Noise temperature and G/T calculations

For many years I have had a noise temp calculator page at

http://www.satsig.net/noise.htm

The way the noise temperature contribution of the waveguide and the  noise temperature contribution of the cable have been the subject of  a number of emails providing alternative methods of doing the calculation. Quite which is right or wrong is unclear to me.

If you are an expert or know someone who is, please ask them to comment.

var antcontribution = 1.0 * form.anttemp.value * Math.pow(10,- form.wgloss.value / 10.0);

var wgcontribution = 1.0 * form.wgtemp.value * ( Math.pow(10,form.wgloss.value / 10.0) -1.0);

var lnbcontribution = 1.0 * lnbtemp;

var cablecontribution = (1.0 * form.cabletemp.value * (Math.pow(10,form.cableloss.value / 10.0) - 1.0))/( Math.pow(10,form.lnbgain.value / 10.0)) ;

systemp = antcontribution + wgcontribution + lnbcontribution + cablecontribution + modemcontribution;


form.gain.value = round(form.antgain.value - form.wgloss.value,3);

form.gont.value = round(form.gain.value - 10*Math.log(systemp)/Math.log(10.0),3);

 

The above are extracts from the javascript code embedded in the html page.  To see all the code, go to the page and do View Source (Right Click in Chrome)

To test the equations I have made two examples below. I would have though that they should both give same result, with cable instead of waveguide in the second example.  The calculation results are different. Why ?



The noise from a 1 dB attenuator is 75K at the output but what at the input ?

Best regards, Eric