PlayStation 4 (codename Orbis) technical hardware investigation (news and rumours)

The motherboard only shows 2 prongs for the PSU connection so I would assume it only has a 12v feed to the motherboard.

Regulator doesn't have to be strong for 5v supply to the SATA, USB and other parts as they consume very little power compared to the GDDR5 and APU.

 

I think you need to have GDDR5 on each side of the board to make use of the clamshell mode anyway.

I'm surprised at how clean the whole thing is. Even tearing it down was very straight forward, comparing it to the fat PS3 where disassembling was a serious task.

 

I mean strong enough that the 5V supply need a regulator large enough to be noticeable on the mother board.

USB3 is 900ma x3
HDD 1 amp
Bluray 1 amp

Up to 25 watts of 5V ?

I thought the Fat was only 12V, but no it had a separate 5 pin header connector for the 5V, it could be one of the two white connectors there.

 

Good point...which explains why the PS4 board is double sided and why the X1 board may not be.

 

You're looking at 3w max for the HDD, some portable Blu-Ray drives only use 5w when reading a Blu-Ray disk...

I would say with modern manufacturing techniques you could easily do the whole I/O section of the machine at 10-15w

 

What is the chip next to the PlayStation sign?

 

So when, if ever, do we find out the final clock rates? Is that even going to be possible to test/discover?

 

I think I can see SCEI on it... so another custom chip?
It doesn't seem to be connected to much other than the SoC. Maybe security?

 

Isn't that the custom ARM based processor for background downloading and low power tasks? It looks very much like the Vita SoC...

Spoiler

I jest of course!!

 

Putting two chips on the same command bus increases noise. The same chips are typically rated to work at 6Gbps at 32-bit configurations and 5.5Gbps at 16-bit (clamshell) configurations. PS4 aimed for 192GB/s when it was supposed to have 4GB, doubling the memory meant they had to drop the speed.

 

The backside of the PS4 PCB is almost empty besides the RAM chips.



It doesn't explain the large difference to the sheer number of components on the XB1 board. Let's see how many of them could be eliminated in the retail boxes.

 

We were shown a retail PS4 board... The One board we've been shown was not final or retail..

 

Note what looks like one lone thermal pad on one of the GDDRs, which means the RF shield doubles as a heatspreader.

The Xbone mobo we have been shown has those diagnostic number segment LEDs and whatnot on it - there could be a fair amount of space freed up in a final retail product.

 

Upon closer inspection the PS4 board and X1 board seem to have similar number of components. The X1 board's white labeling for components is making it appear much more complex even though it isn't. Also notice the backside of the PS4 board has quite a number of tiny surface mount parts.

 

As I explicitly mentioned, when I posted the comparison pictures (and again underneath the picture of the backside of the PS4 PCB you quoted).

 

I think it's going to be a wash when we see the production X1 board but it's an interesting discussion...

 

I don't think so. And there are good reasons to believe the retail XB1 board will lose some of the overhead it has.

You should expect roughly the same amount of small capacitors underneath the XB1 APU and other slightly larger chips on the backside of the XB1 board. There really isn't much on the PS4 board.

 

Well I was of course ignoring the diagnostic 7 segment LEDs on the X1 board.

 

Yes it's like a huge heat spreader, and it has holes and an embossed structure which seems to allow some air flow.

 

The hardware design looks elegant and the little CG with parts coming together to form a PS4 looks great.

They should include a spare playstation screw with every PS4!