I took some measurements with my oscilloscope (Rigol 1054z).
edit: I didn't check until I was done with all measurements, but my unit indeed says
NOSE SOURCE. Heh.
The test setup is as such:
- power supply
- bg7tbl noise source
- SMA dc block (18 ghz)
- SMA 10dB pad (18 ghz)
- sma cable
- sma to bnc
- bnc 50 ohm through terminator
- Rigol 1054z
TIME DOMAINFirst, your measurement:
20us/div, measure rms, vpp, avg.
9V inputNote: Your settings were 10V/div, trigger at 10V, unknown termination.
Note: My settings are 200mV/div, trigger at 540mV, 50 ohm termination.
Second, same measurement but
12V input:
And finally, this one I'm happy about:
Same measurement but input voltage
swept from 12.0V to 8.88V (CH2).
FREQUENCY DOMAINFor the next two, low frequency (first few MHz) FFT:
Measurement Settings:
- 100 ms/div
- FFT (Memory*)
- 500kHz/div
- 10.0dBV/div
- 2.50MHz center
- -60dbV
Noise Source
OFFNoise Source
ON
For the next two, same thing but over the full bandwidth of the Rigol 1054z (approx. 130 MHz)
Measurement Settings:
- 100 ms/div
- FFT (Memory*)
- 25.0MHz/div
- 10.0dBV/div
- 65.0MHz center
- -60dbV
Noise Source
OFFNote the spurs and the noise floor level assuming no spurs:
Noise Source
ON (note: This is accidentally 200 us/div, but it doesn't matter)
Note the relative flatness across the range (horizontal marker 2 is about 130 MHz, the 3db point).
I think the jumping is either due to triggering or FFT window edge effects, or both, or neither. I really don't have a clue. Hence AC triggering to randomize.
Here is the 200us/div version, but with the 20M bandwidth limiter on (as a sanity check):
I will follow up later tonight with actual spectrum analyzer measurements.
I'm not a noise, RF, or any other relevant kind of expert so I'll let others draw conclusions from these pictures, if any can be reached. I think the difference between 12V and 9V is quite significant, though.