Help understanding 802.11b SSID decoder - 5 MHz samp rate??

[rpc0618]

Hello,

I'm trying to understand the receiver architecture of the 802.11b SSID receiver example: https://decibel.ni.com/content/docs/DOC-40650 .

How does it get away with sampling at 5MHz, or less than 1/4th nyquist (I'd expect > 22MHz)? It later seems to drop the rate down to 1MS/sec (iirc) after the fft-based frequency offset estimation and correction. What boggles my mind is that 5MHz is not even an integer decimation of the chip rate, so symbols have to be overlapping in a messy way.

I see it receives 802.11g beacon frames that are transmitted at 1Mbps just fine (ERP-DSSS, which should just be like 802.11 and 802.11b, if I'm correct). This is awesome! (also probably Voo-Doo / Black Magic!) That means less work for the CPU, but I'm having trouble understanding the receiver architecture. It seems it ignores the chipping altogether, which might result in maybe 1/11'th the received power being correct, and the rest noise?

Does anyone understand the receiver architecture enough to explain the basics of what it's doing? (Something even as simple as a list of established methods employed, with included links or names that I can google would be greatly appreciated!) Note: I'm currently an EE senior here, still working on my bachelor's so please understand my blunt lack of knowledge.

Thanks in advance,

Richard

[kevins33]

Hi Richard,

Thanks for your post. I'll have to look into the code more, I'm unsure what is happening and didn't develop this solution. It may be a good idea to also post your question on the community example's comments section so that the person who created the example will receive a notification. 

Have you been able to successfully use the code? And are you just curious about the receiver architecture, or is the behavior incorrect?

Also, what hardware are you using?

Thanks!

[rpc0618]

Yes, I've been able to use it to pick up 802.11g beacons (in what I'm assuming is the backward-compatible DSSS/1Mbps mode based on wireshark captures on another computer).

It seems to be correct, I'm just very curious about the receiver architecture because it seems to violate one of the most fundamental principles of signal processing... but since I'm new to it all, and since the actual data rate (instead of the raw chip rate) is only 1Mbps, I'm sure it's possible. I just wanted to know how it's possible, and what methods are used.

I don't remember exactly, but I'm pretty sure it's a 2920 or N210 with an RF-frontend/dauguterboard that can tune to the 2.4GHz ISM band (CBX I think). In whatever case, it works.

Thanks,

 - Rich

[tcap]

Hello Rich,

Have you tried messaging C-Dav or DameOfDataflow? They appear to have made this example. If you send them a message, they will most likely receive an email with your question. I can see what I can help with, but I am no expert on 802.11b.

With regards to the Nyquist question earlier, you mentioned it needs to sample at greater than 22MHz. I think it will only require 22MHz, if the system is running at the full 11Mb/s throughput. I am guessing in the 802.11 specification, it discussing running at lower speeds.

https://en.wikipedia.org/wiki/IEEE_802.11b-1999

This also might explain your chipping issue.

Regards,

Tom