WARP Project Forums - Wireless Open-Access Research Platform

You are not logged in.

#1 2019-Sep-29 22:43:30

Fpgageek
Member
Registered: 2019-Sep-29
Posts: 3

802.11 warpv3 reference design LTS correlation

Hi,

I am working with Mango_802.11_RefDes_v1.8.0 and its wlan_phy_rx_pmd file in SysGen_Reference. In LTS_correlation system you are feeding coef2 to ROM but I cant see from where coef2 is coming. Is coef2 loaded via some matlab file?? or it is some where in the design itself, if yes it would be great if you can point to that.

One more thing the LTS correlation implentation in system gen is very complicated is it possible to get some algorithmic description from somewhere? Meaning the logic behind which can help understand the implementation flow steps in the system generation model.

Thanks and Regards

Offline

 

#2 2019-Sep-30 10:01:36

murphpo
Administrator
From: Mango Communications
Registered: 2006-Jul-03
Posts: 5104

Re: 802.11 warpv3 reference design LTS correlation

Is coef2 loaded via some matlab file?? or it is some where in the design itself, if yes it would be great if you can point to that.

The correlation coefficients are defined in the wlan_phy_rx_init.m script.

And just to clarify - are you using this design on Mango WARP v3 hardware (per the license requirements)? If not, please contact us to discuss licensing options for our newer 802.11 MAC/PHY FPGA design for third party hardware.

Offline

 

#3 2019-Oct-01 01:34:59

Fpgageek
Member
Registered: 2019-Sep-29
Posts: 3

Re: 802.11 warpv3 reference design LTS correlation

Thanks for your reply. We have used your warp boards in the past and even curently and  have duly acknowledged. We can provide citations as evidence for any concerns and also a list of our works via email. Its been a while I posted on your forum and it seems my old email account will expire due to change of server here so i created a new. We are not using your 802.11 reference design on any other board. We do have warpv3 hardware infact v2 and v1 as well. We are not currently running this design  on any board (not even v3) but only going through system generator model to understand how a real system works. Yes, we do have xilinx system generator license :). I am sure going through your system generator design and understanding it for the purpose  how a real 802.11 system works as an example is not violation of any licensing? Otherwise you should add in the license that you cannot even read or go through the the design??? I am correct right. Hope it helps and clarifies all this.

Regards

Last edited by Fpgageek (2019-Oct-01 01:45:55)

Offline

 

#4 2019-Oct-01 10:12:29

murphpo
Administrator
From: Mango Communications
Registered: 2006-Jul-03
Posts: 5104

Re: 802.11 warpv3 reference design LTS correlation

Thanks for the clarification. Using the PHY models in simulation is of course ok for WARP hardware owners. Providing a design users could extend and test in hardware was a primary goal of the original 802.11 design. Our newer 802.11 design provides the same flexibility for users targeting modern, third-party hardware, just with a different licensing model.

One more thing the LTS correlation implentation in system gen is very complicated is it possible to get some algorithmic description from somewhere? Meaning the logic behind which can help understand the implementation flow steps in the system generation model.

The LTS correlator implements a streaming 64-point complex cross correlation. The 64-point correlation is re-computed in every sample period. The logic uses a quantized coefficient vector so the atomic operations can be implemented as add/subtract instead of multiplication. The complicated control logic allows the 64-point correlation to be computed using 16 add/sub blocks which are clocked at >=4x the sample rate and are time-shared to compute 4 "products" per sample period.

Offline

 

#5 2019-Oct-01 22:58:32

Fpgageek
Member
Registered: 2019-Sep-29
Posts: 3

Re: 802.11 warpv3 reference design LTS correlation

Many thanks for your response providing more details of  LTS correlation.
Regards

Offline

 

Board footer