Amplifier use in antenna pattern testing. 4

[noisy]

I am doing antenna pattern testing and wish to use an amplifier to increase the signal as the antenna pattern occasionally “goes down into noise”. –However, I do not know whether to put the amplifier on the receive side or the transmitter side.

The antenna patterns being acquired are those of a parabolic dish and the frequency is 0.5 to 1 GHz (patterns taken at 100MHz intervals). The test range is 40 metres. I only have one source antenna and so cannot swap for one of increased gain. The antenna_under_test feeds into a spectrum analyser (possibly Marconi 8569B).

My RF power source can output 10dBm Max. Even when the RF power source is set to supply a particular frequency the spectrum analyser shows that it actually transmits a narrowish band (say 40Khz) of frequencies with a central peak at the ‘selected’ frequency. The only amplifier available to me has 13dB of gain (between 0.5 and 2GHz, also Noise Figure of 1.8dB and VSWR of 2) –However, the maximum power that can be input to it is 0dBm (it goes non-linear after that and no increase in output power occurs) . Therefore, if I put the amp immediately after the RF power source then I have to turn the power source down to 0 dBm –and then obviously get just 13dBm out of the amplifier. (With the RF power source alone, I could have got 10dBm, so with the amp on the transmitter side I only get an extra 3 dBm.)

Therefore, I was wondering whether or not to put the amplifier on the receiver side –since then we would be using the amplifier’s full 13dB of gain. –However, since this amplifier is relatively wideband (13dB of gain between 0.5 and 2GHz), and since we are not following it with a bandpass filter, I was wondering whether we would end up amplifying all the noise received between 0.5 and 2GHz –and then possibly get problems with intermodulation products when this is all fed into the spectrum analyser. –Thus possibly giving a false reading of received amplitude at each transmission frequency.

I would be very grateful for advise on whether to put the amplifier on the receive side or transmit side. –Also, when antenna pattern testing, -if the pattern goes “down into noise”, how does one know if this is because of a genuine null in the antenna pattern, or because more signal amplification is needed? -Sorry its so long!

 

[VEBill]

"...the spectrum analyser shows that it actually transmits a narrowish band (say 40kHz) of frequencies..."

That's probably due to a setting on your spectrum analyser.

 

[VEBill]

With a noise figure of 1.8 dB, it sounds like a pretty good receive side preamplifier. Try it and see.

 

[VEBill]

At the outset, decided the scale of your dB (min, max). No point pulling your hair out if you only need a 30dB amplitude range on the final chart.

 

[noisy]

Thanks VE1BLL, replies appreciated. -i must admit i had thought that it would be a waste of time putting the amp on the receive side though, -since if the signal has "gone down into noise", then won't the amplifier just amplify all the noise between 0.5 and 2 GHz?....the amp will not be able to "pull the signal out of the noise", since it is not tuned to whatever the specific frequency being used is.

That's why i favour the transmission side for the amplifier, (despite the problem of lower gain when used here) -though my head tells me i'm missing something!

 

[VEBill]

It depends where the noise is coming from. If the noise is coming from the noise floor of your spectrum analyser (quite likely), then an external amplifier with a lower noise floor may very well help.

At these frequencies (0.5 to 1 GHz), most receivers make their own noise (as can be seen by simply disconnecting the outside antenna and notice that the noise probably doesn't change in the slightest). That proves it's internal noise.

At lower frequencies, disconnecting the antenna makes most of the noise go away. It's possible to get into the external noise floor at very high frequencies, but you need to try hard.

Most spectrum analysers aren't really THAT good receivers (some expensive ones are). Of course, they have other advantages over true receivers (like covering the entire band in milliseconds...).

 

[pstuckey]

Noisy,

Most spectrum analyzers have noise figures in the 30-40 dB range. So putting an LNA in the receive chain before the spectrum analyzer will help because a lot of the noise is introduced to the system from the spectrum analyzer. Note that you are not in this case "pulling the signal out of the noise", you are just amplifying the signal prior to a large noise source.

NF.total = NF.1 + (NF.2 - 1)/G.1 + (NF.3 - 1)/(G.1 * G.2) + ...
(all quantities are linear, not dB)

Peter

 

[BrianR]

A side issue here is what is your local noise environment, like mobile phone towers and two-way radio transmitters operating in your band. Local interference would drive you to amplify the transmitted signal rather than the received signals.

 

[VEBill]

Interference can go both ways. Even 10mW into a moderate gain antenna (accidently aimed in the general direction of a local 'cell phone' tower) might not be the best idea.

Plain white vans with funny antennas might eventually show up. In Canada - a stern warning. Slightly south - a quadrillion dollar fine and seventy-two thousand years in jail.

 

[remzorado]

Please help me solve this: A land mobile base station has 120(degrees)beamwidth in the horizontal plane and a mechanical beamtilt of -3(degrees)in the direction of the city of licence. What is the beamtilt at the edge of the beam, at 60(degrees)?

 

[VEBill]

-1 ?

 

[waveboy]

Hi Remzorado,

I make it -1.5 degrees. It was easyish once a few diagrams that looked like wedges of cheese had been drawn. You provided some nice numbers because it allowed one to work with the 30,60,90 triangle, (which of course has sides with with ratio 1,2,root3).

By the way, do you really finely calculate beam angles etc in your industry, or is it more a case of sending guys into the desired coverage area with RF power meters? (just curious)

 

[VEBill]

-3 at 0 degrees (given)
-2 at 30 degrees
-1 at 60 degrees (the correct answer)
0 at 90 (the axis of tilt)

In summary: -1 !

(It would be -1.5 at 45 degrees.)

Assuming that I understand the question (?).

 

[VEBill]

Or maybe not? Did I over-simplify?

 

[VEBill]

Doh! 'waveboy' was right.
cos(60) and all that.

[flush]