Dr. Ulrich Rohde: Electrically Short Antennas

If I may step in here, consider the following example:

You have a broadcast station nearby at 6.9 MHz, producing an S9+40dB signal while you're listening to the 40M band.
In an analog receiver, the 40M signal of interest eventually passes an IF filter, after which most of the gain is before detection. So the very strong BC signal outside the pass band of the IF filter doesn't pose a big problem as the dynamic range of the front end can be 100 dB or more and the signal never makes it to the part where the highest gain is.

An SDR however, has a theoretical dynamic range of 6 dB per ADC bit. The IC-7300 has a 14 bit ADC, thus having a dynamic range of 84 dB max. Any filtering is done in the digital realm and on the HF side, there are just wide filters, several MHz wide. You have set the attenuation in the 7300 so that the OVF indicator doesn't appear. So this BC signal of S9+40 dB "fits" in the maximum range of the ADC. S9 + 40 dB equals (9 * 6 + 40) 94 dB above S0.
Subtract the dynamic range of the 7300 from this number: 94 - 84 dB = 20 dB. 20 dB above S0 is S3. This means that any signal below S3 can no longer be received by the 7300 in this situation.

I am by no means an RF expert but I am an electronics designer and have designed DSP systems in the past. Anybody who wants to shoot holes the theory above, be my guest

 

The direct translation of "raw" ADC dynamic range into RF dynamic range is only strictly valid if the information bandwidth should be equal to the Nyquist bandwidth, fs/2.

In this case, an ADC having 14 bits indeed has a dynamic range of 73 dB or less.

RF engineers do however know that all practical SDR implementations use some form of decimation, which in effect results in considerable oversampling in the information bandwidth.

The "raw" ADC dynamic range has then to be added to a "process gain" or "decimation gain" which is = 10*log(Nyquist bandwidth/information bandwidth), so for the IC7300 and SSB it becomes about 10log(60000/3) = 43 dB, which has to be added to 73 dB.

This adds up to 116 dB which may be somewhat optimistic, and will only be valid as long as the maximum input level to the ADC is closely held at full-scale but not exceeded.

Signals in the range of about -120 to -10 dBm can then be accommodated, but only a few strong at a time. When there are many strong signals in the RF passband, the ADC will overflow more and more often.

This makes RF level control and preselection in front of the ADC more and more necessary when the full available dynamic range becomes utilised.

73/
Karl-Arne
SM0AOM

 

PA0MHS, the magic word is processing gain

https://www.designnews.com/aerospace/where-does-fft-process-gain-come

73, Peter - HB9PJT

 

This was a very good explanation and education for our readers. I guess with my dynamic range analysis of active antennas I invited this and I am happy that this topic gets expanded.

Military monitoring receivers give up some dynamic range for bandwidth, up to 6 GHZ,( there are wideband active antennas available and a SSB receiver needs about 3 KHz bandwidth in comparison.

The advantages of these wideband aktiv antennas are that they are physically smaller then conventional antennas.

 

But do these SMALLER ANTs still provide the GAIN as well as the DIRECTIONALITY of the usual sized yogis?

 

The electrical field strength at a certain point in space produces the same voltage in any conductor, beit a full size dipole or a 1 inch nail. The trick is to "get" this voltage from the nail into the receiver. Whereas the full size dipole has a low impedance and can be connected to a transmission line directly, the nail however has a very very high impedance and requires an amplfier with a very very high input impedance and a low output impedance.

 

Thanks for the explanation Arne and Peter! This explains some effects I observed back in the day. But I was never able to find answers for that as Internet access wasn't available at that time.

 

The principle is ok but not complete.
The processing gain resulting from the FFT also occurs in the time domaine , using analog filters. In addition the FFT-bin-filters do have a sin(x)/x characteristic and therefore need like the discrete FIR filters a post cleanup with a suitable window function like the Blackman -Harris function. The drawback of the FFT filtering is that the individual bin/Filters have a fixed raster (fs/M) with spaces in between.


Some companies have tried to build FFT based receivers which then have a terrible metallic sound for small signals and poor selectivity because of the sin(x)/x function.
In summary it’s better to emulate the analog receiver using digital capabilities

74 de Ulrich

 

I tried to change 74 to 73 and the program does not allow this, sorry

 

These are not arrangements (Yagi) but simple rod antennas, vertical or horizontal dipoles with electronic gain

 

My limited experience here is that cavity filters and mechanical filters as well as heavy RF shielding and common mode current protection are needed in a high RF environment. Both receive and transmit filtering is often necessary so you don't interfere. The closer the nearby signals are to your desired frequency, the more difficult, expensive and diagnostics-intensive this becomes.

When the RF energy coming off of a tower melts the snow in the immediate area, you have a challenge.

 

For those who like to focus more on SDR and state of the art, here is a presentation that may be of interest :

https://badw.de/fileadmin/members/R/3685/Rohde_SDR_Oct2017.pdf

73 de Ulrich N1UL

 

Nice and interesting article. You guys need a native english speaker to proof your material. Poor spelling and poor grammar make this a difficult read.

 

Not another conspiracy theory, I hope.