View Single Post
Old 15-Aug-2020, 2:02 PM   #13
Retired A/V Tech
Join Date: Aug 2012
Location: S.E. VA
Posts: 2,738
Originally Posted by AKADAP View Post
SNR is the ratio of signal power to noise power. Spectrum analyzers are deceptive in that the apparent noise floor changes with the resolution bandwidth setting. The narrower the resolution bandwidth, the lower the apparent noise floor. One can't just look at the spectrum plot and measure the signal power to noise floor. One must integrate the signal power over the bandwidth, and noise power over bandwidth, and take the ratio of the two results.
The resolution bandwidth used for the scan will definitely affect the power reading. The thermal noise floor of -106 dBm for a DTV signal is determined by its bandwidth of 6 MHz.

However, since the same resolution bandwidth is used for the DTV signal and the adjacent noise floor, the power ratio between the two is preserved.

A $25 SDR dongle used as a spectrum analyzer is not lab-grade equipment, but when the SNR given by the scan agrees with the SNR given by my SONY TV as shown in post #10, that's good enough for me. However, it is necessary to adjust the SDR gain to the proper level. With the RTL-SDR.COM V3 dongle and the free open-source spectrum analyzer software, it should be increased to the point where the noise floor just starts to rise. With the RTL-SDR.COM V3 dongle and SDR# (SDRsharp) software, SDRplay RSP1A and Airspy R2, the gain should be adjusted for max SNR, which is usually not max gain.
Often when people discuss preamps, they focus on a single channel of reception. This is dangerous since preamps are wide band amplifiers and there could be high power signals within the bandwidth of the preamp.
Very true. The front end is wide open; there is no preselection. One very strong signal will cause clipping and gain compression; two or more very strong signals will cause IMD that creates spurs which will raise the noise floor and wipe out weak signals.
All amplifiers have a limited output voltage range. If the instantaneous input voltage times the gain of the amplifier causes the result to be outside that range, the output of the amplifier will clip. During the time that the amplifier is clipping, nothing gets through.
Yes; I learned that lesson the first time I connected two preamps in series; nothing got through.
This means that if there are any high powered signals, it can be a bad idea to try and use a preamp. In my case, I found that local FM stations were hugely more powerful than the TV stations I was trying to receive. I found that I needed an FM trap. There are two types of FM traps, and marketing seldom telly you the kind they are selling. One type kills all FM, but also kills channel 6. The other allows channel 6, but does not kill the low end of the FM band.
That would be the Radio Shack FM filter and the Antennas Direct FM filter:

FM filters are getting hard to find. A HLSJ makes a good FM filter if you don't need VHF-Low. FM filters are made for SDRs, but they are designed for 50 ohms and have SMA connectors.
For digital television I wish I could find a low cost tool that would display the constellation plot of the signal. This can give good clues as to why you are having trouble with a signal. The closest I have seen is the HDHomeRun Tech, but that is not a full constellation plot it is a histogram of the I signal rather than a plot of I vs Q.
Majortom on the Canadian forum has worked out a low-cost method. It requires Linux software manipulation which is a little beyond my ability. I'm happy with just signal strength, SNR and errors. The shape of the signal when finding the best location for an antenna is also useful. Fewer notches seems to be better.
I modified the linux driver for this usb atsc tuner to allow plotting the received constellation, using updateDVB.
Working pretty good here...

a weak distant cblt

strong local wned

Attached Images
File Type: jpg 8vsb-iqplot-cblt_2.jpg (95.6 KB, 922 views)
File Type: jpg 8vsb-iqplot-wned_2.jpg (83.1 KB, 952 views)
If you can not measure it, you can not improve it.
Lord Kelvin, 1883

Last edited by rabbit73; 15-Aug-2020 at 4:00 PM.
rabbit73 is offline   Reply With Quote