RF data baud rate & signal propagation

[Houston_Chuck]

Some co-workers are experiencing degraded signal coverage with RF data transmitted at 900mhz. They have recently converted from a MDC 4800 rate to RD-Lap 19.2. The reports are that the coverage is not as good now as it was using the MDC 4800. Poking around on the web I found one article that stated "doubling the baud rate results in a 3DB reduction in signal strength". I am looking for some other literature dealing with this topic. Any suggestions, or past dealings with a situation like this?

Thanks

[SlimBob]

well... if you take a time-based error, twice as many bits will go by at double the speed.

[Al]

You will probably find that the quantity being referred to is Eb/No rather than signal strength. Eb/No is the ratio of energy per bit to noise power density at the receiver's demod output. Halving the bitrate *or* increasing the carrier power by 3 dB will increase the Eb/No by 3 dB. The decrease in bit error rate accompanied by adjusting either of the two parameters I've noted above will be dependent on exactly where on the Eb/No vs. carrier power curve you are now operating your system. Any good technical book on digital communications should show the curve I've mentioned. Try looking at Communications Satellite Handbook by Walter L Morgan and Gary D Gordon, page 589, ISBN 0-471-31603-2. Another resource is Telecommunication Transmission Handbook, 2nd Ed. by Roger L. Freeman, isbn 0-471-08029-2, page 27.

[Jim202]

You might want to try your hand at doing a search for mobile data. There are a number of companies on the market today that offer fairly good data rates over radio.

The big problem is the higher the data rate, the lower the distance of the high speed data. What this translates to, is given the same RF power out, for every time you raise the data rtate, the lest distance you will be able to go with the higher data rates.

Many of the public safety systems today boast some form of data. The rate of 9600 bps is about normal. as this rate goes up to say 19.2 kbps, the range it will cover shrinks acordingly. This is a major problem for system engineers today.

With the radio users clamering for higher speed data, the FCC is telling the users to cut their bandwidth from 25 Khz. to 12.5 Khz. This has the same effect as trying to increase the data rate. The lower bandwidth causes a lower data rate to be carried over the same radio channel.

Bottom line is high data rates demand high radio channel bandwidths. It is a constant battle for the system engineer to find a solution.

Jim

[psapengineer]

You might try reviewing the "Gain Bandwidth Product" thoery. Essentially it states that for a given system; gain times bandwidth equals a fixed quotient. So, if the bandwidth is doubled the gain is forced to half. This is old time hard core EE stuff.

We can view data modulation on a spectrum analyzer. We then need to think of the power being delivered by the modulator as equal to the "area" under the spectrum's outline (also called a curve; calculus stuff).

So if the spectrum is twice as wide and the area under the curve is fixed (ie: modulator delivers the same amount of power) then the height or amplitude of the "curve" must be half of what it was before since the "area" remains constant.

Enuf........... Uck..........

Hope this helps, Bob

[psapengineer]

And here's one more thought:

Since gain in the modulator is measrued as voltage gain or loss, the difference may be far greater than the 3dB you found in one article, more on the order of 6 to 12 dB for your application; Thus:

Gain/Loss = 20 Log V2/V1 where V2 is the final/new voltage out and V1 was the voltage before.

So, if one doubles the bandwidth twice it's equivilient to halving the gain twice:

Loss = 20 Log (.5)(.5) = -12 dB and that's a bunch of DB; maybe enough difference to make a class C amplifier quit working.

[radiomidwest]

Hi Chuck,
We converted from 4800 to 19.2 here back in 2000 or so. Yes there is a severe reduction in coverage area. The simple fact is that the 4800 had a simple waveform. Easy to generate. Easy to decode. The 19.2 has a 4-level FSK waveform that has to be generated properlly and then on the rx end more signal processing. With 4-level FSK, it is more difficult for the decoder to determine what is noise and what is signal.

That's it in a nut shell. The more complex the waveform, the greater signal strenth required. Because the decoder is less forgiving.