Guys,
Have anyone seen or tested more than 12 layers in 5G NSA?
As per my information we can max go with 32x32 layers MIMO.
I have seen 16 (4 MU-MIMO users with 4x4 each).
Huawei and Ericsson have 16 pdsch layers
32 is the max number of CSI-RS ports, not layers.
There is higher chance to see 8 users each with 2 layers than 4 users each with 4 layers.
Number of layers can be equal or less that number of CSI-RS port.
Isn’t this concept correct?
If gNB supports 8 Beams.
Trials i was involved in were for 4 Beams and 4 layers each.
PMI matrix is PxL where P is CSI-RS ports and L is layers.
They are not related to each other.
Number of csi-ports gives how narrow is the beam.
Ok.
So what is max layer therotically possible?
As there is space multiplexing I don’t think there is a limit other than implementation limit.
Limit is pairing users for mu-MIMO.
Ok.
I knew that in FR1, there can be max 12x12 MIMO.
Which I think is not limitation.
You can find a report of Ericsson demonstrating 16 PDSCH layers.
Read it in Signals Flash 16:
4 beams equal 8 layers.
Remember 1 beam equal 2 layers.
With FR1 we can use 8 SSB beams, but MIMO mode I seen as of now max 4x2 option only.
It’s difficult to get rank4 for an UE, this is why more common is 8 UEs with 2 layers each.
How did you concluded that?
I mean max 4x2?
1 beam can be 4x4.
4x2 was in LTE also.
1 SSB beam can’t we require 2 SSB beam for 4x4.
It’s 4x2 or 4x4 I mean.
With 4x4 MIMO mode and 2 SSB beam I seen rank 4, but very difficult to achieve.