Multicouplers

[Ubbe]

Maybe noise measurement can be done in a poor mans way by first checking a receivers sensitivity from a signal generator and then inserting the amplifier in between and if the NF are 2dB the sensitivity should be 2dB worse.

/Ubbe

 

[k7ng]

Have you done a noise Y factor with spectrum analyzer for noise figure before? It can be very accurate if the noise source has an appropriate amount of noise.

It's been quite a long time but yes, I've done that. I worked at a startup that was building the first C-band VSATs and I tested LNAs that way until we'd made enough money to get a noise figure measurement system. And I found the noise source I thought I still had.

 

[prcguy]

Maybe noise measurement can be done in a poor mans way by first checking a receivers sensitivity from a signal generator and then inserting the amplifier in between and if the NF are 2dB the sensitivity should be 2dB worse.

/Ubbe

That will not give you NF as the gain of the amplifier even with a moderate noise figure will show improved receiver sensitivity. You need a calibrated noise source and if I remember right you measure the change in noise level (delta) on a spectrum analyzer with the noise source on and off then convert that to dB, subtract 1, convert back to ratio then subtract the noise source ENR from the delta. K7ng, let me know if that sounds right.

Typical spectrum analyzer settings I used were 1MHz RBW, 100Hz VBW and its nice to have a long sweep time like 20 seconds then turn the noise source on for 2 seconds, off for 2 seconds and repeat for the duration of the sweep so you have several on off ratios to compare. I'll see if I can find an old NF test I've done.

 

[k7ng]

prcguy, that sounds right. 1980 was a while back and I may have a few less brain cells now in the memory bank.
I never thought I'd do a manual y-factor measurement again.
I'll see if I can set up something. I don't have anything with a 'known' NF to check against, but I really only care to +/- 1 dB anyway.

 

[prcguy]

Here is a noise plot I did years back but I don't have the noise source ENR handy. Anyway this is a good way to set up a spectrum analyzer to show the noise level delta with noise source turned on and off. In this case the noise delta appears to be about 6.65dB before any calculations. This was measured at the output of a C-band satellite LNA with the dish pointing into space at a 25deg angle above horizon where there were no signals and a quiet background.

 

[k7ng]

Seems to me I used a 0 MHz sweep at fairly fast rate, just turned the noise source on & off and watch the flat line jump up & down. HP 181T and 8559 plug-in I think.

 

[prcguy]

Seems to me I used a 0 MHz sweep at fairly fast rate, just turned the noise source on & off and watch the flat line jump up & down. HP 181T and 8559 plug-in I think.

That will work but I had to provide data to a customer so the slow sweep is good for that.

 

[k7ng]

As long as an airplane with a radar altimeter doesn't fly thru your field of view.

 

[prcguy]

As long as an airplane with a radar altimeter doesn't fly thru your field of view.

Well then I would see it on the spectrum analyzer display and do another measurement, but unlikely, since its never happened. And I'm usually trying to cram a weeks worth of measurements into 4 or 5 days so I can get back home from wherever I've been sent to do the measurements.

Getting back to sticking an unknown LNA in a home made multicoupler. Better to buy an LNA with the specs you want so you know what you will get. Otherwise it could be worse than using a cheap CATV amplified splitter. And that's the whole reason to make one from scratch, to get the best performance possible at less cost than a commercial unit.

 

[Ubbe]

That will not give you NF as the gain of the amplifier even with a moderate noise figure will show improved receiver sensitivity.

I where thinking of doing like this: Measure directly from a signal generator to the scanner to see its RSSI level when you get a little noise in the audio that equals to 12dB S/N.

Then connect an amplifier and reduce the signal generator so that you still get 12dB S/N.
If the amplifier introduce noise then the RSSI signal level indicated by the scanner will be higher and would be the noise level that the amplifier adds. The dB in difference would be the added NF from the amplifier? Or at least you'll know how much degradation you get from its NF.

/Ubbe

 

[mikewazowski]

Original question was asked and answered.