Issue with Circuit analyser?

Hi, I recently got a PCSU1000 to complement my PCGU1000. Very happy with the intstrument but I ran into an issue with the Circuit Analyser.

Running this on a Win10-64bit PC and software version 4.07.

If I make the following settings:
PCGU output at sine, 0.10Vrms = -20dB output signal level.

PCSU1000 at 50mV input range, both channels.

Output a of PCGU1000 to input 1 of PCSU1000 via coaxial cable.
Output 2 of PCGU1000 to input 2 of PCSU1000 via coaxial cable.

When I run a bode plot without the option ‘Options -> Normalise output by input’, I get the following results:

Hz mVrms dBV deg
1000 100.812 -19.930 -0.1
2000 100.952 -19.918 -0.2
3000 100.915 -19.921 -0.3
4000 100.681 -19.941 -0.2
5000 100.826 -19.929 -0.1
6000 100.825 -19.929 -0.3
7000 101.196 -19.897 -0.4
8000 101.457 -19.874 -0.3
9000 101.465 -19.874 -0.5
10000 101.625 -19.860 -0.3

etc…

The measured output is at approx. -20dB, as expected.

Now, When I run a bode plot with the option ‘Options -> Normalise output by input’ enabled, I get:

Hz mVrms dBV deg
1000 0.100 -80.004 -0.1
2000 0.100 -79.998 -0.2
3000 0.100 -80.004 -0.2
4000 0.100 -79.998 -0.3
5000 0.100 -80.002 -0.3
6000 0.100 -80.003 -0.3
7000 0.100 -79.997 -0.4
8000 0.100 -79.993 -0.3
9000 0.100 -79.992 -0.5
10000 0.100 -79.995 -0.3

etc…

From what I expected, the -20dB output signal should be normalised by the -20dB input signal, resulting in -20dB(out) - (-20dB(in)) = 0dB transfer from my coaxial cable.

So why am I seeing -80dB? Are my expectations incorrect or am I missing something???

Thank you for reporting this issue.
I’m sorry for the inconvenience.

Indeed, the problem seems to occur only if option “Normalize Output by Input” is selected and CH2 Volts/Div setting is 50mV/div or lower.
As a workaround you can use 0.1V/div setting on CH2.
In “Normalize Output by Input” mode the CH2 is only used to monitor the generator output voltage.
The accuracy on 0.1V/div range is good enough for this purpose. On CH1 you can use all Volts/Div ranges.

Thanks for the reply, I had some time to test this option again…

PCGU output at sine, 0.10Vrms = -20dB output signal level.

PCSU1000 at 0.1V input range, both channels.

Output 1 of PCGU1000 to input 1 of PCSU1000 via coaxial cable.
Output 2 of PCGU1000 to input 2 of PCSU1000 via coaxial cable.

When I run a bode plot without the option ‘Options -> Normalise output by input’, I get the following results

Hz mVrms dBV deg
1000 100.909 -19.921 -0.1
2000 100.793 -19.931 0.0
3000 101.215 -19.895 -0.3
4000 100.970 -19.916 -0.2
5000 101.032 -19.911 -0.1
6000 101.285 -19.889 -0.3
7000 101.546 -19.867 -0.3
8000 101.876 -19.839 -0.3
9000 101.818 -19.843 -0.5
10000 101.916 -19.835 -0.4

As expected, as the function generator outputs -20dB signal.

And the same measurement again, but with the option ‘Options -> Normalise output by input’ enabled:

Hz Vrms dBV deg
1000 0.101 -19.927 -0.4
2000 0.101 -19.926 -0.1
3000 0.101 -19.932 -0.3
4000 0.101 -19.930 -0.5
5000 0.101 -19.926 -0.1
6000 0.101 -19.933 -0.3
7000 0.101 -19.922 -0.3
8000 0.101 -19.923 -0.2
9000 0.101 -19.917 -0.5
10000 0.101 -19.917 -0.3

A better result compared to -80dB, but it lists the transfer again as -20dB, while I would expect it to be -20dB output Ch1 measured signal, normalised by -20dB Ch2 Reference signal, = 0dB transfer from the coaxial cable (or a device that has its transfer measured) between PCGU1000 and PCSU1000.

So are my expectations incorrect, or is it still an issue with the software?

In “Normalize Output by Input” mode the input signal is monitored by CH2 during the scan and the voltage level is compared to the generator output voltage setting.
The recorded CH1 values are corrected according to the variations of the CH2 level during the scan.
If there is no difference in the input signal compared to the generator setting, the output voltage values are plotted without any correction.
This is why you get similar voltage values in both cases.

In both cases the plotted values are the voltages, not transfer ratios of the circuit.

Anyhow, using “Normalize Output by Input” mode the shape of the plot corresponds better the transfer function of the circuit tested. Especially if the input impedance of the circuit is low and dependent on frequency.