Hello Marcus, Thank you for your reply.
Is there support today in software for direct register writes to the AD9361 device for the E310? If there is, I think I can use SYNC and a modified firmware to implement Multi-chip synchronization. Regards, Ofer Saferman On Sun, Jan 3, 2021 at 7:00 PM <[email protected]> wrote: > ---------- Forwarded message ---------- > From: "Marcus D. Leech" <[email protected]> > To: [email protected] > Cc: > Bcc: > Date: Sat, 02 Jan 2021 15:18:06 -0500 > Subject: Re: [USRP-users] Hacking E310 units for synchronization > On 01/02/2021 02:44 AM, Ofer Saferman via USRP-users wrote: > > Hello, > > > > I must say that I am a bit frustrated. > > When I purchased multiple E310 units it was with the hope I will be > > able to synchronize them mainly based on: > > > https://kb.ettus.com/Synchronization_and_MIMO_Capability_with_USRP_Devices, > > > it shows the E310 as MIMO capable and having a reference clock input. > > Of course, now I know that not to be true. At least the reference > > clock part. It seems that the MIMO part might be incorrect as well. > > > > If I hack the clock and patch an external clock connector to be used > > instead of the internal TCXO will I be able, using existing UHD > > software and firmware, to synchronize multiple E310 units? > > > > I will be able to provide a common clock to multiple units and I will > > be able to provide a 1-PPS signal through the SYNC pins of multiple > > units that is generated from the same common clock. > > I need these E310 to transmit and have their base-band samples > > perfectly time aligned. > > > > How is multi-chip synchronization of the AD9361 devices being done by > > UHD? By using the SYNC pin and setting > > usrp->set_time_source("external")? Is it enough? Do I need to handle > > it on my own? How? > > > > I figured the internal TCXO frequency of E310 is 40 MHz. Can I provide > > an external clock of different frequency? Maybe 10 MHz? Or does it > > require a major change in software or firmware? > > > > Does anybody have a different idea how to get E310 units synchronized > > at sample level without voiding the warranty and hacking an external > > clock? > > > > Please help. > > > > Regards, > > Ofer Saferman > > > > > The E310 was never designed for multi-unit MIMO, although it does 2x2 > MIMO just fine, and there are plenty of folks using it for that, > including stacks for various wireless protocols. The AD9361 > inherently supports 2x2 MIMO due to the shared LO architecture between > the two TX and two RX channels. > > The AD9361 SYNC_IN pin does get tied to the Zynq FPGA, but the UHD > code-base doesn't use it. I'll note that for most telecom-type > MIMO implementations, zero-phase-offset is NOT a requirement, just > ongoing phase coherence. > > Similarly the B2xx series does 2x2 MIMO just fine, but it was never > designed for multi-unit MIMO, although due to the clocking architecture, > the B2xx could likely (no guarantees) be made to do a multi-unit MIMO > implementation. > > The SYNC input connector on the E310 is designed to: > > (A) Provide a higher longer-term quality than the on-board clock. > (B) Allow for TDM synchronization on system deployments where that is > important. > > Due to the high residual phase-noise of the DPLL implementation on the > SYNC input on the E310, it cannot be used to produce a working > multi-unit MIMO (or any other kind of phase-coherent) application. > -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
_______________________________________________ USRP-users mailing list [email protected] http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
