Silver 7700K IHS Prototype Testing

My first silver IHS (Intel heat spreader) has come . This one is the .925 solid silver variant with a cheaper finish. It is also UK Hallmarked (registered) so it is classed as an equity for an easy resale for the future. Tests will be coming soon...

There is another one being made elsewhere in a softer .999 (Hallmarked) solid silver for an easy squash against the heatsink and die (hopefully if it's that soft). This one will have a better finish, but it is more expensive to be made.

Some may call this stupidity! but someone had to do it
I know the bottleneck of the temps is the weld between the IHS and the heatsink. But I'm going to give it a shot!
These silver IHS prototypes cost me a kidney to be made. But if I ever need another, they will be a lot cheaper to buy as I have already been through the moulding, spruing, Hallmarking, registering, etc, etc setup.

I will like to apologize for the photo quality. It was done with my Galaxy S5 with a cracked camera lens... I don't have a camera, sorry.

The design has been taking from the BitsPower Kaby Lake IHS lid (left). The BitsPower lid looks a bit off due to the casting procedure.



The opposite side of the lids. Both sides of the BitsPower lid has been lapped before the silicone moulding procedure. The scruffy colour is from the torture of the moulding work. The next coming soon .999 silver IHS will have the 2 holes covered.The rough side with a less cheaper finish as it is not needed on this area. Take note: the small stamp, is the Hallmark .925 stamp.



Dark room photo with the USB flash stick as a reference. Shines bright like a diamond

The .999 silver version will be coming soon. It will be much more superior. See you all after the first .925 prep work

 

That's one lit LID you have there.

 

It's been quite a while since I saw the first silver lugged heatsinks ^-^ Never took off due to the expense but i'm interested to see your results.

 

Hedonismbot would approve. Lol. Let us know what comes of the testing.

 

A 2c difference on max CPU load. It's now 80c max compared to my previous max 82c. On "full" fans the CPU minimum temp is now 39c compared to the previous 40c after a full stress test. So the silver lid made a 1c difference on minimum temps. Please note that this is after when the system has been heated up and then brought back down again. Using 1+ hour of Prime95 and 1 hour of gaming on Battlefield: Hardline and other games on 7680x1440 Nvidia Surround.
Please see HWiNFO64 for the reference of 4.5+ hours uptime.

Using prime 95:
First I tried "full" fans on Clevo dash with U3+ (x3) Varder fan mod on full. This got me 74c on the CPU.
Then I used just "full" fans on the Clevo dash without the U3+ (x3) Varder fan mod. This got me to 77c tops.
I then finally used "Overclock" fan settings on the Clevo dash with no fan mod, and I got a maximum CPU temp of 80c.
I have tried everything to put the CPU to a full load. It seems to only go to 90.044W max.

BIOS: Stock Clevo Bios
Ambient temperature: 22c-24c through out the full day.
CPU Max Temp (4.5Ghz 7700K): 80c
Max CPU Power Package: 90.044 Watts
CPU Under-Volt: -145mv
GPU (1080 Stock) Master Max Temp: 66C
GPU (1080 Stock) Slave Max Temp: 66C
Samsung 960 M.2 SSD Max Temps: #1: 48c, #2: 60c
NVidia Driver: 385.41
3x2560x1440P (7680x1440P) + 1x1920x1080P displays.

The screen shot below has been finished of with "full" fans on Clevo Dash to see the lowest it will hit.


I'm now waiting on a .999 silver lid to finish. Hopefully that will show a better result. Right now its only a 2c difference. I'm going to try everyday to see if I can break the 80c CPU temp barrier through gaming or tests.

 

Interesting experiment, how much does it cost?

 

The answer can be seen below

 

Lol I hope my partner doest find out... I'm casting her something. So if she finds out, I will pull the gift out of my secret stash to make her calm! This will drown her thoughts about the silver IHS lids

Both experiments the .999 silver and .925 silver lids first initial setups (there was a few) cost me a bomb. I can't remember exctly, but it all depends on the finishing, Hallmarking etc. But I do know that; if I repeat the order for the .925 silver lid using the same initial setup it will be much cheaper, it will cost me only £65 each then, as long as I don't change the setup. This all depends on the silver prices at that time too. The .999 silver lid will cost me a bit extra though, I'm still waiting for the first prototype. It is the finishing process that costs the most. Hopfully after the first .999 lid test, I will sell a few here (without profiting) to curious individuals. And who knows
... Maybe I will make other intel lids for kicks. I'm starting to like this hobby

By the way: hopfully the .999 silver lid will be so soft, it will warp or wrap it's self to the heatsink, making it have better contact.

Does anyone know if the P870km1 CPU heatsink is a vapour chamber or not. Because if it isn't.... well I got ideas ​

 

Please check out my prep work on a Clevo P870km1 with more photos for this experiment. Some where on this thread (I think on this page). I don't know how to specifically link it to the start of the first post, sorry.

http://forum.notebookreview.com/thr...876-owners-lounge-phoenix-3-0.800081/page-229

 

I think many of us can be interested and can pay for any alloys or metallic mixture allow us to enhance the heat problems on all DTR's even for the heatsinks as well, the cost is an important factor but it's not the most. If you succeed to give us a solution makes a real difference, I'll be your first buyer. Good luck.

 

Thank you. I would just like to inform this forum that gold has a lower thermal conductivity than silver. Diamond and graphine have a lot more thermal conductivity than silver, but is more expensive. And it's hard for them two materials to conduct heat on all directions. It depends on how the atoms are placed in them materials. And when the atoms have been placed probably in line, heat can only travel in one direction. So silver is the best thermal and electric conductor ever. This includes best light reflector too.

The .925 silver lid that I have is 92.5% silver and 7.5% copper. The .999 silver lid that I have in the making is 99.9% silver. I think this is the best metal for controlled thermal conductivity for now.

I will stay in touch with this forum. good day.

 

I'm interested, let me know how the .999 works

Sent from my XT1585 using Tapatalk

 

How about an 1 piece IHS / Heatsink ?
The big problem as I see it is making the contact between the two.

 

That's what I'm thinking, but my head is telling me that I need to narrow the the heatsink connection on where it connects to the naked (no lid) CPU die, due to the LGA socket clamp needed to be clamped down to hold the cpu in place. This way, it will be one single connection. I'm still doing my research on it. Just the lid is confusing enough because of the casting shrinkage. I have already lost to moulding setups on the .925 version. Hopfully this .999 will end all this lid hassle. And thats when I will be thinking on going to heatsinks. I will keep you informed pal. Thanks for your thought

 

You run into socket issues and good pin contact going direct die.

 

Thank you Meaker. Do you think there is a way around this?

 

Mill the IHS until it is level with the core and press down on it too to spread the pressure.

 

Gameplay temps and results for the .925 silver IHS lid and 17w/mk thermal pads on both. "Auto" fans and U3+ (3×) vardar fan mod with all fans set to 40%. More quieter that way.

 

Nice work, 68c max is something really impressive, please give some light about the BIOS settings you are using in this test.

 

Stock Clevo BIOS for the P870KM1-G

On Clevo CPU overclocking dash:
CPU ratio override: 45x on all cores.
Cache Ratio Override: 40x
Power limit 1 at 91w
power limit time at 8s
power limit 2 disabled
CPU voltage: Adaptive@ -145mv

I hope this helps.

 

I saw your video as well, nice channel you have.

 

Thank you. It's doesn't have the best enhanced edited videos but i thought its the best place to share some live results. Hopfully, some live results on the .999 silver IHS

 

next up, we just need to test IHS made of diamond, pair it up with liquid metal

 

fun fact: caseking here in germany started to offer advanced, pro and ultra versions of cherry picked coffee lake cpus. 8700K goes up to 5.2 ghz with the ultra version sporting....*drumroll*... a silver ihs!

reminded me of this thread

Sent from my HUAWEI NXT-AL10 using Tapatalk

 

Thanks for sharing. This is will save me a lot of money in the future.
Here is the link?:
https://www.caseking.de/en/der8auer-core-i7-8700k-5-2-ghz-ultra-edition-cpbu-162.html

Need one for Kaby Lake now. If anyone spots one; then share it here please.

I might stop the making of the .999 silver BitsPower IHS lid to save money. I want to move on to the P870KM1 heatsink, which I love to be in solid .999 silver. Also it be great if the heatsink be bareback, meaning no IHS on the CPU, just a straight heatsink clamp to the die. Does anyone know where to buy a P870KM1 CPU grid heatsink please?

 

how is the silver IHS vs the old IHS? also does silver last and not oxidize like copper?

 

Well, in the link above; Der8auer are stating:

"An additional reduction of 8 to 12 °C temperatures can be expected due to the exchange of the business standard Integrated Heatspreader (IHS) with a silver heatspreader. The der8auer Ultra Edition is the absolute apex in all things related to CPU modding and offers the very best overclocking performance."

On my tests, I am only getting a 2c drop on the 7700K CPU! but mine is in .925 silver and not in .999 silver like theirs.

As for the oxidizing; I have not opened up and checked my .925 Silver lid yet. I am going to wait another month at least. I will post photos here for future reference.

EDIT: The down fall to their silver IHS lid, is the height. The P870 CPU heatsinks needs that extra 1mm of height to fill in the gap for a better contact. For a desktop though, I bet their ace

 

As requested by some:
I have finished the video of the same run as last test on Battlefield: Hardline using the .925 Silver lid, but with "Full" fans on Clevo dash and full fans on the x3 U3+ Vardar mod cooler. The last test was on "Automatic" fans through the Clevo dash and 40% 1100RPM x3 U3+ Vardar mod cooler.

Last video (Auto Fans, 40% 1100RPM x3 Vardar cooler)
2x 1080 GPU: 57c max
CPU 7700K 4.5GHz: 68c max


This Video (Full Fans, 100% 3000RPM x3 Vardar cooler)
2x 1080 GPU: 49c max
CPU 7700K 4.5Ghz: 62c max

 

thats a beauty, 40% auto fan at only 68C is pretty nice, how much of an improvement do you think it helped? at 40% auto fan

 

Thank you.
Just to confirm; that's 40% fan speed on the U3+ fan mod, that consists of 3x FF5 120mm Vardar fans running at 1100-1200RPM when on 40% power and the Clevo dash is set to "Automatic" fan setting.

I think the "automatic" or "overclock" Clevo fan menu settings, is an unfair test. Because the CPU and GPU fans come on and work separately/independently depending which hardware reaches a certain amount temperature point. This cools down the overall heat on the whole heatsink that are attached together. This is why the "full" fan settings through the Clevo dash is more accurate for test on max temperatures IMHO.

The .925 silver IHS only helped me by 1-2c I think. It's not that much of a difference, but I loved the experiment.

I'm now trying to get a p870km cpu heatsink (or maybe the new Coffee lake) to see if I can create it in piece's of silver, then build it up. If I do, then I will make sure its a bareback heatsink, meaning the CPU lid has to be removed and the 1151 clamp to be down on top of the CPU by using some kind of housing. This will improve temps big time I think.

I'm turning this into a hobby lol.

 

.999 Silver heat spreader lid, is now going through its initial setup process. This is 1.1mm thicker than the Kaby Lake lid to fill in the P870 heatsink gap. This is also Coffee Lake compatible too. The small hole is for anyone who wants to turn it into a necklace afterwards


The Coffee Lake die is slightly bigger than the Sky Lake and Kaby Lake dies.

 

That's about the extra size the two extra cores take. It's still mostly IGP in size.

 

I always wondered how much of a big thermal difference the IGP makes to the overall CPU Temperature.. I know that the 7740X has no IGP, but it has a base clock of 4.3Ghz, a higher TDP at 112W and supports its stock speeds of 2666Mhz of DDR4 RAM so comparing it to the 7700K won't be accurate.
Because the 7700K is mostly made for gaming, it would be great if Intel just left the IGP out.

 

The IGP is still there jus disabled. Thermally it does not make much difference when disabled like in our desktop systems.

 

As requested by some:
Witcher 3: Wild Hunt gameplay on P870KM1-G using the .925 Silver lid, but with "Full" fans on Clevo dash and full fans on the x3 U3+ Vardar FF5 fan mod cooler. Everything is maxed out. Nvidia HairWorks enabled too. Temperatures and speeds are displayed in-game and before and after the video. On desktop the CPU drops to 800mhz on idle. This is because of Intel's Speedstep technology. In-game stays at 4.5GHz constantly though.

Thank you @Mr. Fox for showing me how to display my CPU speeds in-game by using HWiNFO64 with MSI Afterburner.

This Video (Full Fans on Clevo dash, 100% 3000RPM x3 Vardar FF5 fan mod cooler)
2560x1440
Nvidia Drivers: 387.92 Notebook
#1 1080 GPU: 52c max
#2 1080 GPU: 51c max
CPU 7700K 4.5Ghz: 60c max



*This video is rated mature*





Sorry about my Steam username, (unless you are a committed BGA lover) as I don't want random adds. Please take a joke.

 

As I have quoted at the P870KM1 owners lounge:

"What's taking so long is that .999 silver is hard to work with. It's hard to cast because it shrinks, it's hard to laser cut because silver is the most light reflective metal there is, and it's hard to chop because .999 silver is too soft. I have already lost a lot of money doing this, as I have tried multiple designs and process. This will be the last process I do. It's going to be done in Germany, with some Hi-Tec machinery."

To get this mirror finish and flat down to 0.0002mm accuracy (unlike the cheap finish Der8auer have to offer) it would cost £2364 for the initial setup and the £788 each for the .999 silver lid. Which is absolutely ridiculous. I'm not going ahead of this, and now I'm looking else where.

My block of .999 silver is coming back in bits too as they had to do a quote. Gutted!!

I'm going to try another manufacturing process. I want to know how Der8aur have done this?

This is the final update of the lid. It now has curved edges instead of corners:
Top View:



Bottom View:

 

It's always nice to make prototypes but boy can mass production come along and give you a good kick, just look at carbon nano tube stuff.

 

Carbon nano would kill me!
You'd need $1000's for a initial setup.
But don't worry. I have not given up yet with this .999 Silver lid. I'm still looking...
I need to find a way anyways, as I need them to build my new invention as well

 

Silver water cooling block?

 

There's already a .925 solid Silver water cooling block.

https://shop.aquacomputer.de/produc...=3573&XTCsid=6hsu3alk3rk1l7u5tm74vvcfs5gha6gq

Users are asking me for the Silver lid to go with it.

So its not that...

 

Lol, I wonder what the metal implications of so much silver are.

 

So how’s the .999% silver one been brought to completion? I will gladly buy a Kaby lake one From you if so.

 

Hi Andrew,
I'm still waiting for more quotes, as 5mm 99.99% silver sheets are hard and expensive to buy. Also working from bigger blocks cost much more and is a hassle to work with.

I'm no longer going through the much cheaper casting procedure like what Der8auer have done, because of the shrinking of the original master, sprue marks from ejection of the silicon mould, and warped (bad finish) uneven face. I am now trying to get it done through CNC milling machines. As I have stated on my older posts, that .999 silver is really hard to work with when it comes to cutting.

My quotes so far are:
1) £2364 for the initial setup and then £788 each for the CNC .999 silver lid.
2) My latest and cheapest CNC .999 silver lid quote is £550 each with no initial setup fees. The price does not change even if I buy in bulk. This company build top of the line space and aviation equipment and the manager guarantees "it will make Der8auer lids look silly" (his words).
3) .925 silver lid which I have now (which does not fit on a coffee lake CPU). I have already paid the initial setup and a single duplicate will only cost £65. The only down fall with this is that the lid is inaccurate and I had to place 2x 0.5mm shims on the 1151 bracket to hold down the lid because of the shrinking done in a casting procedure. See photo below:

I have also had a fall out with this company. I don't think I will be seeing them again...

Any new updates on quotes I will post it on this thread. I have not quit my journey on this universal socket 1151 lid just yet. It's just the companies have to see if it can be done before they give me a quote.

The bare-back heat sink for the P870KM1 will NOT be done. As is cost thousands to be made before it even goes into production. So I'm only sticking to the IHS lid for now... I only got one kidney left

@ole!!! It has been over 9 weeks now that I have used the .925 lid with liquid metal on the die of the CPU and on top of the lid to the heat sink. I have just pulled it out today so I can keep this thread up to date. This is the tarnishing that the liquid metal has done (photo below):
Note: this has now been cleaned back to its original finish... if not even better (please see next photo)


And here is my IHS lid collection including the .925 silver lid which I have just took out and cleaned. The blur image on the silver is the reflection of my pear tree. Sorry about the photo quality.

From first row, left to right: 1: 3D printed plastic prototype with 1mm extra height (top face), 2: BitsPower shining lapped lid (the finish is horrid due to the moulding procedure when the first .925 lid was casted), 3: Copper CNC machined lid with no 1mm extra height, 4: .925 silver lid after it has been cleaned from liquid metal. The marble effect on the face is a reflection of a tree.

From the second row, left to right: 5: 3D printed plastic prototype with 1mm extra height (bottom face), 6: Original Kaby Lake IHS lid from a 7700K CPU, 7: Original Sky Lake lid from a Pentium G4400 CPU, 8: Original Sandy Bridge IHS lid from a Intel Celeron G530.

 

Ah the path of science Cool pictures.

 

Wow, a company charges that much money and then gives you some BS! They wouldn’t go back and re-make it worth the extra couple millimeters you needed so you didn’t have to use shims? I can see why you had a fall out. LOL I read this whole discussion this whole thread and I really appreciate what you’re doing here. I’m a material sciences myself so I know what you’re going through.

Have you tried contacting Bitspower and using whoever they use for their CNC process for their “silver shining” IHS. If you have not seen one in person I urge you to pick one up As they are nickel plated copper but not for this reason for the reason that they are CNC machine I believe in Taiwan. Will cost you a whole lot less than any European route you choose to go with or any route. Call EK water blocks, Aquacomputer, etc as they run CNCS ALL DAY.

I’m sure you’ve thought about a lot of these things and these are just suggestions to help you get your end goal quicker.

 

The fallout was the manager saying that I wasting his time on a small order, and I should learn about casting procedures on the internet before I got in contact with them. I found out he was a nasty boss from other stories . But one of his workers helped me and guided me to the right and better direction. The 1mm-2mm loss they did on the .925 silver lid was my fault as casting procedure is inaccurate because on how the silver shrinks when it gets cooled down. I was thinking making a bigger prototype (master) so when it does get casted, it will shrink to the size I wanted it to be in the first place. But like I said; casting is uneven and does not create a super flat surface. Also there is spur marks when removing the master from the mould.

And thank you for you suggestion anything might help, but I have already contacted BitsPower and asked them for the CAD file or PDF drawing of the silver shinning. They declined very hard. I already own a few Bitspower Silver shining IHS lids by the way. There is one on the photo above, but it is brown due to me lapping the face straight for a better heat sink contact. Its labelled "2: BitsPower shining lapped lid (the finish is horrid due to the moulding procedure when the first .925 lid was casted)".

And you are correct! The BitsPower IHS lid is made from low grade copper then electroplated with a nickel finish for protection. The nickel finish slows down the thermal conductivity by the way. CNC machining copper is very cheap. I can CNC the copper IHS for around £20 each after the first initial payment.

It's silver that we need my friend. It's an enthusiast dream, We want to be able to hold an equity while getting a few degrees of temperature drops all at the same time. I have mentioned this silver experiment before this thread was created, stating that i will 100% make one and I got a feeling I gave other companies the go, go (race) on making a silver IHS lid. Before this, I used to google search "silver IHS lid" and no one ever created one. It was only talked about... So I will just like to inform every one I have another better idea on thermal conductivity. But I will leave it until it is fully completed before I say anything...

I will try EK water blocks, AquaComputer, etc when I have time. Thank you for sharing .​

I have contacted other major PC companies, and their secret to cheap prices per unit is: mass production, cheaper materials, Taiwan or China cheap labour, contract agreements on the amount of units to be made, also some companies license their designs to factories and then the factories sells them under a different name. These tech giants are eating the little fish you see .
While we are on this subject: Its only other day I just found out that AlphaCool and Fujipoly buy their thermal pads from the same company/factory, but have the thermal pads packed by their own packaging that's saved in the companies/factory warehouse where the thermal pads are being built ... If only I had extra money I could have "Stress Tech 17w/mk thermal pads" on sale now (£78 for 100mm x 100mm 17w/mk 0.5mm unit).

The moral of the story; is that successful companies will never help me... But its worth a try
I hope to reach my goal soon. I will keep you guys informed. Peace out

 

Well, I think you are headed down the right alley, you just need a little help. I did my PhD at UCSB in the materials department and may be able to offer some help. I am not going to disclose the information on the thread though because this is considered intellectual property and a trade secret for my company AM Thermal Innovations. If you are interested in making this a joint venture or just need professional advice and supply channels to achieve your end goal, I would be glad to offer my help. shoot me an email at andrew,[email protected] and we can talk.

 

That's excellent. I will send you a private message with my Email address.
And just to give you a heads up. My new thermal conductivity idea which is not stated on this thread is with a certain man (an investor which I can not state his name on here) hes helping me to go through the international patent procedure. So I can not really share this idea for now.
Please share with me any directions or thoughts through my Email address. Anything is greatly appreciated. Thank you.

 

A thin plating of nickel is negligible to thermal performance.