi under the load of 140 watts actual.

NOL I dont know why you have this idea of 0.036" stuck in your head. It will cause more problems than anything else.

Basically capillary tubing should be kept int eh 6 foot to 12 foot range to use 0.036 on a 1/3 hp compressor and get the proper evaporating temps you would need more than 12 feet
0.031 or 0.028 is fine... 0.031 if you need more HX space for the larger displacment compressor

Quote:

---

Originally Posted by johann

NOL I dont know why you have this idea of 0.036" stuck in your head. It will cause more problems than anything else.

---

0.036' and ost of flash gas is what i think will happen.
As chilly said 0.031" and 0.028" all we need for these tiny systems. From a builders point of view best to go with 0.028" as its cheaper and at the correct length do the same job as 0.031". i cant get 0.028" thats the main reason i use 0.031" :stick:

0.07mm is about 0.028 in