N210 MIMO

[kaosun]

 Does anyone succeed in establishing a 2x2 MIMO by using four N210?  I've checked the link: "2X2 MIMO With NI USRP" at http://www.ni.com/white-paper/13878/en . The article said they used 4 USRP to establish a MIMO system but didn't mention the model of the USRP. My lab bought 4 Ettus Research N210 for MIMO experiment, not the N2920 used in the video. I downloaded the VI from NI's website with all toolkits installed. In my test, the MIMO has never successed. The error information always says that "Invalid input symbol array; Input symbol array is empty.", or once a while very weak received signal with disordered constellation. I tried coxial connection and wireless connection, and confirm there's no error on both my hardware and software configuration. My question is : whether N210 can establish a MIMO system to test Alamouti?

Also, I noticed the USRP used in the Video is N2920, then I found N210 and N292X's FW and FPGA file in "NI USRP Configuration" and I updated the N210 by N292X's FW and Fpga files. Tested no improve.

Does anyone can help me? Thanks in advance!!!

[ErikL]

There are several MIMO examples of varrying complexity.

Using an N210 should work well.

First please review the following white paper for an overview of how to perform MIMO with USRP.

http://www.ni.com/white-paper/14311/en/

2x2 can be done with the MIMO cable.  Higher order mimo can be done by distributing clocks to mulltiple devices.

Secondly, Try the following VIs:

2x2 MIMO QAM with Alamouti Coding

or

6x6 MIMO-OFDM System with NI USRP

[kaosun]

Thank you Erik for your quick reply!

The 2x2 solution you mentioned is just as same as I discussed in my question, the VI doesn't work with my N210.

The 6x6 solution bring me into another nightmare with it's complex antenna configuration. Still studying...

[ErikL]

The errors: "Invalid input symbol array; Input symbol array is empty."

Can you describe your setup?  Do you have 2 TX USRP's connected by MIMO cable and 2 RX USRP's connected by MIMO cable?

If memory serves me correctly that VI is expecting the USRP's to be locked together in frequency using an external clock.  If that is the case you may be missing the tx signal with the RX devices because they are not sharing the time base.

the 6x6 mimo on the other hands is actually made to operate without sharing clock between TX and RX (more like a real world system)  Another great feature of that code is that you can run the 6x6 code in 2x2 mode.

Erik

[kaosun]

For 2x2 MIMO, I have two branches connected with Netgear 5 port Gigabytes switch, also connects my desktop. For the TX branch, one N210 connects directly to switch and the other one connects this one by MIMO cable. Same with RX branch. For the software setting, the N210 which connects switch directly in Tx branch, also means 0 in Per Device, uses internal ref frequency and internal timebase, the MIMO (1 in Per Device) one uses MIMO ref frequency and MIMO timebase. Same as the RX branch. 

Doesn't work still.....

[kaosun]

Do you mean even we use 2x2 MIMO, we have to connect a external "ref in" to 4 N210s? I didn't find such connection in the tutorial video at all at: http://www.ni.com/white-paper/13878/en

Anyone can help me?!

[James-K]

I believe Erik is correct about having the clocks synched on frequency and phase.  If you read section 3 of the white paper (http://www.ni.com/white-paper/14311/en/), this is explained in a little more detail.

[kaosun]

I tried connecting one signal generator to get PPS signal and another signal genarator to get 10M signal. Both are sent into the primary Tx and Rx transceivers. The slaves are using MIMO. But, failed. I just wondering whether the two signal genarators need sychronization too? Thank you!

[James-K]

Are you still receiving the original error or a different error?  How have you made the connections to each of the USRPs?

[kaosun]

The errors: "Invalid input symbol array; Input symbol array is empty."