Tuesday, September 1st 2020

NVIDIA GeForce RTX 3090 Ampere Bare PCB Pictured

Here are some of the clearest pictures of an NVIDIA GeForce RTX 3090 "Ampere" Founders Edition PCB, albeit bare (components not yet placed). The PCB goes in this form to another production line, where a PCB placer machine with reels of components places them along both sides. We can still make out quite a few things. To beign with, there are six high current traces near where the 12-pin Molex MicroFit power connector goes in.

NVIDIA seems to have utilized both sides of the PCB effectively. The card features traces for a mammoth 20-phase VRM. The obverse side has the chokes and DrMOS, the reverse side has the capacitors. Twelve memory pads for the 384-bit wide GDDR6X memory interface surround the GPU pad. There are pads for four display output connectors, which could be a mixture of DisplayPort, HDMI, or even USB-C VirtualLink. The PCB still has its spacers on, which will be cut out after component placement, after which the PCB heads for testing and product assembly. NVIDIA is expected to announce the GeForce RTX 3090 Ampere at an online event later today.
Sources: ChipHell Forums, VideoCardz
Add your own comment

36 Comments on NVIDIA GeForce RTX 3090 Ampere Bare PCB Pictured

#1
Dux
Very short PCB. That gigantic cooler makes the card long.
Posted on Reply
#2
maxitaxi96
well finally we have undeniable proof that this taversal coprocessor stuff was nonsense...
also the amount of phases the RTX 3090 would have was allready calculated by Igor from igorslab.de in June (!)

only things missing now are the exact structure of the cooler and perfomance numbers. Gimme!
Posted on Reply
#3
BoboOOZ
DuxCroVery short PCB. That gigantic cooler makes the card long.
Very nice design. Before, you always got huge empty pcbs for big cooler cards, this way the pcb is only the size it needs to be, letting the radiator breathe better and do a better job. I don't know how big is the added complexity of manufacturing this, but it's definitely a step in the good direction.
I mean, GPU TDPs are goind up anyway, it would be better if the cooling design allowed for better cooling capacity in the future. CPU coolers have it way easier.
Posted on Reply
#4
Lionheart
I wonder if there will be a RTX 3090 Ti
Posted on Reply
#5
BoboOOZ
LionheartI wonder if there will be a RTX 3090 Ti
You mean wearing that exact name? That's not sure, it seems Nvidia are trying to clean up their naming scheme a bit. But if you mean a stronger Ampere card than the 3090, there's a high chance for that.
Posted on Reply
#6
mobiuus
i think 3090 is titan but maybe im wrong but nevertheless i think there will be 3080 ti to fill the space in shader units between 3080 and 3090
Posted on Reply
#7
MrAMD
DarkStalkeri think 3090 is titan but maybe im wrong but nevertheless i think there will be 3080 ti to fill the space in shader units between 3080 and 3090
I think there's a 48GB full-fat GA102 Titan coming later
Posted on Reply
#8
ebivan
BoboOOZYou mean wearing that exact name? That's not sure, it seems Nvidia are trying to clean up their naming scheme a bit. But if you mean a stronger Ampere card than the 3090, there's a high chance for that.
Of course there will be. For all the enthusiasts that spent four digits for a 3090 now and don't care for spending another 2k a couple of months later just to keep having the biggest... I mean fastest GPU.

Anyways, I like the size of that PCB. Is it true, that all tiers will use the same PCB like with the 16xx cards? I will probably get a 3070 or maybe the AMD equivalent after RDNA2 is released. With watercooling attached, something this small could be a great little addition to my ITX system.
Posted on Reply
#9
ShurikN
Ampere, worst kept secret ever.
Posted on Reply
#10
ebivan
ShurikNAmpere, worst kept secret ever.
Its called marketing. Do you really think that most of these leaks are involuntary? Its all part of the release process.
Posted on Reply
#11
Dammeron
DuxCroVery short PCB. That gigantic cooler makes the card long.
There's still a possibility of a separate PCB containing some of the stuff (I don't see any mounts for power supply in the pics above), there's also this:


Until I see the whole card I'll wait with my opinions about how small the card is. :)
Posted on Reply
#12
steen
btarunrTo beign with, there are six high current traces near where the 12-pin Molex MicroFit power connector goes in.
The 12-pin microfit connector area is indeed interesting.
There are pads for four display output connectors, which could be a mixture of DisplayPort, HDMI, or even USB-C VirtualLink.
They are double stacked pads.
Posted on Reply
#13
Sihastru
What is interesting is that the 12-pin PCIe Molex connector capable of 600+ gazillion watts only has 4 receptacle through holes on the PCB (2 ground, 2 +12V).
Posted on Reply
#14
medi01
ShurikNAmpere, worst kept secret ever.
You still know very little about how it performs, hours before its release.
Posted on Reply
#15
kiriakost
Hmm about 20 VRM's stacked closely to its other.
Next thing we need to know this is : How many layers and thickness of the PCB.

A large number of VRM's this translates that the GPU has the ability to turn off engines that the software can not use them.
Its like an active power saving mechanism.
Posted on Reply
#17
TheLostSwede
News Editor
SihastruWhat is interesting is that the 12-pin PCIe Molex connector capable of 600+ gazillion watts only has 4 receptacle through holes on the PCB (2 ground, 2 +12V).
I count six, which I would guess is four for +12V and two for ground.
steenThey are double stacked pads.
Uhm, only one allows for either HDMI or DP, the other three are DP, although there is something going on with the top most port, as there's space for a chip underneath the connector, which is highly unusual.
However, none of them are double stacked, if you compare with this. It also makes it clear that the 3090 doesn't have a USB-C port for VirtualLink.



@btarunr the PCB has space for 24 memory ICs, not 12, as there are 12 pads on each side.

It's interesting to see a graphics card with more power phases than most high-end motherboards. That must be one power hungry chip.
kiriakostNext thing we need to know this is : How many layers and thickness of the PCB.
Sadly it's a bit too blurry to see, but it seems to be a 12 layer PCB if you look at the rear picture of the PCB and up top there's a yellow-ish square where I'm guessing the PCB layer number is present and it looks like a 12 to me.
Posted on Reply
#18
EarthDog
kiriakostA large number of VRM's this translates that the GPU has the ability to turn off engines that the software can not use them.
Its like an active power saving mechanism.
?????

No.
Posted on Reply
#19
iO
4x DP with the option of swapping the bottom one for a single HDMI port.

If they can cram the GA102 on ~205mm, a sub 170mm card should be trivial for GA104...
Posted on Reply
#20
Sihastru
TheLostSwedeI count six, which I would guess is four for +12V and two for ground.
The two smaller ones are reinforcement pins for a partial metal plate.
Posted on Reply
#21
Dredi
DammeronThere's still a possibility of a separate PCB containing some of the stuff (I don't see any mounts for power supply in the pics above), there's also this:


Until I see the whole card I'll wait with my opinions about how small the card is. :)
The picture is just some fanfiction. What would possibly be in the additional pcb anyway, when you can clearly see that everything necessary is already in the news articles pcb picture.
SihastruWhat is interesting is that the 12-pin PCIe Molex connector capable of 600+ gazillion watts only has 4 receptacle through holes on the PCB (2 ground, 2 +12V).
300 Watts before it will start to heat up in an unhealthy manner. The PCB mounting points can carry that no problem, the thin wires and connection points inside the plug & receptacle are the problem areas. It’ll be interesting to see what figure nvidia will tout in the evenings presentation.
Posted on Reply
#22
ShurikN
medi01You still know very little about how it performs, hours before its release.
The hype machinery has to start somewhere. What's interesting about all of it is, I can't remember when has this last happened with an Nvidia release. The amount of" leaks" rivals some of the biggest Amd's hype trains.
Which is weird because there's not much known about RDNA2, and I'm guessing that card is launching very soon as well.
Posted on Reply
#23
TiN
Layout routing of this board must be real fun :). Placing of big VRM so close to sensitive high-speed traces is SI nightmare :)
Posted on Reply
#24
medi01
ShurikNWhich is weird because there's not much known about RDNA2, and I'm guessing that card is launching very soon as well.
Sometime in November, according to that weird twitter dude that was spot on on number of CUs in 3xxx lineup.
Posted on Reply
#25
Gungar
Short PCB means HBM, that card is going to be expensive!
Posted on Reply
Add your own comment
Nov 21st, 2024 12:25 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts