7700X lapping, replace IHS or direct die cooling?

[ChrisM255]

Start lapping, and test it yourself!

GL.

I've always done hi-grit and it served me well. But you're absolutely right but again I'm asking for others experiences if they have any if not then... You're just posting to post but you're not helping the conversation

 

[ShrimpBrime]

If you read the original post I'm asking for anybody that has specifically had experience in 7000 series I've done this many times and I know the process. Wins/losses experienced in ryzen 7000, that kind of thing
But I definitely appreciate the level of detail here and I know you're just trying to help out so I don't mind these type of posts. Thank you!

Won't matter. It's a nickel plated copper plate. Haven't done a Ryzen 7000 series chip, nor would I spend the time because I already know the results.

As I mentioned, probably direct die would make more of a measurable difference.

You're welcome, but after 20 years of the same fkn conversation with people about this subject, I think I'm worn out on it.

I wish you luck as I will watch from the sidelines if you actually lap or de-lid. Either way, I'd practice on any cpu you have laying around or can fine cheap or free. Remember practice will make perfect.

Good Luck!!!

 

[ChrisM255]

Won't matter. It's a nickel plated copper plate. Haven't done a Ryzen 7000 series chip, nor would I spend the time because I already know the results.

I really don't know what to expect from this series versus previous because it's a thicker IHS than they have ever used before it's higher power with a smaller contact patch for your cooler in terms of square millimeters. I can't say you're wrong, but if they didn't do the IHS thickness to match the am4 cooler height so you could carry over, it would have been thinner... AMD even admitted to the IHS height for cooler compatibility to am4 so it's not just a theory or someone's opinion. They announced it as if it was a feature.
From a thermal exchange standpoint, this just doesn't sound right to my old under-educated ears... the more I say "thicker IHS + higher power + smaller contact area" it sounds more and more wrong each time. There might be something about just the thermal mass being optimal and all us skeptics are just simply wrong, but I haven't ran across anyone saying that. One of the reasons I wanted to talk about this. Can't be everywhere on the internet. I can watch all of the videos, 3rd party testing, thermal exchange and equilibrium and engineering documentation I want and would never be able to cover all of it by the time I moved on to my next PC build. I get only a few hours to dedicate to this weekly, if I am lucky.

I'm glad somebody is rooting for me...

 

[ShrimpBrime]

I really don't know what to expect from this series versus previous because it's a thicker IHS than they have ever used before it's higher power with a smaller contact patch for your cooler in terms of square millimeters. I can't say you're wrong, but if they didn't do the IHS thickness to match the am4 cooler height so you could carry over, it would have been thinner... AMD even admitted to the IHS height for cooler compatibility to am4 so it's not just a theory or someone's opinion. They announced it as if it was a feature.

I'm glad somebody is rooting for me...

Higher power than what?

AMD used the same old plate all the way up. Fx-9590 was 220w at stock.

The height for coolers could be what everyone wanted to hear, but likely near the same mass as previous plates. This means it would hold, store, dissipate similar thermals. However the surface area looks reduce from the notches.

Processors are a bit more transistor dense on a smaller package.

The next problem is removing heat in only 2 dimensions. The top of the cpu, up and out.

Then trying to get everyone to research what coefficient means. Temp gradients and such.

Then, on top of all that, boosted to max from the box with no headroom.

Experimenting with a lidless 2700X, cold plates and Peltier cooling, this is where I stopped using AMD ryzen processors. I have 5000 gen, but it's my kids, and he uses it normal.

I like just doing cooling modifications. There's not a lot of guys that can say they've cooled a 2700X at 4ghz under 1.2v at load with a peliter.

No, this isn't a Ryzen 7000. But the laws of thermal dynamics don't care what cpu you have. They laws don't change. Not for you, not for me and certainly not for boggledbeagles.

 

[ChrisM255]

Good point on the power. But as you pointed out density and also die architecture has changed and due to variations throughout the platforms they have all run differently, some being super hot runners and others able to cool to reasonable temps at higher powers. I feel like this is a little like comparing oranges to tangerines.
I mainly think that the transfer area can be improved between the IHS and the cooler, there's just not much surface area there for transferring heat off of the IHS and with everything being so densely packed on those dies the height may help to spread it out but you're still may be oversaturation on an IHS with that much z height above the die. But you cannot improve the surface area without replacing the IHS. Lapping or height reduction I think will only make things worse due to the heat not being able to spread out properly at the contact surface.
The only reason I think this series hits that TJ Maxx right away at stock is because the IHS may be getting oversaturated.
I wouldn't put it above a manufacturer to do things to hide what's really going on behind the hood. They claim no throttling but then how can it reach 95 so quick and just cap there if something wasn't throttling. When I watch my numbers as I'm doing a stress test based on what I see I would expect to see throttling coming up but it doesn't I think they may be just hiding it. Manufacturers get caught doing this all the time.
I think AM6 is going to come a lot sooner than we think due to this issue and it's unpopularity. In a couple years from now I suspect there will be no more IHS's like this. At some point they're going to address everybody's hatred towards the high temps.
I know I don't know everything and probably don't know nearly enough. Hence the conversation about the platform looking for anybody who has experience or has seen something that I haven't before I just go diving in half blind with only references to previous gens and known mounting issues with cooling modifications on this platform. I'm honestly not seeing hardly anything but theories and mounting issues when I try to look it up...

Shrimp, I honestly Love your input keep it coming if you're willing, you're giving me at very least food for thought and I'm not saying you're wrong I just have a different view.

I just looked it up just to make sure and the IHS is 1.7mm thicker than previous gen with more than 20% less surface area for cooler contact. Trying to find the square mm for am4 vs am5 to get the true area reduction. I have seen it just don't remember that figure.
Am5 IHS is 3.7mm
Am4 was 2mm

 

[mechtech]

Better than a copper plate, but not by a huge great deal because back to copper waterblock. Maybe a 2c drop at best. Marginal I suppose would be a good way to word it.

I still have it, but seldom use it.

View attachment 330235

1879 coin?! Think I would have bought a 1 or 2 oz bullion bar for $30

 

[ChrisM255]

Surface area was reduced by ~30% AM4-AM5
AM4 ~1300 sq mm
AM5 ~910 sq mm
And IHS thickness increased by 1.7mm

That is just too drastic for me to believe the thermal exchange is "perfect" due to just thermal mass. Years ago I saw an engineer from Intel interviewed talking about how important the z-axis is in terms of heat saturation and thermal resistance.

Higher power than what?

I digress on the wattage... sorry. I really thought they upped it for AM5 across the board but didn't have any doubt about it till you said that, but that was only for the X3D AM4-AM5
Good call out!

Still don't like the increased z-axis and contact reduction... there is definitely still a problem with the IHS.
It could be argued that with the increased density on the dies requires an increase in z-axis to spread out the heat, but the density increase was not enough to justify a 1.7mm increase almost doubling the amount of copper it has to pass through. No-one has ever done an increase that drastic for even close to that much density change. That's a lot of resistance increase that has to be saturating the IHS given the fact that it can't spread out horizontally as much with a 30% reduction in the surface plate and most of the sides cut out.

Am I way off thinking this?

 

[ShrimpBrime]

1879 coin?! Think I would have bought a 1 or 2 oz bullion bar for $30

My brother insisted. The coin wasn't worth more than its weight. We have a bag of em in a special place, so no worries and definitely not one of a kind.

I've a few cold plates to work with.

My home made Dice pot is Copper base from a waterblock and brass handrail. Brazed the pieces together.

More pictures? I feel they say 1000 words I don't feel like typing.....

 

[Launcestonian]

Surface area was reduced by ~30% AM4-AM5
AM4 ~1300 sq mm
AM5 ~910 sq mm
And IHS thickness increased by 1.7mm

That is just too drastic for me to believe the thermal exchange is "perfect" due to just thermal mass. Years ago I saw an engineer from Intel interviewed talking about how important the z-axis is in terms of heat saturation and thermal resistance.


I digress on the wattage... sorry. I really thought they upped it for AM5 across the board but didn't have any doubt about it till you said that, but that was only for the X3D AM4-AM5
Good call out!

Still don't like the increased z-axis and contact reduction... there is definitely still a problem with the IHS.
It could be argued that with the increased density on the dies requires an increase in z-axis to spread out the heat, but the density increase was not enough to justify a 1.7mm increase almost doubling the amount of copper it has to pass through. No-one has ever done an increase that drastic for even close to that much density change. That's a lot of resistance increase that has to be saturating the IHS given the fact that it can't spread out horizontally as much with a 30% reduction in the surface plate and most of the sides cut out.

Am I way off thinking this?

No, your making sense with the laws of physics here. The AMD engineers had to make a decision about backward compatibility with AM4 coolers & being "adequate" for the uniqueness of AM5. That's why they raised the official heat limit to 95C for Zen4. With a figure like that it sent shivers up the spines of enthusiasts no doubt, but the engineers assure us that 95C is fine...

My Rocket lake i7 K series has a 100C limit & that is based on 14nm tech. There is no getting away from these laws of physics.

I also think the 8000 series will be physically the same outward size as the 7000 series, it makes no sense that AMD would suddenly change it but with the same socket compatibility for AM5. So an investment in specific tools to get serious about lapping the IHS is not a bad idea for enthusiasts to consider at all.

There is lots of other discussions about this topic around on the net for those interested...

 

[ShrimpBrime]

Surface area was reduced by ~30% AM4-AM5
AM4 ~1300 sq mm
AM5 ~910 sq mm
And IHS thickness increased by 1.7mm

That is just too drastic for me to believe the thermal exchange is "perfect" due to just thermal mass. Years ago I saw an engineer from Intel interviewed talking about how important the z-axis is in terms of heat saturation and thermal resistance.


I digress on the wattage... sorry. I really thought they upped it for AM5 across the board but didn't have any doubt about it till you said that, but that was only for the X3D AM4-AM5
Good call out!

Still don't like the increased z-axis and contact reduction... there is definitely still a problem with the IHS.
It could be argued that with the increased density on the dies requires an increase in z-axis to spread out the heat, but the density increase was not enough to justify a 1.7mm increase almost doubling the amount of copper it has to pass through. No-one has ever done an increase that drastic for even close to that much density change. That's a lot of resistance increase that has to be saturating the IHS given the fact that it can't spread out horizontally as much with a 30% reduction in the surface plate and most of the sides cut out.

Am I way off thinking this?

Not over thinking. Nah, pretty accurate. But I don't feel the old style IHS plate wouldnt be that much different.

Because to do "PROPER" testing, you would have to solder the old style plate back onto a chip you removed a plate from.

Not impossible, but also way high risk of roasting a chip. I do not have a proper station for such procedures and would expect a 50% failure rate just on the first try with a hand torch. But the nice thing is there's no chance of loosing pins, but those pads can come off just the same as well.

The thicker plate is an issue because it lacks the surface area to transfer wattage. It's only good for so much, just like any copper plate.

Copper plating. What have people really tested with. Reviewers actually put effort into thermal transfer? Actual testing? Maybe it was designed to be thicker. Maybe they wanted to run the cpu with a higher temp gradient. Many air coolers, as I've said before many times, are more effecient at higher temps. So the designed the power envelope around certain coefficient numbers with their own models and testing. The thicker IHS plate is acceptable.

Grain of salt??
No way. Everyone's ambient temps are different. Humidity also plays a role. Fan placement, case size, static pressure, quiet fans, water cooling. Air cooling.

You all move the same wattage around the same rates. The temp only matters at throttle point. Right?

Oh no, your AMD at 5.2ghz hit 90c. So what? That's what gamers demanded for desktop processors and overclocked the shit for ya. They used to release processors at a bit lower frequency and power, and they also supplied the coolers too.

So, overclocking on ambient died. It's not a thing. 4 billion gaming PCs and only a few thousands would consider themselves competitive benchers. Cause for you to beat my scores (some of them) you'd have to graduate to extreme cooling.

So, no. The IHS plate on a 7700X is fine. I don't belive modification is necessary. Put the cpu all core to its base clocks, and you'll see what I mean. Turn the "auto boost oc" off and tell me these plates are an issue. BS. The plates are fine.

 

[ChrisM255]

I get your point, but no-one "wants" to set it at base and turn off boost... not willingly... definitely not me.
But if for instance this was a concern for me and I wanted an 8 core CPU on this platform, I could've just gotten a 7700 and been much more comfortable or just go get a better cooling solution. Have not seen anyone complaining about the non-X skus unless it is a terrible pre-built (big box shelf product) or overclocked and lets be honest those things are most often nerfed at the bios level and locked. I would really like to converse with an AMD engineer to see what they calculated the IHS to be optimally before they were forced on the AM4 cooler compatibility path.

I agree to do proper testing as in a from the factory equivalent would need to be soldered and would likely cook the chip without proper setup that uber controls the environment. But I think we can still get an idea form using liquid metal in place of the solder. Taking the numbers from liquid metal and add a couple/few extra degrees and this is what should be expected form solder. Not an exact science but I bet I can find a chart showing the differences in stock solder to liquid metal and get fairly close.

I agree with the the possibilities of needing a thicker design and higher temperature gradients, but that's a lot of maybes and speculative thinking. I don't know, one of the reasons I want to experiment on it.

Everyone's temps are different due to the overabundance of options and builders... including builders throwing crap together that should never be combined in cases that should never have been made. Different cooling solutions and different environments. But you are also absolutely not wrong that just about everything can throw off results from one user to another in comparison.

Temps don't just matter at throttle point though. Rate of climb effects the algorithm's behavior while ramping, it is going to adjust it's boost behavior if too much saturation occurs.

Never really considered that they are basically overclocking to the max to satisfy gamers and enthusiasts and not to mention to beet the other guys on the performance numbers... Some do still provide the cooler, the AMD non-X skus with lower clocks, thermals and freq... Hmmm

Overclocking on ambient died? Maybe mainstream, but not for me! I don't need to make a leaderboard to feel good about what I am doing or to feel that feeling of success when I hit a PR. If your only incentive is to be on a leaderboard than yeah ambient is not for you. The last few systems I have done stuff like this with, I was able to consistently get results equal to a lot of AIO overclockers with my wimpy air coolers and that is still a win for me and that is all I need. I am fueled by PRs and the love of the process.

Plates are not fine... aaaahhhh!

I am going to be clear for anyone that might stumble upon this later and not read the mile long thread we have going here (long for me anyways for a casual read). I do not "need" to do this, My AK500 does sufficient cooling for the 7700X and a water cooler would do even better. I am experimenting and testing my theory that is shared by many that the IHS is garbage on the AM5 platform, at least for the X skus. I know that there is some content around this out there but I want to experiment for myself.

This is what I am going to do...
I will not be using an AIO for testing... the ones I have need to be cleaned if used for testing for good results but pumps are not what they once were either and I am not going to spend ~$150 dollars or more on something that I am not going to use long term, sorry I know this will likely upset some. I apologize.

This weekend:
1. I will be pulling my stock #s and results with the current stock 7700X, Kryosheet and AK500 Zero Dark with dual Arctic static pressure fans
2. Replace the Kryosheet with Kryonaut and run again at stock. ding this so I don't have to invest in a million Kryosheets... degradation with each re-use, not ideal for testing
3. Lap surface of IHS, lap surface of the AK500 install with Kryonaut and run again

Tracking and bench tests:
HWinfo64 v7.68 cleared before each run
Cinebench R23 10 minute run single and multi
Maybe some gaming numbers, might not have time day of release of info but will eventually do gaming numbers

Following:
Do incremental IHS reduction and track at 0.2 MM increments, lapping and doing runs at each increment, stopping at likely a 1.0-1.7 mm reduction depending on where/if I start to see diminishing returns.
I will likely have to mess with the mounting with the different heights, but I have a plan to help with this. Just need to find some 0.2mm spacers, likely nylon or some other non-conductive plastic or change out the posts for something more adjustable. I am pretty meticulous so I should be able to figure this out.

Following that:
De-lid and replace with a performance IHS using Conductonaut liquid metal between die and IHS and Kryonaut to the lapped AK500
Replacement IHS = TBD

Direct Die:
TBD - saving for investments needed for this endeavor... I would do custom water cooling for this

Keep in mind this is being done on my free time. I am not a content creator that gets paid or reimbursed through add revenue for their time. I run my own small business and work a 50+hr/wk job and have 4 kids in the house. Life happens... Just be patient and I promise I will deliver. Gotta go in today for an emergency electrical service and deal with server power backup and shutdown if needed.. but tomorrow I should be able to get started.

I will except all suggestions, but will likely not be persuaded from my course of action above unless I see something really compelling that does not require a large investment.
Thank you for any and all input so far, hope you guys are excited to see the results as I am!

 

[ShrimpBrime]

It's not about wanting to turn off the auto OC because gaming. You really need to jump past that and realize I am only talking about moving BTU around at different rates/watts ect.

That's what I mean. Why I get tired on an ages long old conversation. The minds wonder and miss Shrimpy's point.

The rest, I throw to the wind.

Here. A put some stuff on a box. These are some of the various types of waterblocks and cold plates I've played around with.

The XBOX plate is really nice and thin. I used it with a couple lidless s939 chips with dryice or LN2. It didn't make any difference in temps. I could have used the cold plate there to the left. That's an opteron 148. And I can use that plate on just about anything. Have another in a different box which has the raised sides ground off. I used a grinder, not lapped.

The other large plates are feom server heat sinks. I removed all the fins from l, similar to the top left heatsink there with the mounting screws sticking out.

The waterblocks are a variety I use depending on the wattage for any cpu and any IHS plate. The center left is an old ThermalTake block that the cpu melted the acrylic off. This happened feom a switching power supply failure and the cpu baked for an unknown amount of time until I came home from work. I'm lucky to not have had a fire started. That's actually a plate of steel with barbs I brazed on.

The center Supreme HF is my favorite. Believe it or not, the 1366 slots actually fit well on 1700 LGA mounting holes. And it's big, heavy and great with all high wattage processors.

The heatkiller though looks large is about 30% lighter. It does pretty well with most chips, but not as good as the first 2 blocks described above.

The cold plate I use for TEC cooling only. The large ones there.

All this you see, are ALL within 10c end temps because of only the ambient air temp. Any other time, I will have applied chilling.

All the copper here before you is the same. The rate of thermal travel in copper does not change.

Now @ChrisM255, if you are really interested, like seriously deeply interested, I'll be happy to send you a cold plate, maybe a waterblocks to play with. The plates I can let you keep, the waterblocks you want to test with must be returned. I don't thing I have any Peltiers left that are any good. I'd send you a couple of those too. But I'll look.

 

[ChrisM255]

Out of box settings
untouched stock IHS
Kryosheet + Stock AK500 Zero Dark w/ dual Arctic 120mm pwm static pressure fans in push/pull

Idle: ambient 19.1°C
Let it come up, waited for background loading to finish, reset, gave it a few to capture data and screenshot

I was confused when I saw "enhanced" next to the CPU label because I just reset to stock, but didn't reset expo, but I am OK with this. It is running at the rams rated speed so I will keep this the same for all testing. Stock freq and voltage with expo enabled. ASUS "standard" fan profile in bios

Cinebench single core

Cinebench multicore

I just want to say that this is in a cold basement I would hate to see this thing running in like a hot bedroom during summer... I imagine lots more red! I literally have to wear a hoodie and my feet are freezing. I'm going to check ambient with each set of runs I do but seasons change the temperature down here not my PC it's in the main space in a large basement, PC doesn't change the temp down here even after 4 hours of gaming.

 

[BoggledBeagle]

This 140W at 92°C, what is the frequency and CPU voltage during running this load?

 

[ChrisM255]

Yesterday was pretty nerve-racking for me with the electrical project for various reasons. This is my only day off this week and I need a mental break so I'm gonna game for a bit and pick this up later today. I will post my game temps at current settings.

This 140W at 92°C, what is the frequency and CPU voltage during running this load?

Good call hang on
Touched 5.4 all core effective but was 5075-5100 on all cores steady during the run. I'm running multi again and will repost the score with expanded view on HWinfo64

Average effective all core looks to be around 5060 to 5080 during the run so far

5.4 was Target not effective sorry my eyes aren't working right this morning posting new screenshot in a second

during 2nd run

after 2nd run

heading to MW3 Zombies
TerribleGamer33 if anyone is interested!

~45 mins gaming MW3 stock IHS, stock settings with expo enabled, ambient still that steady 19.1°C

 

[AVATARAT]

You can see CPU temps.

Water temp 20C, ambient 21C.
PBO -17 all core, Offset -10, LLC 3 (lowest):

Single:

Multi:

5550MHz all core, auto voltage.

Single:

Multi:

Of course, this is not Cruncher's sable, but you can see how CPU temps go lower and the result goes upper.
In the beginning, you need to do better airflow in your case or something similar if you want better results.
Delidding will offer you a small bump, but your main problem is higher temps in your case/cooling.

 

[BoggledBeagle]

Average effective all core looks to be around 5060 to 5080 during the run so far

I was curious about my Intel CPU under a small Freezer 34 aircooler, I enabled just 8 P cores, limited it to 145W, it ran at 5100 MHz, voltage was 1,16V and maximal temperature 76°C (19°C ambient), score 21400.

So very similar conditions, only temps are much lower and score a bit higher...



Then I enabled all cores, frequency fell to 4600 MHz, voltage to 1.04V and max. temp to 62°C. So while pushing 145W in an air cooler, the CPU can run pretty cool WITH LOW ENOUGH VOLTAGE. The point I am trying to make is that even with a CPU that has higher thermal resistance, you can reach pretty good temps when you run it slow enough with low enough voltage.

 

[ChrisM255]

Prior to this morning I was running CO-20, additional offset (can't remember how much), +500MHz boost and temps weren't too far off from my current with this AK500, but frequencies and scores were much higher. Wish I would've taken screenshots at the time... you guys definitely have me beet with the water cooling for sure, but not bad for an air cooler, no?

I tweak constantly so I never remember where I settle at. All gets fuzy... legalizing you know what in my state really didn't help with that at all!

Boggled, that's a freaking 14900K! moneybags...
But interesting, thank you for sharing!
I would like to see this on a 14600 or 700 sku for a ~closer comparison though
Still talking about 10nm vs 5nm though so different densities and core counts and I feel this will still make a difference and in your case will be easier to cool if capped at the same power.
In my opinion it is hard to compare intel vs ryzen 100%. I mean do you compare by price or by cores? Really can't on cores anymore so its all just apples to oranges to me when comparing AMD to Intel.

I never said you were wrong on lower voltages, limiting or any of that... I play around with that too
I've gotten some great results with this CPU without meaningful performance degradation and around 70C peak temps but that is just not the point of this exercise
I want to see how the temps change as I reduce and change out the IHS out of sheer curiosity all using stock settings and then I will likely play around with the settings to see what I get there.

Avatarat, that's not bad! Surprised with your water temp that it still touched 92 what pump are you running? is that custom or AIO?
Sorry if you said it and I missed it.

Honestly I didn't bother with a X variant of the Ryzen 7 7700 that's why I took the non-x it's 65W it's easy to cool, I can use my Noctua NH-D15 chromax.black and if I need the performance I can just unlock it to a X variant.

Am I missing any performance? Not really if looking at tests

This was why I almost paid the same for the non-x version of my CPU and I didn't want to pay up for a 7800X3D at the time and I still don't for my 1440p gaming.

I almost took this route after hearing about the temps on the x skus. Was hard to pass up a 7700x for $205 though...

Then I enabled all cores, frequency fell to 4600 MHz, voltage to 1.04V and max. temp to 62°C. So while pushing 145W in an air cooler, the CPU can run pretty cool WITH LOW ENOUGH VOLTAGE. The point I am trying to make is that even with a CPU that has higher thermal resistance, you can reach pretty good temps when you run it slow enough with low enough voltage.

didn't realize you took it this far, NICE!
I really need to stop skimming! Reading is not my strongest skill...

I also have a thick mirror that came from an old infrared alignment machine.

And it's what you see in the picture I've already demonstrated.

The SMDs are approximately 3 slices of paper away from that mirror. I removed quite a bit. It started as soldered actually.

So here, I'll demonstrate, again, how you can see lapping brings it much much flatter. Nice and flat.

This is going to be around .4 to .6mm removal only.

Final product mirror finish 2000grit and higher. Some people go nuts with leather polish, but F all that.

________

Wait till I tell you that very course finish actually does well for thermal transfer...... Don't ask why, it's probably a dumb answer.

View attachment 330255
View attachment 330256
View attachment 330257
View attachment 330258
View attachment 330259


Oh I see what you're trying to say.

The picture above, Athlon 5000.... On that mirror?

I lapped the entire IHS plate off. it wasn't a conventional de-lid. Many people have done this in the past. Intel chips too.

But 200 grit, I'll take off 1.5mm of material in just a few minutes. By hand. On the mirror. On a flat enough table surface, BAM! be no problem, and flat. Flatter than stock from the factory.

Here's Lapping Solder out of the silicon.

(I haven't looked at these pictures in a long time. It cost me 5 bucks to photobucket to get my photos I uploaded that many years ago. They hold my pictures for ransom to get my money)

View attachment 330265

I don't know if you added after posting and I didn't catch the additions or maybe it was my terrible skimming but didn't see all this.
AWESOME!
the compound polishing is what I have heard actually adds some insulation value to the surface. DerBauer is who I am quoting here, but have heard it other places as well

 

[AVATARAT]

Avatarat, that's not bad! Surprised with your water temp that it still touched 92 what pump are you running? is that custom or AIO?
Sorry if you said it and I missed it.

Custom loop, but that's not the point.
You need good airflow if you want to use an air cooler.
You can set a TDP limit, this will help too.
You just need to optimize your CPU for your current cooling.

If you do anything on the CPU (lapping/delidding/etc) you need to manage the heights between the cooler and the CPU, I am not sure how you can do this with an air cooler

 

[ShrimpBrime]

Prior to this morning I was running CO-20, additional offset (can't remember how much), +500MHz boost and temps weren't too far off from my current with this AK500, but frequencies and scores were much higher. Wish I would've taken screenshots at the time... you guys definitely have me beet with the water cooling for sure, but not bad for an air cooler, no?

I tweak constantly so I never remember where I settle at. All gets fuzy... legalizing you know what in my state really didn't help with that at all!

Boggled, that's a freaking 14900K! moneybags...
But interesting, thank you for sharing!
I would like to see this on a 14600 or 700 sku for a ~closer comparison though
Still talking about 10nm vs 5nm though so different densities and core counts and I feel this will still make a difference and in your case will be easier to cool if capped at the same power.
In my opinion it is hard to compare intel vs ryzen 100%. I mean do you compare by price or by cores? Really can't on cores anymore so its all just apples to oranges to me when comparing AMD to Intel.

I never said you were wrong on lower voltages, limiting or any of that... I play around with that too
I've gotten some great results with this CPU without meaningful performance degradation and around 70C peak temps but that is just not the point of this exercise
I want to see how the temps change as I reduce and change out the IHS out of sheer curiosity all using stock settings and then I will likely play around with the settings to see what I get there.

Avatarat, that's not bad! Surprised with your water temp that it still touched 92 what pump are you running? is that custom or AIO?
Sorry if you said it and I missed it.


I almost took this route after hearing about the temps on the x skus. Was hard to pass up a 7700x for $205 though...


didn't realize you took it this far, NICE!
I really need to stop skimming! Reading is not my strongest skill...


I don't know if you added after posting and I didn't catch the additions or maybe it was my terrible skimming but didn't see all this.
AWESOME!
the compound polishing is what I have heard actually adds some insulation value to the surface. DerBauer is who I am quoting here, but have heard it other places as well

He's kind of perfected a bit of what was already going on when he was a young lad, new to the scene.

The shaving razor trick is something he had gotten from an old how to thread created many years ago at Extremeoverclockingforums. I may have had a hand in his procedures. Not sure, you tell me....

The De-Lidding of Phenoms I-&-II Tutorial. - EXTREME Overclocking Forums

The De-Lidding of Phenoms I-&-II Tutorial. Essential Tutorials, Guides, & Info

forums.extremeoverclocking.com

 

[ChrisM255]

Custom loop, but that's not the point.
You need good airflow if you want to use an air cooler.
You can set a TDP limit, this will help too.
You just need to optimize your CPU for your current cooling.

If you do anything on the CPU (lapping/delidding/etc) you need to manage the heights between the cooler and the CPU, I am not sure how you can do this with an air cooler

My case has "decent" ventilation, 3) 140mm and 2) 120mm. After doing this build I stuck a thermal couple in the case right in front of the CPU fans with my multimeter and it was always within 10° of ambient if I had a load on my GPU, if I was just doing testing or something on CPU only it was just about ambient. I typically run a fairly aggressive custom fan curve for each fan too making sure it's always at positive pressure within the case. But to be honest my old thermaltake v100 had the best cooling I've ever had when I did my high air flow mod to the front of the case and opened that up. Having a 140mm fan pumping air on the side of your case right at your GPU is just way better in my opinion. I'm honestly thinking of going back to it but the cable routing and space behind the motherboard tray is just poo! It's doable but way tighter than you want it to be.
Although now that you mention it, I defaulted to Asus standard fan curve I probably need to look into this for case balancing. Good call!
Unless I see something terrible I'm not going to balance it just because everything else right now for this testing is stock settings. And I already removed the Kryosheet...
With my particular cooler my idea is that with a millimeter or so of heat spreader before the heat pipes I feel like I can do an over reduction on the IHS and kind of hit a sweet spot between the current IHS and current cooler and after that I am curious about a replacement performance IHS to get a better contact patch on the cooler interface and just see what I get. My guess at this point is that I will have more luck with a reduced IHS and the performance IHS is going to be more viable for a water cooling or direct pipe air.

The De-Lidding of Phenoms I-&-II Tutorial. - EXTREME Overclocking Forums

The De-Lidding of Phenoms I-&-II Tutorial. Essential Tutorials, Guides, & Info

forums.extremeoverclocking.com

Nice post! I only skimmed some of it so far but I'll go back when I have more time to focus. One thing I did pick up as a vibe to me, people were a lot more ambitious and curious.

 

[ShrimpBrime]

My case has "decent" ventilation, 3) 140mm and 2) 120mm. After doing this build I stuck a thermal couple in the case right in front of the CPU fans with my multimeter and it was always within 10° of ambient if I had a load on my GPU, if I was just doing testing or something on CPU only it was just about ambient. I typically run a fairly aggressive custom fan curve for each fan too making sure it's always at positive pressure within the case. But to be honest my old thermaltake v100 had the best cooling I've ever had when I did my high air flow mod to the front of the case and opened that up. Having a 140mm fan pumping air on the side of your case right at your GPU is just way better in my opinion. I'm honestly thinking of going back to it but the cable routing and space behind the motherboard tray is just poo! It's doable but way tighter than you want it to be.
Although now that you mention it, I defaulted to Asus standard fan curve I probably need to look into this for case balancing. Good call!
Unless I see something terrible I'm not going to balance it just because everything else right now for this testing is stock settings. And I already removed the Kryosheet...
With my particular cooler my idea is that with a millimeter or so of heat spreader before the heat pipes I feel like I can do an over reduction on the IHS and kind of hit a sweet spot between the current IHS and current cooler and after that I am curious about a replacement performance IHS to get a better contact patch on the cooler interface and just see what I get. My guess at this point is that I will have more luck with a reduced IHS and the performance IHS is going to be more viable for a water cooling or direct pipe air.


Nice post! I only skimmed some of it so far but I'll go back when I have more time to focus. One thing I did pick up as a vibe to me, people were a lot more ambitious and curious.

There was a purpose to overclock back then. The processors weren't boosted to the max like they are today.

You don't have to do anything to your setup.

I'm only here because you seem to have a genuine interest. Otherwise, it's talk for nothing. Which so far, it's just a bunch of talking. I've already done all the doing. Someone else's turn.

 

[ChrisM255]

I was taken back by the engagement of the audience and them saying they wanted to try stuff. It was nice to see.
But wife kind of put my play time on hold... Like I said, life... Happy wife happy... wife!?
I swapped the Kryosheet for Kryonaut and when I turned the PC back on my wife was glaring at me from the stairs with crossed arms... I've just been sneeking in replies and posting on my phone, lol!

I know you have no reason to believe anything I say, that's ok. You'll see.

 

[ShrimpBrime]

I was taken back by the engagement of the audience and them saying they wanted to try stuff. It was nice to see.
But wife kind of put my play time on hold... Like I said, life... Happy wife happy... wife!?
I swapped the Kryosheet for Kryonaut and when I turned the PC back on my wife was glaring at me from the stairs with crossed arms... I've just been sneeking in replies and posting on my phone, lol!

I know you have no reason to believe anything I say, that's ok. You'll see.

I believe what you say. I don't have a reason not to. Never implied it. I'm usually forthcoming.

All the replies people have given you are valid.

They have not lapped a 7700X because the cooling is adequate. And so is the IHS plate.

You may shave off a couple of Celsius. It's all good. As you've seen, I've tested many high watts. With plates of all sizes, for a few different style cooling methods.

Cause if you come back with some miracle 12c shaved off suddenly, then there are other physical changes made during the process.

Truthfully, a larger thicker cold plate does better for extreme cooling conditions.

I think of you had a custom loop, nice heavy waterblock, you'd have phenomenal temps and never so much as just clean your IHS plate.

 

[ChrisM255]

Sorry Shrimpie, you totally called it earlier, please don't get mad...
Sooo, I calculated the difference in mass and thermal resistance from AM4-AM5 IHS's

The IHS top plate, like just the top not including the side mass, it went from:
AM4 5800X - TDP/PPT 105/142w, 1300 sq mm @ 2mm thick = 2600 cu mm
AM5 7700X - TDP/PPT 105/142w, 910 sq mm @ 3.7mm thick = 3367 cu mm
That is a 29.5% increase in mass with an 85% increase in z-height
Using the calculations for the thermal conductivity of copper using the dimensions:
AM4 IHS results has a thermal resistance of 0.0038484 degrees C/W
AM5 IHS results has a thermal resistance of 0.01018414 degrees C/W

Its 2.6463309 times the thermal resistance and 30% less contact area to transfer to the cooler.

I'm not saying your wrong about any of your cooling solutions, or that there are no other ways to take care of it or even that it can't run at 95C.
Just for the love of Pete, stop saying that the IHS is "adequate" and "fine"
ALL I'm saying here is:
They F'd it up, come on, ADMIT IT!!! They put a bad IHS on it...