Signal Noise of Rigol vs Keysight Scope

[bb1]

It is interesting to compare lowest noise levels of Rigol DS2000A and Keysight DSOX1000.

Lowest noise can be observed at the highest real sensitivity, which for both scopes is equal to 1mV/.

It is instructive to get noise as a function of frequency. The plots of noise density vs frequency are shown on two figures
semilogy2GSs_fullBW.png and loglog_2GSs_fullBW.png.

These figures show the same data in different ways. Semilogy uses linear frequency scale more suitable for higher frequencies,
while loglog shows more details at lower frequencies.

Maximum sampling rate of 2GSs was used for both scopes.

We can see that the noise of Keysight scope significantly drops above 200 MHz, which is the scope bandwidth.
On the other hand, Rigol noise declines more gradually at frequencies above 300 MHz bandwidth value.

It is interesting that in a wide range of frequencies inside bandwidth the Rigol noise is significantly lower than Keysight noise.
For example, at 10MHz Rigol noise is about 3 nV/sqrt(Hz), while Keysight noise is about 9 nV/sqrt(Hz).

At low frequencies, below about 100 kHz, noise of both scopes rises with decreasing frequency.
Below about 20 kHz Rigol noise becomes much larger than Keysight noise.

For each frequency point RMS of 1000 FFT amplitude samples was calculated for Keysight scope,
but only 100 FFT amplitude samples were used for Rigol scope.
For this reason Rigol curve at large frequencies seems to be wider than Keysight curve.
The lower number of Rigol samples is due to much slower communication between Rigol and PC.

Python program keysight_getNoiseDensity.py used to download data from Keysight scope to PC is attached.
It fully explains how the noise density was calculated.

Exactly the same noise density calculations were used for Rigol data,
while the code for downloading data from Rigol is much more convoluted.

In all cases the scope inputs were externally terminated by 50 ohm load.
Internal Rigol 50 ohm termination was not used.

It is interesting to compare observed noise densities to the 50 ohm resistor
thermal noise density which is about 1 nV/sqrt(Hz).

[tautech]

It is interesting to compare lowest noise levels of Rigol DS2000A and Keysight DSOX1000.

Below about 20 kHz Rigol noise becomes much larger than Keysight noise.

To properly understand what you see with your tests you need to study this thread:
https://www.eevblog.com/forum/projects/project-yaigol-fixing-rigol-scope-design-problems/