Playstation 5 [PS5] [Release November 12 2020]

[Barrabas]

Wow, impressive!!, better than watching porn. This has to be silent. That crazy MF heatsink and the rumored liquid metal was true. One thing I noted at the ending text is "The dust catcher does not guarantee the hardware clogging from dust". Now the only thing I regret is that I did not order two so that one could be dedicated next to me in my bed

 

[anexanhume]

For an interesting comparison.
5700xt is 251 mm² also at 7nm
so the CPU is taking up about 35mm^2

Doesn’t work like that since area sharing is different compared to a stand alone GPU. Note the L3 is only 8MB, too.

 

[cheapchips]

From the PS blog entry on the teardown:

We began conceptualizing PS5 in 2015, and we’ve spent the past five years designing and developing the console.

https://blog.playstation.com/2020/1...k-at-our-most-transformative-console-yet/amp/

Five years. Wonder how real and how much effort went the 2019 release version?

 

[function]

For an interesting comparison.
5700xt is 251 mm² also at 7nm
so the CPU is taking up about 35mm^2

So it's looking like the same 8 MB of L3 for the CPU that the XSX and Renoir APUs have. Makes sense, as that's the design that AMD has and on an APU the silicon is better spent on CUs (or simply using a smaller and less costly die).

360 mm2 is the die size of XSX vs 286 mm2 the die size of PS5... Only a 25%. I thought it was much more. Can be they are realized on different silicon processes ?

Well, the GPU on XSX is 40% bigger (assuming basically the same architecture), but the memory bus is only 25% larger and the CPU is probably the same size. So about 25% larger for XSX seems about right.

Given how tied to a particular process architectures are I think it's basically a give that PS5, XSX and PC RDNA2 (and probably Renoir) are all on the same process.

 

[iroboto]

Doesn’t work like that since area sharing is different compared to a stand alone GPU. Note the L3 is only 8MB, too.

Yea that's true, I honestly expect the CPU to be a tad bit larger. I was expecting the PS5 to hit 300mm2, 286 sounds slim.

 

[Unknown Soldier]

I saw the sarcasm but wanted to comment on the value cost. If it was all copper that heatsink would probably be 2+kg. I was talking about the amount they did use.

And yet my reply was for the comment about the heatsink and that it had a lot of copper, which it doesn't really if you have a look at the picture. The copper seems to be minimalistic imo and only used where needed.

 

[HBRU]

I think we need another APU size measurement. Because 286mm² seems small. Some people were also wrong when they tried to measure the XSX die like this.

Others says 303

 

[Lalaland]

And yet my reply was for the comment about the heatsink and that it had a lot of copper, which it doesn't really if you have a look at the picture. The copper seems to be minimalistic imo and only used where needed.

Ryzen Stock cooler (stolen from https://lanoc.org/review/cooling/7630-ryzen-itx-cooler-roundup?showall=1 )

It's a lot of copper truly, look for example at this stock Ryzen cooler, the Cu element is a "penny" contact area inserted into what is otherwise an entirely Al block, that is a minimal use of copper. What Sony have done is here is build the entire base of this heatsink out of Cu and connected the Al fins to it directly. Any use of copper is going to raise questions from production engineering, which is why we get contact pennies like the above, for them to sign off on fairly large area sheet of it suggest there are good engineering reasons for it. If there weren't I would expect an Al sheet with contact pennies for the hot elements

 

[function]

second edit: I no longer think it's actively cooled, or at least there is no dedicated channel. The air will come in however it can from the fan, but no dedicated channel, or at least the channel isn't as wide as the SSD.
At first found some evidence on the plastic black case that it would channel. I mean that arrow made sense. That looks like a hole there that would channel air over the SSD. But upon further review, you can actually see the reflection of the screw there against the plastic. And I didn't see that the first time. So I think that black area is blocked off. It's only going to get air from the sides... which is very much blocked off by the metal. Honestly it's really hard to tell, someone might need to do some photoshop colour inversions to see what's there.

You can see for yourself. the Video should be setup at the right time stamp, but the reflection of the screw is visible to me. That isn't dust imo, and you can see that plastic blocker when he removes the plastic shell. The metal casing is revealed underneath that, and there are a couple areas where air can filter through from the corner and sides.

There appear to be two dedicated grooves for air to flow through that won't be blocked by the metal top plate.



I'd also guess that there will be some air flow between the PCIe slot and the case around it, so some air should move over the top part of the SSD too.

So probably not masses of air flow, but they've definitely made provision for some.

 

[Love_In_Rio]

The complexity in applying and isolating it was one of the downsides which I saw noted for a use case like this.
I'm curious about the comment about getting vapor chamber performance out of heat pipes, since it's not like a vapor chamber isn't successfully used for this class of chip.

Here the priority is not only keep the temperatures for the main SOC. I think it is a high priority in the design to keep the SSD modules cold plus the added M.2 SSD.

 

[rabidrabbit]

I'm away from internet for a few hours and during that everything happens

 

[eastmen]

That black plastic is likely where the SSD will slot into. You see the black veins on the right? Most likely where the air can go out.

if you look at the post i quoted you can see the other side. It looks like there is no opening and its in its own slot

 

[Deleted member 11852]

liquid metal confirmed and 120 mm fan

PS5 is ludicrously over-engineered and I don't meant that in a bad way. The 4K drive is bonkers chunky, the heatsink is frankly absurd, the fan is insane and the volume of liquid metal is just crazy.

I like it. Those reports from over the weekend about it running cool and quiet, now we know why. I bet whoever designed that stand is very pleased with themselves.

 

[eastmen]

really shows the contrasting design

 

[Ika]

Looks like a decent quality build, the wide fan blades probably push a lot of air, and that big sink must be very expensive.
Some of the questions I have after the first watch:
- No dust filters? Those whitish grills might block too much air, but they won't stop much dust imo.
- Somehow I just can't believe that the position of the fan is optimal for the heatsink position. Looks like some fins will get very little fresh air, and I mean some of the fins around heat pipes.
- Liquid metal is only 3-5C cooler compared to a good thermal paste and usualy not worth the trouble considering how badly it can stain copper over time, but it doesn't touch the alimunium parts at least (because it would eat into those very quickly). I can only think of one reason to use it, and that's longevity.
- perhaps the RAM could be a little closer to the SOC to reduce signal degradation and voltage drop (based of the distance to tha VRM).
- Wonder if the PSU has a dedicated fan or not, because those small fans use to get very loud when they spin up, but if there is no dedicated fan, then the PSU will get air through the heat sink?

 

[DuckThor Evil]

There appear to be two dedicated grooves for air to flow through that won't be blocked by the metal top plate.

Those do look blocked based on other pictures though, in any case I think the overall design is very nice and appears to be very performant.

 

[Unknown Soldier]

I'm away from internet for a few hours and during that everything happens

The same thing happened with me. Really nice surprise to be honest.

 

[Unknown Soldier]

if you look at the post i quoted you can see the other side. It looks like there is no opening and its in its own slot

Yep, already noticed that when I saw the top pic. However function did provide route for air to escape the enclosure.

 

[Globalisateur]

Others says 303

Yes I have seen 293 mm² elsewhere.

 

[function]

Those do look blocked based on other pictures though, in any case I think the overall design is very nice and appears to be very performant.

Coming from this angle, I think you can see that while the area for the top plate is slightly recessed, the two grooves I mentioned go lower and that they'll remain clear even with everything in place. I also think there'll be at least some gaps around the PCIe slot.

The second groove further away is even deeper than the first, presumably to account for the fact that some air will already have exhausted and the pressure will be less.

Actually, I think there might be a gap between the metal top plate and the plastic case above the PCIe slot too.

There's definitely a great deal of thought been put into the PS5. You have to respect their focus on cooling this time round.