# Testing an LNA on receiving a weak signal with the RTL-SDR

Over on YouTube Adam Alicajic the designer of the LNA4ALL low noise amplifier has uploaded a video showing the effect of an LNA on reception of a weak signal. He shows an example of how a very weak signal cannot be received by the RTL-SDR even when the gain is set to maximum unless an LNA is connected.

Adam has posted this video in regards to some statements saying that an LNA will only increase the noise floor and cannot bring signals out of the noise floor. There is a discussion about this on this Reddit thread.

DVB-T dongle + LNA = Myth or Truth

Subscribe
Notify of

Inline Feedbacks

I the video I see that with adding the LNA the noise floor increases 2-3 dB ( probably as much as NF of the LNA) but in other posts you have mentioned that the LNA can’t increase the SNR at all and will decrease it as much as it’s noise. It is very important to me as I need to have some links working at noise floor and need to know if it is possible to get some SNR using LNAs .

Hello all,

It may look very basic but I wanted to ask you is it practically possible to have wireless board designed with LNA to receive signals lower than the noise floor in a regular or noisy area?

Thanks

Big mistake!
The S/N ratio must be related to the rf noise. You relate it to the ADC noise, which is a completely different thing.
For a correct S/N measurement the indicated noise floor has to be well (by at least 15dB) above the ADC noise to make the tuner dominate the total noise.
Setting the tuner gain to a very low level will show the noise floor of the ADC. Increasing the gain will lift up the tuner noise. At some point it will override the ADC noise. At 15dB above the ADC noise floor, the sum of both will be dominated by the tuner noise. And the tuner noise is the correct “N” in the S/N term.
Applying Frii’s Formula, your LNA4ALL will improve the sensitivity of a R820T based SDR-RTL dongle by roughly 3dB, depending on the frequency.

Do we agree on the following formula for the sensitivity:

S = -174 + NF + 10log(BW) +S/n

NF stands for noise figure
If you decrease the NF for let say 7 dB, how much will Sensitivity increase ?

LNA in the example is lowering the overall noise figure from 10dB to 1dB roughly, hence improving the Sensitivity for 9db. The 3db increased sensitivity will be the case if the NF is improved only 3dB. Declared 3.5dB NF of DVB-T used tuners is far from reality and truth, more likely 7db and higher up to 15dB (on the maximum gain).

Where did you get the “real and true” NF values, being 7dB and up to 15dB, from?

All my measured doubtless results are NF=3,5dB to 4,0dB. All my 7 dongles are RTL-SDR receivers applying a R820T or R820T2 tuner.

You can see, hear and calculate that figures from the Leif, SM5BSZ recent measurements with the videos published online.
More over there are other German guys, who made the CANFI software that came out with the same figures.
Myself also, mine DVB-T dongle has NF of 10dB.

But let assume that you are right, that the NF of that dongles are 3.5-4dB how do you then explain the video I made. The dB scale on the video is calibrated up to +/- 1db and the presented difference is too big for the 3dB only improvement.