Noctua Presents NA-TPG1 Thermal Paste Guard for AMD AM5

[Lionheart]

So what? Let it fill the space in between. Why bother at all? All my CPUs have residue on the sides of the heatspreader, ist doent not matter at all, as long as the upper surface is clean when applying new grease.

I guess this is for those folks who RGB-illuminate their cases and put their GPUs in the window to compensate for their small p... -I mean power supplies!

Jesus christ dude, projecting much! If someone wants to have a clean CPU, let them have a clean CPU, wtf is the big deal. God this site is full of negative egotistical dweebs.

 

[Hofnaerrchen]

If you need a guard when applying TIM, 99% you do it the wrong way.

I would agree it use as caution when using Liquid metal... I don't know how from a swab can something drip off... but anyways.

Actually this is a problem even professionals (Igor from igorslab for example) recognized already. This went as far as applying insulating tape to prevent thermal paste getting into the gaps. A company making profit out of such a situation is not new. Did you write something accordingly when intel had bending problems with it's new socket, because some third party companies also had a solution on hand quite fast?^^

 

[vmarv]

They have SMD's there now, and clearly had nowhere else to put em
View attachment 264386

They could have designed a regular squared IHS and cover all the surface below it, just leaving 0.5 mm of free space over the SMDs.
Even if these need some sort of cooling, that space could have been enough.

Also, according to the video of Derbauer, they made it thicker than the old ones, and this is the cause of the bad temps. It's a bad overall design.

 

[cvaldes]

They could have designed a regular squared IHS and cover all the surface below it, just leaving 0.5 mm of free space over the SMDs.
Even if these need some sort of cooling, that space could have been enough.

Also, according to the video of Derbauer, they made it thicker than the old ones, and this is the cause of the bad temps. It's a bad overall design.

It's given that AMD tried a number of IHS designs during development including the IHS design from the AM4 generation and probably some modified designs of those from the competition (like Intel products).

It's also probable that the people who gave the final approval for this design have a couple hundred years of semiconductor design experience combined, with advanced degrees from some of the world's best engineering universities. The AM5 IHS design wasn't from some 22-year old snot-nosed intern from Foothill Junior College.

It is possible that it's a bad design. That would bode ominously for AMD's future as a semiconductor company though.

 

[SLObinger]

The thickness of the IHS is not why the CPU runs at 95c. The chip is just hot by design. Any CPU with direct die water cooling runs much cooler vs a chip with an IHS. If Zen 4 had a full square IHS that was half the thickness it wouldn't have significantly cooler temps. Zen 4 increased IHS thickness plus the larger backplates on X670 motherboards are likely to protect the 5nm CCDs pressure or torsion related damage

 

[cvaldes]

The thickness of the IHS is not why the CPU runs at 95c. The chip is just hot by design.

This brings up a point that some people here just don't understand. AMD could have programmed this chip to run at 90, 85, 80 °C, etc.

They decided to let it run at 95 °C. And they probably tested various thermal ceilings in the lab.

It's worth pointing out that a lot of mobile CPU designs also run around this temperature. I have an Intel Core i7 in my Mac mini 2018 and during Handbrake encodes it runs around 95 °C on average. This is nothing new.

 

[SLObinger]

This brings up a point that some people here just don't understand. AMD could have programmed this chip to run at 90, 85, 80 °C, etc.

They decided to let it run at 95 °C. And they probably tested various thermal ceilings in the lab.

It's worth pointing out that a lot of mobile CPU designs also run around this temperature. I have an Intel Core i7 in my Mac mini 2018 and during Handbrake encodes it runs around 95 °C on average. This is nothing new.

Yup. I have a Dell XPS with an i9-12900HK and it runs between 90-100 C at load all day long

 

[Veseleil]

This IHS design is a compromise that shouldn't happened. They've (AMD) earned so much consumer points for compatibility already, might've tried something new for a change. I mean, it's a completely new platform, all parts (CPU, MB, RAM) are already way too much expensive, so why the hell did they bogged down the thermals just for the sake of a damn compatibility? Someone has the money for the AM5 upgrade, but will cheap out on a CPU cooler by reusing the old one? WTH AMD!?

 

[vmarv]

AMD may have designed the chip to run hot, and they may be cool with that.
But the chip passed from 90°C to 70°C after being delidded, using an AIO and the stock frequencies and voltage. This mean that they could have done better. One doesn't need to be an engineer to understands the benefits of having that chip running that cool at stocks, instead of 95°C. And this difference in temps is caused by the IHS.

 

[trsttte]

If Zen 4 had a full square IHS that was half the thickness it wouldn't have significantly cooler temps. Zen 4 increased IHS thickness plus the larger backplates on X670 motherboards are likely to protect the 5nm CCDs pressure or torsion related damage

Pressure or torsion related damage has nothing to do with it, like nothing at all, that's what the extra layers and finishing on top of the silicon is there for. There's not really a difference in pressure/torsion risk between a 5nm or a 7nm, they both are packaged with similar processes after the lithography is done. A 5nm chip is not more fragile than a 7nm.

This brings up a point that some people here just don't understand. AMD could have programmed this chip to run at 90, 85, 80 °C, etc.

Yes, that's true, but a thicker IHS will still present a higher thermal resistance than a smaller one. Higher thermal resistence, less heat transfer, higher temp for the same power (the -20ºc or whatever result from delided also shows this, if that was even a question).

They could have set a smaller or higher max temp regardless of the ihs thickness, but they could also get a lower stable temperature for the max power they set with a smaller ihs. One doesn't contradict the other.

Someone has the money for the AM5 upgrade, but will cheap out on a CPU cooler by reusing the old one?

It's not even like mount adapter brackets aren't a thing, even Intel with a very different footprint is using them (though keeping the same hole pattern is also an option), I don't think this was a necessary compromise. Oh well, too late now

 

[cvaldes]

Again, I'm convinced that AMD engineers had enough foresight to test IHS designs in a variety of thicknesses and decided this was the best compromise in terms of physical protection, thermal performance, manufacturing cost, durability, etc.

Unlike a lot of people here, I don't feel like I really should be playing armchair EE Ph.D. My guess is that the average age of the senior engineers involved in semiconductor thermal solutions at AMD have an average age of 50 and probably have been doing this for 20-25+ years.

I have a lovely liberal arts degree. I generally leave the engineering to people who went to school for it and have practiced in the field for decades.

 

[vmarv]

Again, I'm convinced that AMD engineers had enough foresight to test IHS designs in a variety of thicknesses and decided this was the best compromise in terms of physical protection, thermal performance, manufacturing cost, durability, etc.
Unlike a lot of people here, I don't feel like I really should be playing armchair EE Ph.D. My guess is that the average age of the senior engineers involved in semiconductor thermal solutions at AMD have an average age of 50 and probably have been doing this for 20-25+ years.
I have a lovely liberal arts degree.

I guess you missed some videos recently. They made it thicker to allow the customers to reuse their AM4 coolers. It was a compromise, not the best possible option.

 

[trsttte]

Again, I'm convinced that AMD engineers had enough foresight to test IHS designs in a variety of thicknesses and decided this was the best compromise in terms of physical protection, thermal performance, manufacturing cost, durability, etc.

Unlike a lot of people here, I don't feel like I really should be playing armchair EE Ph.D. My guess is that the average age of the senior engineers involved in semiconductor thermal solutions at AMD have an average age of 50 and probably have been doing this for 20-25+ years.

I have a lovely liberal arts degree. I generally leave the engineering to people who went to school for it and have practiced in the field for decades.

The strategic decision to keep cooler compatibility is not really an engineering decision though, it's a business decision that the engineers had to work with.

Does it work? yeah sure, i'm not questioning that, but could it work better? Also yes.

 

[[XC] Oj101]

Dunning Kruger in full effect...

 

[cvaldes]

The strategic decision to keep cooler compatibility is not really an engineering decision though, it's a business decision that the engineers had to work with.

Does it work? yeah sure, i'm not questioning that, but could it work better? Also yes.

That's a larger question. What is the value of backwards compatibility versus a clean new design?

In any case, the business makes the money that pays the engineers' salaries. This isn't just some college dream project that has no existing business behind it.

There are business parameters that are just as important as some engineering decisions.

If the engineers don't like it, they can go elsewhere and start their own company. That's how Silicon Valley came to be. Shockley > Fairchild > Intel > others. Absolutely nothing wrong with that, in fact that entrepreneurial spirit is highly prized in Silicon Valley.

However, as long as these engineers have AMD employee badges, AMD management calls the shots. Dr. Lisa Su and the rest of her senior management team has increased shareholder value over her tenure so I doubt if the board will fire her anytime soon.

California is an at-will employment state. Anyone of those AMD engineers is free to "pursue other opportunities."

 

[onemanhitsquad]

they did this , they should have done that , it would be better if , ...then just don't buy it.

 

[Minus Infinity]

It was very predictable - thicker metal higher thermal resistance, pretty simple. They could have kept the layout and socket keep out areas but change the height (conversion kit would be new mounting screws or just washers for crafty users ) and the thermals would be improved a lot just as with deliding, maybe now they have an excuse to change sockets next year

I don't think there would be any outrage, pretty natural to need at least a mounting adapter with a new socket, maybe they feared it would take longer for manufacturers to adopt the new socket but they're no longer the underdog pseudo-newcommer they were back at the start of AM4, this was just a poor decision

Why they chose a poor thermal alloy given the extra thickness of the IHS is pretty weird. Is it penny pinching, or they really don't care that the cpu runs at 95C and can't really achieve the promised boost clocks for long.

 

[A Computer Guy]

This IHS design is a compromise that shouldn't happened. They've (AMD) earned so much consumer points for compatibility already, might've tried something new for a change. I mean, it's a completely new platform, all parts (CPU, MB, RAM) are already way too much expensive, so why the hell did they bogged down the thermals just for the sake of a damn compatibility? Someone has the money for the AM5 upgrade, but will cheap out on a CPU cooler by reusing the old one? WTH AMD!?

I think the bigger problem is NOT all the AM4 coolers will be compatible especially those that used their own backplate design. It will be interesting to see a list of what coolers (and their relative price point) are compatible vs. not compatible to determine if retaining some semblance of cooler compatibility was a good choice.

 

[cvaldes]

I think the bigger problem is NOT all the AM4 coolers will be compatible especially those that used their own backplate design. It will be interesting to see a list of what coolers (and their relative price point) are compatible vs. not compatible to determine if retaining some semblance of cooler compatibility was a good choice.

I believe that's why there are compatibility kits for many of those coolers that aren't compatible.

But if I understand correctly, AMD made a deliberate business decision to maintain certain physical dimensions to minimize the number of thermal solutions that would be incompatible.

To my knowledge AMD never promised the rest of the industry 100% backwards compatibility between AM4 and AM5.

Hell, Intel sure doesn't.

 

[A Computer Guy]

I believe that's why there are compatibility kits for many of those coolers that aren't compatible.

But if I understand correctly, AMD made a deliberate business decision to maintain certain physical dimensions to minimize the number of thermal solutions that would be incompatible.

I think that is a reasonable business decision especially if the IHS thickness doesn't impact actual cooling capability. I am patiently waiting for someone to release some test results with IHS vs. Direct Die to see if you can actually boost performance under direct die. If you take something like buildzoids chart (

) will direct die yield any gains over 200W? The new CPU's are designed to run hot at 95c so I suspect not but I am not a CPU engineer.

 

[crazyeyesreaper]

If i use AM5 for cooler testing anytime soon I will definitely want one of these. Just to make my life easier.

 

[SL2]

But if I understand correctly, AMD made a deliberate business decision to maintain certain physical dimensions to minimize the number of thermal solutions that would be incompatible.

I don't believe that for a second.

Why does the lid have to be that thick, if it affects thermal performance in such a bad way? AMD says Ryzen 7000 will hit a temp limit before a power limit, so why make it extra hot when you need every MHz?

Why didn't AMD just make the socket itself ~1.5 mm higher instead? (Or whatever height we're talking about.)


No, there's something else going on here. My bet is still preparations for die stacking of some kind in future products, something that uses up more space/height than current 3D cache. (The added cache in 5800X3D are combined with extra thin CCD's as a height compensation, maybe that's something they want to avoid, although I guess the new lid is even thicker than that..)
That's the only reason for why AMD must use a thick lid that I can come up with.

(Or, maybe it's not that easy to design an LGA socket in plastic that has added height that can withstand the mounting force while maintaining proper contact with the CPU.)

 

[pressing on]

I don't believe that for a second.

Why does the lid have to be that thick, if it affects thermal performance in such a bad way? AMD says Ryzen 7000 will hit a temp limit before a power limit, so why make it extra hot when you need every MHz?

Why didn't AMD just make the socket itself ~1.5 mm higher instead? (Or whatever height we're talking about.)


No, there's something else going on here. My bet is still preparations for die stacking of some kind in future products, something that uses up more space/height than current 3D cache. (The added cache in 5800X3D are combined with extra thin CCD's as a height compensation, maybe that's something they want to avoid, although I guess the new lid is even thicker than that..)
That's the only reason for why AMD must use a thick lid that I can come up with.

(Or, maybe it's not that easy to design an LGA socket in plastic that has added height that can withstand the mounting force while maintaining proper contact with the CPU.)

During the Zen 4 launch there was a slide showing that AM4 cooler compatibility was a key feature of Ryzen 7000 processors (see the slide below) so it was a deliberate choice. Making the IHS thicker was part of that choice because due to the AM5 LGA socket the CPU sits slightly lower than it did with AM4. If the plastic pillars that form part of the standard hook type fitting had been reduced in height to allow a thinner IHS there would have been a risk that AM4 coolers would not have enough tolerance to fit correctly. An adapter would have solved this issue but that simply solves one problem by replacing it with another.

For example, the need for an LGA 1700 adapter for previous coolers is a feature of Intel's 12th and 13th Gen processors (Alder Lake and Raptor Lake). Some cooler manufacturers will supply them free on proof of purchase of one of these CPUs. LGA 1700 adapters are on sale at most retailers and obviously users are buying an adapter rather than waiting for one, or because they hadn't realised one was needed.

All of this inconvenience and expense has been avoided by AMD for the vast majority of AM4 coolers. Not all, because the AM5 ILM is screwed into the backplate so AM4 coolers that rely on removing/replacing the backplate will have an issue but the number of coolers affected is small and most of the manufactures concerned are already doing something about this.

 

[Mussels]

So what? Let it fill the space in between. Why bother at all? All my CPUs have residue on the sides of the heatspreader, ist doent not matter at all, as long as the upper surface is clean when applying new grease.

I guess this is for those folks who RGB-illuminate their cases and put their GPUs in the window to compensate for their small p... -I mean power supplies!

No it wont matter at all - but if you care about it for OCD reasons or whatever, you can now get a little bracket.
Heck, that bracket might be really handy if you're lapping the CPU to hold it nice and neatly with, and give you a perfect height to lap it down to.


It's optional and cheap, which is how it should be.

Just spend some money to slim the card

lap the credit card

 

[Ferrum Master]

Actually this is a problem even professionals (Igor from igorslab for example)

Let's get us one thing straight. Who is a professional or not. A professional internet diva, tech journalist, whos earning money with this? Or or a engineer, a manufacturer official and dev sitting behind the desk and don't giving a F* about anything and just making his stuff?

No internet tech diva is really a pro. Just give the man the right tools, instruments and review items, you can make anyone a diva or Jesus with a freaking opinion. That counts also for Igors lab, whose articles often caused my eyebrows rise, also the toxic competition between der8auer that ends up as a circus.