Diversity Combiners 2

[stockdam]

Can anyone help me to understand the differences between pre and post detection diversity combiners.

I'll describe the application.

I've a digital PCM signal that is being transmitted at S Band (say 2GHz). It is transmitted using left and right hand circular polarisation.

The receiving antenna then passes the two received signals to two receivers and they are connected to a diverstity combiner. The combiner can either be a pre detect or post detect. WHich is better and why?

 

[VEBill]

Pre-detection can improve the signal-to-noise ratio just as would an antenna with more gain (like an array). It can help you to get under the noise at the most critical stage, the RF. Once you detect the two signals, then any common RF noise (like that coming from the wrong direction) becomes indistinguishable from the signal. Pre-detection is better, but more difficult.

I'm curious about how you send the same (?) signal with two opposite circular polarities. Are they on the same RF channel? If so, then you might just have made slant linear polarization the hard way.

 

[stockdam]

Thanks. I already "knew" that pre-d was better than post-D - do you have any links or articles that would explain why?


I don't know much about the application yet so I don't know if they are transmitting on different frequencies (I guess they are).

 

[VEBill]

I can't find any.

 

[stockdam]

No I can't find anything either......it seems to be a "secret"

 

[tbshmkr]

Here is an address to a site with a paper on pre/post-detection combining.

http://citeseer.ist.psu.edu/rahman05pre.html

While in the USAF, I worked on troposcatter (2GHz) systems with analog post-detection quad diversity combining receivers (tubes). These were upgraded to pre-detection quad diversity combining receivers (solid-state).

The modernization from tube to solid-state was the significant improvement.

Check library for books on radio communications, especially tropospheric scatter radio.

"The Lenkurt Demodulator" was a trade magazine that covered microwave/troposcatter communications. You might be able to find issues in a library or a used bookstore.




The rent is due.

 

[tbshmkr]

Troposcatter was the primary longhaul multi-channel communications network for the world-wide Defense Communications System. Sometimes called radio relay.

BTW - The transmit string was baseband feeding 1 IF modulator (70MHz) feeding two up-converters (2.1 & 2.4 GHz, for example) feeding RF Power Amps (1KW). The receive string was 4 receivers (2.1V, 2.1H, 2.4V, 2.4H), each down-converted to 70 MHZ, combined to a single 70MHz channel, and detected to baseband.

The baseband was analog FDM (frequency division multiplexing).

The same system "supposedly" could be used for digital TDM (time division multiplexing) with filter changes.

The rent is due.

 

[VEBill]

ViaSat offers the QDC-100 for pre-detection combining of UHF MilSatCom in the area of 300 MHz. But I don't recall seeing any simple explanations of why pre- is better than post-detection combining.

http://www.google.com/search?q=ViaSat+QDC-100

One simple explanation is that with multiple antennas being actively combined, you could (for example) steer a null onto a noise source based on the antennas and wavelengths at the RF carrier frequency. But once you detect the signals and noise, then the noise (now baseband) cannot be canceled without also canceling the signal.

The fundamental advantage is the differences in the several signals at RF. Many of those differences disappear once you detect the baseband signals. By detecting, you're tossing away critical information that exists only at RF.

That's the explanation programmed into my gut. But I suspect that it's only one aspect of a larger truth.

 

[tbshmkr]

Reference (1977)

Design Handbook for Line of Sight Microwave Communication Systems

http://www.everyspec.com/MIL-HDBK/MIL-HDBK+(0300+-+0499)/download.php?spec=MIL_HDBK_416.2088.pdf

para 4.4.25.2 - "...the earlier the signals are combined the better is the ultimate performance."

sec 4.5.34 Diversity - more explanation of diversity and combining with digital/analog system.

--

The rent is due.

 

[tbshmkr]

New Reference (2008)

I knew there were newer references. This is from a computer science paper.

http://www.tucs.fi/publications/attachment.php?fname=TR884.pdf

The paper concludes pre-detection maximal ratio combining is optimal to maximize SNR of the received signal.

Post-detection combining may perform better under some conditions, especially impulse noise.

--

The rent is due.

 

[stockdam]

Thanks for the links. As it says pre-d is better but post-d is cheaper and easier - it's difficult to decide whether you need pre-d or wheher post will be sufficient.