You are not logged in.
I am a newbie to WARP. I have following questions.
I read the thread (http://warp.rice.edu/forums/viewtopic.php?id=366) to shut off some subcarriers. As I am in my initial stage of development, I am currently doing simulation of OFDM code rather than actually implementing the design, by setting some data sub-carriers masks (full rate) to 0 (no change with the base rate yet) in the matlab file ofdm_tx_init.m.
If I do only this, 1) Will this result in data loss during transmission or the OFDM code is smart enough to actually shrink the data to the active sub-carriers? 2) How can I verify this in simulation, preferably by using scopes? and; 3) If the answer to the question 1 is "data loss will occur" then, could you please give me some direction or insight on how to modify the design to prevent the loss?
Thank you!
Offline
The PHY is designed to handle an arbitrary number of active subcarriers per OFDM symbol without data loss. if you disable subcarriers the resulting transmission will last longer, but every payload byte will eventually be transmitted.
Offline
Is the PHY implementation in the 802.11 reference design also flexible enough to handle an arbitrary number of active subcarriers without too much modifications ? I would like to implement an 802.11-like system which uses fewer subcarriers. Thanks.
Offline
Our 802.11 PHY implementation would support different subcarrier maps, but it's complicated by the interleaver. The 802.11 spec defines a "block interleaver" where all the coded bits in a single OFDM symbol are interleaved together. Thus the interleaver logic and subcarrier assignment logic are tightly integrated. If you disable a subcarrier, you must also update the the interleaving pattern to use fewer bits and to not assign any bits to that subcarrier. I built the interleaving logic with ROMs to support arbitrary interleaving patterns (the 802.11 standard pattern can be implemented more efficiently than with a ROM), so it would be straightforward to experiment with different subcarrier maps and interleaving patterns. It's definitely worth testing in simulation first - this is the easiest way to verify you make compatible changes to the Tx (interleaver) and Rx (de-interleaver).
Offline
Thanks for the quick response!
Offline