Martin,
I’ve been searching through the mailing list for more information
about synchronizing two USRP boards to create the transmitter for a
4x4 MIMO node. From what I’ve found, it seems that this is not
currently implemented (see links below for relevant threads). What is
the current status of this effort, and when (if at all) do you think
that this functionality might be implemented?
links:
http://www.nabble.com/Question-about-4-x-4-MIMO-transmitting-and-receiving-tf3605863.html#a10074272
http://www.nabble.com/Re%3A-USRP-Tx-Question-tf1452214.html#a3924517
–
Ketan M.
http://hydra.ece.utexas.edu
Martin,
Thanks for your response. It is disappointing to hear that the MIMO
support for 4 antennas will not be available for some time.
Could you provide some more details into why this problem of aligning
the transmit samples will be easier in the new m-block framework? Will
these new packet oriented m-blocks utilize a different interface to
the USRP which will make this alignment-task easier?
On 8/29/07, Martin D. [email protected] wrote:
I am the author of the usrp_multi code (and the modified usrp firmware)
After this project it also depends on the amount of work I have, how much time will be left for implementing 4x4MIMO.
links:
This means that every time you start a gnuradio script the phase relationship between the multiple USRPs will have a fixed unknown phase difference.
Martin DvH
Olifantasia
Software Defined Radio, GnuRadio, USRP development
–
Ketan M.
On Wed, Aug 29, 2007 at 09:40:58PM -0500, Ketan M. wrote:
Martin,
Thanks for your response. It is disappointing to hear that the MIMO
support for 4 antennas will not be available for some time.
Addressing your disappointment: This is a free software project.
Code gets written because some actual human being writes it 
If you’re waiting for somebody else to write it, you may be waiting
for a long long time…
Could you provide some more details into why this problem of aligning
the transmit samples will be easier in the new m-block framework? Will
these new packet oriented m-blocks utilize a different interface to
the USRP which will make this alignment-task easier?
There will be a new interface. More on that when it becomes available.
Each frame to be transmitted will have a timestamp that specifies the
time that it is to be transmitted. You still have to sync the sample
clocks and LO’s on the separate USRPs, but there will be a way to
communicate, in-band, when you want the frame to be transmitted.
Low-level design ideas are described here:
http://gnuradio.org/trac/browser/gnuradio/trunk/usrp/doc/inband-signaling-usb
They are subject to change…
See also:
http://gnuradio.org/trac/wiki/MessageBlocks
Eric
Ketan M. wrote:
Martin,
Thanks for your response. It is disappointing to hear that the MIMO
support for 4 antennas will not be available for some time.
Could you provide some more details into why this problem of aligning
the transmit samples will be easier in the new m-block framework? Will
these new packet oriented m-blocks utilize a different interface to
the USRP which will make this alignment-task easier?
Sorry for the late response. I get so much mail, that I miss a mail once
in a while.
With the m-block interface you can specify in a packet when the samples
should be sent.
As long as the time reference in the USRPs are synced this means that
all samples will be synchronised.
In the old framework there is no sense of absolute or relative time.
This means you have to build from scratch a system which can hold-off
samples untill the time comes for them to be sent out.
In essence this IS what the m-blocks will do. So why invent the wheel
twice?
Greetings,
Martin