i9-9900K CPUs are Starting to Ship from Newegg

FrgMstr

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,634
There has been a lot said about Intel's latest Core i9-9900K launch, and you can read our review here, and even yesterday Newegg was telling some of its customers that their pre-orders would be put off a bit, until the 26th. Expected ship date was originally the 19th. Elmy has just informed us that Newegg shipping his brand new 9900K today, and he even got a shipping number. Now to see him over on his Twitch channel getting it installed into his awesome water cooled rig.


After my tweet went Viral. newegg decided to speed up my order. ( J/K ). My @IntelGaming 9900K is on the way. Can't wait to see what this thing can do.
 
From what I understand from Der8auer the i9-9900K is one hot tamale with a pretty low overclocking tolerance even though it has a sTIM. It looks like Intel decided to play it safe with a thicker silicon package to better tolerate the sTIM soldering heat (which would naturally be quite a bit higher than running temps).

He delidded one for shits & giggles and found a really thick core and that if he lapped a few mm off the thickness of the core package he was able to get far better cooling results. So even though non-unlidded the 9900K runs cooler than it's earlier generation counterpart (not unlidded as well), I guess Intel felt that a slightly hotter chip is a good tradeoff to using sTIM and having less incentive for people delidding their CPUs to get them cooler.

Personally I would not be too keen on grinding down the core silicon...
 
Customers Also Bought.png
 
Actually, Der8auer was a lot more adamant about the bad temps than that. He said he was very surprised at just how BAD the temps were.

Just removing the solder helped a lot. Which he really couldn't understand.

If this is going to be the case, then unfortunately Intel tried to give enthusiasts what they wanted... and made the problem worse because now the average user can't delid without extreme risk and even after that it would have to be cleaned off to apply something else.

Honestly, this is getting kind of stupid. If soldering the chips extremely well is just not justifiable by Intel anymore from a technical standpoint then how about they put the liquid metal tim on and design the heat spreader and cpu top to be resistant to it migrating. That would make everyone happy right?
 
So somebody who has a twitch channel gets his order expedited. How about regular joes, they still have to wait?
 
Teenk9 said:
So somebody who has a twitch channel gets his order expedited. How about regular joes, they still have to wait?
Click to expand...

Kyle_B swings the big D around and things happen. Don't hate him for being beautiful :p
 
good to see these shipping...bad that they raised the price of the 8700K
 
Well it's Newegg, so your item will either arrive at the speed of light or will circle Missouri in the back of a truck for 2 weeks. I shop with them a lot but their shipping inconsistency is frustrating.
 
Soldering the heat spreader is only part of the overall thermal equation.

Note the thermal conductivity differences in this table:

https://electronics-cooling.com/wp-content/uploads/2006/08/2006_August_TechData_Table1.jpg

The conductivity ranges from 17 to 78 w/mk.

Watts per meter kelvin means applying 1 watt across a 1 meter thick layer raises the temp on the heated side 1° k, if the other side is held constant.

So, 17w/mK sux 4.58 times worse than 78w/mK.

At 100W, and a layer 1mm thick, it's either a 0.0058° e or a 0.0012° rise. (Power/w/mK value/thickness) 100/17/1000=0.0058



Armed with the power value from a monitoring program, the die temp, the temp of the cooler base, the radiator, and the ambient temp, you can calculate the overall thermal conductivity of your thermal solution.

Example:
My 3930k running this browser is using about 72W
My hottest core is 55° C (C and K are the same size degrees, C just has an offset from absolute zero.)
My package reads 55°C too, so that's not useful.

The base of the cooler, a Capitan 360 water loop, is 33°C. Die to base difference=55-33=22°C
The Radiator is 30°C. Die to Radiator is 25°C
Room is 26°C. Die to room is 29°C.

So current conductivity is: (Low Power)
22°C/72W=0.305 (Temp rise/power to the base °C/W)
25°C72=0.34 Die to Radiator
29/72=0.402 Die to Ambient

If we now go to higher power, like running euler3d, we see:

My 3930k running this benchmark is using about 153.29W
My hottest core is 75° C (C and K are the same size degrees, C just has an offset from absolute zero.)
My package reads 75°C too, so that's not useful. Again. (This means the IHS is working properly)

The base of the cooler, a Capitan 360 water loop, is now 33.7°C. Die to base difference=75-33.7=41.3°C
The Radiator is 30.7°C. Die to Radiator is 44.3°C
Room is 26°C. Die to room is 49°C.

Now, conductivity is: (High Power)
41.3/153.29=0.269 to the base
44.3/153.29=0.288 Die to Radiator
49/153.29=0.319 Die to Ambient

Note how these numbers have not changed a great deal; the difference is the increased rate of thermal transfer, made more efficient by the increased thermal difference.
(The higher the temperature difference, the higher the thermal flux, which makes it flow faster.)
If you establish a high flux, and end the run, Watch the die temp drop Below Ambient, as the flux continues to run for a bit, once established.

(I had to show this to a physicist once, and he still didn't believe me. I made a product out of it, lol.)

A heatsink in boiling water will never go much above 100°C as long as it's covered in water, so the heat exchange rate is most efficient for water at its boiling point. :)
As is Freon, Ammonia, and every other liquid.
The "Target Working Temperature" is the big difference. :)

I did NOT heat soak the cooler solution, as this is just showing how to measure it; allowing my system to achieve thermal equilibrium would give better results.


If you see a large difference in the numbers you get, (I see it go from .402 to .319, chip to air, going from 72W to 153W) then you are going across a boundary, either local boiling in the cooler, or something is changing.
The number should all go down with increasing power, unless you are undercooled. :)

When you reach the limit of your cooling solution, you need more power!!
(A 2000W ice chiller would be nice, lol.)

The bottom number is w/mK/thickness, and is a sum of ALL thermal 'impedance'; it's hard to unfold ALL the variables, but a lump sum works for this.

Thermal 'barriers add like resistors, alternate thermal paths are like parallel resistors. :)
Cooling the bottom of the die is next, lol. Just wait...


I'd love for someone to post the Raw temperature results as above on one of the 9990k systems; we can see where the difference is, and how it changes with temperature.

Sorry if this is obvious, but I've never seen this calculation published; this would eliminate the crappy coolers from the market Real Quick. :)
 
DrBorg said:
Soldering the heat spreader is only part of the overall thermal equation.

Note the thermal conductivity differences in this table:

https://electronics-cooling.com/wp-content/uploads/2006/08/2006_August_TechData_Table1.jpg

The conductivity ranges from 17 to 78 w/mk.

Watts per meter kelvin means applying 1 watt across a 1 meter thick layer raises the temp on the heated side 1° k, if the other side is held constant.

So, 17w/mK sux 4.58 times worse than 78w/mK.

At 100W, and a layer 1mm thick, it's either a 0.0058° e or a 0.0012° rise. (Power/w/mK value/thickness) 100/17/1000=0.0058



Armed with the power value from a monitoring program, the die temp, the temp of the cooler base, the radiator, and the ambient temp, you can calculate the overall thermal conductivity of your thermal solution.

Example:
My 3930k running this browser is using about 72W
My hottest core is 55° C (C and K are the same size degrees, C just has an offset from absolute zero.)
My package reads 55°C too, so that's not useful.

The base of the cooler, a Capitan 360 water loop, is 33°C. Die to base difference=55-33=22°C
The Radiator is 30°C. Die to Radiator is 25°C
Room is 26°C. Die to room is 29°C.

So current conductivity is: (Low Power)
22°C/72W=0.305 (Temp rise/power to the base °C/W)
25°C72=0.34 Die to Radiator
29/72=0.402 Die to Ambient

If we now go to higher power, like running euler3d, we see:

My 3930k running this benchmark is using about 153.29W
My hottest core is 75° C (C and K are the same size degrees, C just has an offset from absolute zero.)
My package reads 75°C too, so that's not useful. Again. (This means the IHS is working properly)

The base of the cooler, a Capitan 360 water loop, is now 33.7°C. Die to base difference=75-33.7=41.3°C
The Radiator is 30.7°C. Die to Radiator is 44.3°C
Room is 26°C. Die to room is 49°C.

Now, conductivity is: (High Power)
41.3/153.29=0.269 to the base
44.3/153.29=0.288 Die to Radiator
49/153.29=0.319 Die to Ambient

Note how these numbers have not changed a great deal; the difference is the increased rate of thermal transfer, made more efficient by the increased thermal difference.
(The higher the temperature difference, the higher the thermal flux, which makes it flow faster.)
If you establish a high flux, and end the run, Watch the die temp drop Below Ambient, as the flux continues to run for a bit, once established.

(I had to show this to a physicist once, and he still didn't believe me. I made a product out of it, lol.)

A heatsink in boiling water will never go much above 100°C as long as it's covered in water, so the heat exchange rate is most efficient for water at its boiling point. :)
As is Freon, Ammonia, and every other liquid.
The "Target Working Temperature" is the big difference. :)

I did NOT heat soak the cooler solution, as this is just showing how to measure it; allowing my system to achieve thermal equilibrium would give better results.


If you see a large difference in the numbers you get, (I see it go from .402 to .319, chip to air, going from 72W to 153W) then you are going across a boundary, either local boiling in the cooler, or something is changing.
The number should all go down with increasing power, unless you are undercooled. :)

When you reach the limit of your cooling solution, you need more power!!
(A 2000W ice chiller would be nice, lol.)

The bottom number is w/mK/thickness, and is a sum of ALL thermal 'impedance'; it's hard to unfold ALL the variables, but a lump sum works for this.

Thermal 'barriers add like resistors, alternate thermal paths are like parallel resistors. :)
Cooling the bottom of the die is next, lol. Just wait...


I'd love for someone to post the Raw temperature results as above on one of the 9990k systems; we can see where the difference is, and how it changes with temperature.

Sorry if this is obvious, but I've never seen this calculation published; this would eliminate the crappy coolers from the market Real Quick. :)
Click to expand...
Bro, can we get a TL;DR on this?
 
DrBorg said:
Soldering the heat spreader is only part of the overall thermal equation.

Note the thermal conductivity differences in this table:

https://electronics-cooling.com/wp-content/uploads/2006/08/2006_August_TechData_Table1.jpg

The conductivity ranges from 17 to 78 w/mk.

Watts per meter kelvin means applying 1 watt across a 1 meter thick layer raises the temp on the heated side 1° k, if the other side is held constant.

So, 17w/mK sux 4.58 times worse than 78w/mK.

At 100W, and a layer 1mm thick, it's either a 0.0058° e or a 0.0012° rise. (Power/w/mK value/thickness) 100/17/1000=0.0058



Armed with the power value from a monitoring program, the die temp, the temp of the cooler base, the radiator, and the ambient temp, you can calculate the overall thermal conductivity of your thermal solution.

Example:
My 3930k running this browser is using about 72W
My hottest core is 55° C (C and K are the same size degrees, C just has an offset from absolute zero.)
My package reads 55°C too, so that's not useful.

The base of the cooler, a Capitan 360 water loop, is 33°C. Die to base difference=55-33=22°C
The Radiator is 30°C. Die to Radiator is 25°C
Room is 26°C. Die to room is 29°C.

So current conductivity is: (Low Power)
22°C/72W=0.305 (Temp rise/power to the base °C/W)
25°C72=0.34 Die to Radiator
29/72=0.402 Die to Ambient

If we now go to higher power, like running euler3d, we see:

My 3930k running this benchmark is using about 153.29W
My hottest core is 75° C (C and K are the same size degrees, C just has an offset from absolute zero.)
My package reads 75°C too, so that's not useful. Again. (This means the IHS is working properly)

The base of the cooler, a Capitan 360 water loop, is now 33.7°C. Die to base difference=75-33.7=41.3°C
The Radiator is 30.7°C. Die to Radiator is 44.3°C
Room is 26°C. Die to room is 49°C.

Now, conductivity is: (High Power)
41.3/153.29=0.269 to the base
44.3/153.29=0.288 Die to Radiator
49/153.29=0.319 Die to Ambient

Note how these numbers have not changed a great deal; the difference is the increased rate of thermal transfer, made more efficient by the increased thermal difference.
(The higher the temperature difference, the higher the thermal flux, which makes it flow faster.)
If you establish a high flux, and end the run, Watch the die temp drop Below Ambient, as the flux continues to run for a bit, once established.

(I had to show this to a physicist once, and he still didn't believe me. I made a product out of it, lol.)

A heatsink in boiling water will never go much above 100°C as long as it's covered in water, so the heat exchange rate is most efficient for water at its boiling point. :)
As is Freon, Ammonia, and every other liquid.
The "Target Working Temperature" is the big difference. :)

I did NOT heat soak the cooler solution, as this is just showing how to measure it; allowing my system to achieve thermal equilibrium would give better results.


If you see a large difference in the numbers you get, (I see it go from .402 to .319, chip to air, going from 72W to 153W) then you are going across a boundary, either local boiling in the cooler, or something is changing.
The number should all go down with increasing power, unless you are undercooled. :)

When you reach the limit of your cooling solution, you need more power!!
(A 2000W ice chiller would be nice, lol.)

The bottom number is w/mK/thickness, and is a sum of ALL thermal 'impedance'; it's hard to unfold ALL the variables, but a lump sum works for this.

Thermal 'barriers add like resistors, alternate thermal paths are like parallel resistors. :)
Cooling the bottom of the die is next, lol. Just wait...


I'd love for someone to post the Raw temperature results as above on one of the 9990k systems; we can see where the difference is, and how it changes with temperature.

Sorry if this is obvious, but I've never seen this calculation published; this would eliminate the crappy coolers from the market Real Quick. :)
Click to expand...


Boiling water doesn't exceed 100°C without being under pressure, no matter how much heat is applied.
Freon and other refrigerants are gases, not liquids.
 
cyberguyz said:
From what I understand from Der8auer the i9-9900K is one hot tamale with a pretty low overclocking tolerance even though it has a sTIM. It looks like Intel decided to play it safe with a thicker silicon package to better tolerate the sTIM soldering heat (which would naturally be quite a bit higher than running temps).

He delidded one for shits & giggles and found a really thick core and that if he lapped a few mm off the thickness of the core package he was able to get far better cooling results. So even though non-unlidded the 9900K runs cooler than it's earlier generation counterpart (not unlidded as well), I guess Intel felt that a slightly hotter chip is a good tradeoff to using sTIM and having less incentive for people delidding their CPUs to get them cooler.

Personally I would not be too keen on grinding down the core silicon...
Click to expand...

It's likely thicker to deal with flexure and dissipation under extreme thermal cycling. E.g. video editing batches. Remember all the doomsday team IDF saying how AMD soldering would make microfracturing, based on some intel study where that happened on relatively thin, sandy bridge or whatever era chips (one I'm running).
People think intel has continuously improved their process. I guess they forgot 5GHz 2600ks in 2011? Intel has just bumped the remaining GHz of clock up and done a few process improvements and uarch tweaks. Nothing much more. Just like when AMD stuff doesn't clock much, now Intel is there too. They're using 100% of what their process allows now.
 
N4CR said:
It's likely thicker to deal with flexure and dissipation under extreme thermal cycling. E.g. video editing batches. Remember all the doomsday team IDF saying how AMD soldering would make microfracturing, based on some intel study where that happened on relatively thin, sandy bridge or whatever era chips (one I'm running).
People think intel has continuously improved their process. I guess they forgot 5GHz 2600ks in 2011? Intel has just bumped the remaining GHz of clock up and done a few process improvements and uarch tweaks. Nothing much more. Just like when AMD stuff doesn't clock much, now Intel is there too. They're using 100% of what their process allows now.
Click to expand...

5GHz 2600k CPUs in 2018, too!

:)
 
DrBorg said:
Soldering the heat spreader is only part of the overall thermal equation.

Note the thermal conductivity differences in this table:

https://electronics-cooling.com/wp-content/uploads/2006/08/2006_August_TechData_Table1.jpg

The conductivity ranges from 17 to 78 w/mk.

Watts per meter kelvin means applying 1 watt across a 1 meter thick layer raises the temp on the heated side 1° k, if the other side is held constant.

So, 17w/mK sux 4.58 times worse than 78w/mK.

At 100W, and a layer 1mm thick, it's either a 0.0058° e or a 0.0012° rise. (Power/w/mK value/thickness) 100/17/1000=0.0058



Armed with the power value from a monitoring program, the die temp, the temp of the cooler base, the radiator, and the ambient temp, you can calculate the overall thermal conductivity of your thermal solution.

Example:
My 3930k running this browser is using about 72W
My hottest core is 55° C (C and K are the same size degrees, C just has an offset from absolute zero.)
My package reads 55°C too, so that's not useful.

The base of the cooler, a Capitan 360 water loop, is 33°C. Die to base difference=55-33=22°C
The Radiator is 30°C. Die to Radiator is 25°C
Room is 26°C. Die to room is 29°C.

So current conductivity is: (Low Power)
22°C/72W=0.305 (Temp rise/power to the base °C/W)
25°C72=0.34 Die to Radiator
29/72=0.402 Die to Ambient

If we now go to higher power, like running euler3d, we see:

My 3930k running this benchmark is using about 153.29W
My hottest core is 75° C (C and K are the same size degrees, C just has an offset from absolute zero.)
My package reads 75°C too, so that's not useful. Again. (This means the IHS is working properly)

The base of the cooler, a Capitan 360 water loop, is now 33.7°C. Die to base difference=75-33.7=41.3°C
The Radiator is 30.7°C. Die to Radiator is 44.3°C
Room is 26°C. Die to room is 49°C.

Now, conductivity is: (High Power)
41.3/153.29=0.269 to the base
44.3/153.29=0.288 Die to Radiator
49/153.29=0.319 Die to Ambient

Note how these numbers have not changed a great deal; the difference is the increased rate of thermal transfer, made more efficient by the increased thermal difference.
(The higher the temperature difference, the higher the thermal flux, which makes it flow faster.)
If you establish a high flux, and end the run, Watch the die temp drop Below Ambient, as the flux continues to run for a bit, once established.

(I had to show this to a physicist once, and he still didn't believe me. I made a product out of it, lol.)

A heatsink in boiling water will never go much above 100°C as long as it's covered in water, so the heat exchange rate is most efficient for water at its boiling point. :)
As is Freon, Ammonia, and every other liquid.
The "Target Working Temperature" is the big difference. :)

I did NOT heat soak the cooler solution, as this is just showing how to measure it; allowing my system to achieve thermal equilibrium would give better results.


If you see a large difference in the numbers you get, (I see it go from .402 to .319, chip to air, going from 72W to 153W) then you are going across a boundary, either local boiling in the cooler, or something is changing.
The number should all go down with increasing power, unless you are undercooled. :)

When you reach the limit of your cooling solution, you need more power!!
(A 2000W ice chiller would be nice, lol.)

The bottom number is w/mK/thickness, and is a sum of ALL thermal 'impedance'; it's hard to unfold ALL the variables, but a lump sum works for this.

Thermal 'barriers add like resistors, alternate thermal paths are like parallel resistors. :)
Cooling the bottom of the die is next, lol. Just wait...


I'd love for someone to post the Raw temperature results as above on one of the 9990k systems; we can see where the difference is, and how it changes with temperature.

Sorry if this is obvious, but I've never seen this calculation published; this would eliminate the crappy coolers from the market Real Quick. :)
Click to expand...

You are now my favourite poster on this site. Please 'spam' it as much as you can with your excellent knowledge and experience, thanks.
The 'flux flow' is highly interesting indeed. Almost like it has inertia..

Cooling the bottom of the die is next, lol. Just wait...
I'm pretty sure IBM was working on this a few years ago with some weird water spray system, but it may have just been the top. Very few CPU packages on older CPUs are even mounted on the bottom too, when heat wasn't so much of an issue of course.



Slash3 said:
5GHz 2600k CPUs in 2018, too!

:)
Click to expand...

Wow, I have never been lucky to get one that clocked well, or maybe I just sucked at tweaking them after mastering the A64 haha.
Has it degraded at all? What voltage? You're doing this on air too which is even more impressive.. Nice work!
 
seems like a very nice processor but at this time, if I was to build/upgrade something I'd be more interested in a ryzen 2700/2600.

N4CR said:
Has it degraded at all? What voltage? You're doing this on air too which is even more impressive..
Click to expand...
It wasn't a rare thing to hit 5.0ghz with those 2600k. One of my friend still has one that used to do 5.2ghz, now it can't even do stock clock speed without increased voltage and its getting replaced by a ryzen soon.
I had a 2500k that would do 5.0ghz too.
 
Last edited:
alxlwson said:
Boiling water doesn't exceed 100°C without being under pressure, no matter how much heat is applied.
Freon and other refrigerants are gases, not liquids.
Click to expand...

I laughed SO hard at this. (y)

I'm going to give you a hint: pv=nrT is for gasses.

Freon and other refrigerants are gases AT STP.

Water is a liquid at STP.Something about "Vapor Pressure".

Boiling water is at the Triple Point. Most refrigerants are used at their triple point.

Can you see where I'm going with this? :)

As opposed to most of my fellow men, I actually use and understand all those boring classes I sat thru. I do occasionally screw up, tho.

I was measuring a new set of pistons last night, and they were all too small by a really small amount. Then I realized, they were really cold from being outside; I gave it 20 minutes, and they were perfect. (Trust but verify)

I'd have never believed years back that 0.0005" would make a difference (Or that I would own tools that measure 0.0001"), but Mod Motors are picky; it only has to stick once.

:)
 
Last edited:
DrBorg said:
I laughed SO hard at this. (y)

I'm going to give you a hint: pv=nrT is for gasses.

Freon and other refrigerants are gases AT STP.

Water is a liquid at STP.Something about "Vapor Pressure".

Boiling water is at the Triple Point. Most refrigerants are used at their triple point.

Can you see where I'm going with this? :)

As opposed to most of my fellow men, I actually use and understand all those boring classes I sat thru. I do occasionally screw up, tho.

I was measuring a new set of pistons last night, and they were all too small by a really small amount. Then I realized, they were really cold from being outside; I gave it 20 minutes, and they were perfect. (Trust but verify)

I'd have never believed years back that 0.0005" would make a difference (Or that I would own tools that measure 0.0001"), but Mod Motors are picky; it only has to stick once.

:)
Click to expand...

I'm talking about sea level pressures, PSIA.

I work with tools that measure down to .00001gon, .001mV, .01mA, .000001". See, I have cool tools as well.
 
STP means "Standard Temperature and Pressure."

It's a standard way of measuring gasses.

(PSIA is 'gage pressure', and has a way more limited usage. Like Gauges. :) )


A 'working fluid' is used near it's triple point; the compressor in an air conditioner compresses the freon into a fluid, so that it can efficiently absorb heat; it then passes thru an orifice, where it transitions to a gas, and cools off.

Freon works well at up to about 120F, or a 50 degree temperature difference.

Ammonia can freeze Freon, Freon can freeze water.

LN2 can make LO2, which is fun to play with, and magnetic, lol. (You can pull LO2 out of LN2 with a magnet. No shit. :) just blow air into the ln2 )

If it gets too hot, it doesn't work, because the compressor cant make it a fluid again.

Ammonia works over a wider range of temperatures, but it works much better at lower than freon temperatures, which is why industrial freezers use Ammonia, if they can.

Water is pretty limited in range, as it freezes at 0C, which is where we'd Really like to run our processors for that 5+GHz OC.

:)
 
PSIA is not a guage pressure. PSIG is guage pressure. PSIA is actual pressure, which includes variances due to elevation, humidity, temperature, etc... Notice the G and A

I thought you said you paid attention in those boring classes? I know I paid attention in my fluid power classes.

Also, Freon is a brand, and you won't find it hardly used in anything any longer. The proper term is refrigerant, and there are hundreds of them.
Your understanding of how evaporative cooling works appears to be quite limited, based on your inaccurate description.
 
DrBorg said:
I laughed SO hard at this. (y)

I'm going to give you a hint: pv=nrT is for gasses.

Freon and other refrigerants are gases AT STP.

Water is a liquid at STP.Something about "Vapor Pressure".

Boiling water is at the Triple Point. Most refrigerants are used at their triple point.

Can you see where I'm going with this? :)

As opposed to most of my fellow men, I actually use and understand all those boring classes I sat thru. I do occasionally screw up, tho.

I was measuring a new set of pistons last night, and they were all too small by a really small amount. Then I realized, they were really cold from being outside; I gave it 20 minutes, and they were perfect. (Trust but verify)

I'd have never believed years back that 0.0005" would make a difference (Or that I would own tools that measure 0.0001"), but Mod Motors are picky; it only has to stick once.

:)
Click to expand...

Some of your points are correct but a lot aren't, at least in terminology.

But I'd start with looking up triple point. That's where the solid, liquid, and gas phase diagrams intersect. Fun for experiments but definitely not being used in heat pipes or any standard refrigerant systems. You're sitting along the liquid/gas intersection as the heat of vaporization is really high, ergo a huge amount of heat capacity and therefore heat transfer with minimal mass transfer. Still ultimately gotta get the heat out.

Water under pressure stays liquid at 0 C. Takes a lot of pressure, though. http://www1.lsbu.ac.uk/water/water_phase_diagram.html

-Someone ELSE who paid attention to his thermo classes.
 
PSIA and PSIG are what you're going to read on a gauge; STP is what you do calculations with, as you have to have a reference point.

None of those take into account humidity or other things, like different gas mixes. Then you go to vapor pressure.

I was being funny, but not everyone get my sense of humor. :)


I have ran a heat pipe cooler to the triple point; it works great, but it won't go below 0°C very much. :) And it doesn't move any heat until it warms back up.

But it keeps a cpu from going to -80°, which is useful. :)

You can insulate the heat pipe fins from the mobo much easier than submerging the whole thing in mineral oil, for one.

Do that on a water loop, and you get really bad results. :D

The triple point is Not where you want to run, and You were the one that noticed. (y)
 
The whole heat thing with the 9900K reminds me of some of Intel's earlier enthusiast CPUs. Can't remember which one it was, but I remember Kyle reviewing a CPU that ran so hot that even under his old water cooling set up the beast was hitting 90+c. The heat memes from that time were pretty good.
 
DrBorg said:
PSIA and PSIG are what you're going to read on a gauge; STP is what you do calculations with, as you have to have a reference point.

None of those take into account humidity or other things, like different gas mixes. Then you go to vapor pressure.

I was being funny, but not everyone get my sense of humor. :)


I have ran a heat pipe cooler to the triple point; it works great, but it won't go below 0°C very much. :) And it doesn't move any heat until it warms back up.

But it keeps a cpu from going to -80°, which is useful. :)

You can insulate the heat pipe fins from the mobo much easier than submerging the whole thing in mineral oil, for one.

Do that on a water loop, and you get really bad results. :D

The triple point is Not where you want to run, and You were the one that noticed. (y)
Click to expand...


Crazy! I guess I wasn't doing calculation when I was using PSIA calculations calculated off of PSIG and barometric conditions to determine minimum suction necessary for various hydraulic pump/tank configurations and suction tube lengths. I guess it was all guess work and I was playing with a calculator for no reason at all!! Thanks for the edumacashun
 
Those aren't gasses, and on the suction side, depending on how tall the lines are, vapor pressure comes into play.

Ever wonder what it's hard to pump water over ~32 feet high? Hydraulic fluid does the same thing.

:rofl:
 
DrBorg said:
The triple point is Not where you want to run, and You were the one that noticed. (y)
Click to expand...

To my defense, it's hard to tell when someone's being funny by saying a plausibly true but not really statement and when they don't actually know better. Anyhow, cheers.
 
Cheers!

I just want to see Data.

I am probably going to be buying a new system, but I already have a Central Air unit with Gas heating; assing another one for the processor isn't in my plans.

I have some systems with 1000W power supplies, but the video cards are drawing more than the CPU in all cases. :)

I can't imagine a CPU that draws twice that, and has maybe a 50% faster performance level.

My 3930k has never drawn more than 250W, and that was at 5.2GHz On All Cores, lol.

If I'd added the cooler Intel demoed their 28 core system on, I could still be running that, lol.

It's now what, 7 years old? My last upgrade improved performance, and lowered power by replacing 2x 7970's with one RX480, and it actually draws less than half the power, even overclocked, and has more memory.

I think Intel has hit the wall, and AMD 32 cores Still Overclock well.

I wanna see what the 9900 draws at 5GHz on all cores. :)
 
What the hell was wrong with bare dies? Just put a shim around the core really close to it to prevent chipping and be done with it. If someone cracks the core, tough titties. Not that I have ever seen that happen.
 
aokman said:
What the hell was wrong with bare dies? Just put a shim around the core really close to it to prevent chipping and be done with it. If someone cracks the core, tough titties.
Click to expand...
They'd be screaming bloody murder demanding an RMA for Intel's "bad design".
 
I don't remember anyone bitching about breaking their Barton Core processors, and I know several people here that chipped them.

It's part of being [H]ard.

To steal and distort an old T-shirt meme, "If I can't overclock it, it ain't worth a F***. " (Stiff Records, BTW.)
 
Anyone want to see some 8700K temps at 5.1ghz? That's delidded with a refurbished Corsair h115i water cooling setup that I got off Newegg for something silly like $50 with an Ebay coupon. I have four 1200rpm 140mm fans running in a push/pull configuration drawing in cool air from the front of my case. The fans are whisper quiet. It's hard to believe just how badly Intel messed up the thermals on the new chips.
8700k 5.1ghz temp.PNG
 
N4CR said:
Wow, I have never been lucky to get one that clocked well, or maybe I just sucked at tweaking them after mastering the A64 haha.
Has it degraded at all? What voltage? You're doing this on air too which is even more impressive.. Nice work!
Click to expand...

Yeah, I certainly can't complain. Bought it at launch, paired it with an Asus P8P67 Deluxe, the bigass Noctua and some nice G.Skill RAM and went to town. Most I got out of that setup was about 4.8GHz 24/7 stable. It wasn't until I switched to the Asrock Z77 Formula OC board that I was able to both reach 5GHz @ 1.40v 24/7 stable as well as kick my 4x8GB 1866MHz RAM up to 2133MHz - the Asus didn't like running four sticks. I haven't run into any issues with degradation, luckily. It's been obnoxiously solid. It replaced a Q9550 that I had been running under phase change on a DFI x38 LanParty board at about 4.4GHz, which was no slouch itself, even in 2011 when I switched. That said, the 2600k curb stomped it. Haaarrrd. Man, memories.
 
  • Like
Reactions: N4CR
like this
You must log in or register to reply here.
Back
Top