Has there ever been a frequency VS phase application written?
Download location?
1 Like
Hi @CESAUDIOPRO, to see phase info use the Automated Tests → AMP Frequency Response Chirp plugin.
In that plugin, select “Use Right Channel as Reference” and “Plot Phase”.
Externally, you’d connect the QA40x output to the DUT, and the QA40x output would also be routed to the right channel. The right channel connection is needed so that your desired phase reference can be displayed.
The setup you’d use is show below:
In the drawing above, note that the R+ output is treated the same as the L+ output. But you could also use a splitter on the L+ output.
Then, in the plug-in, the settings are follows:
And then run the plugin and you’ll see something like the plot below:
In my case, the DUT was a passive RC filter made from a series 3.3K and shunt 33nF, which gives a corner frequency of 1.46 kHz.
Note that the above is using the 0 dBV input range, which makes sense because the sweep I ran was at -10 dBV. If you set the input range too high, you’ll see noise in the measurement. In the plot below, I set the full scale input to 42 dBV, and then press F3 to the run the same sweep again, and I get the following plot:
Note the noise at the higher frequencies. In this case, the end of the sweep (out near 100 kHz) is coming in around -50 dBV, which is about 90 dB below the full scale, and thus the uncertainty in the amplitude measurement. And since phase is derived partially from comparing the amplitudes, the phase gets noisy too.
1 Like
My interest is primarily concerning active circuits-example profiling active equalizers, looking for phase shift- This discussion as about passive components- Does how does this discussion apply to active devices?
Hi @CESAUDIOPRO, it should be the same. The issue with all phase measurements is that requires a clear definition of your reference. If you want to measure the phase of a digital EQ with, say, ~3 mS of delay, that gets a bit sticky. In that case, you need to know the amount of delay precisely and an incorrect delay estimate and/or a poor selection of the reference frequency will give you a phase error that effectively results in a non-sensical phase reading. So, what can be done for digital filters (not in the QA40x software) is you tweak the delay until you get the plot you want. But that brings about it’s own set of problems.
Below is a digital biquad filter with about 5mS of processing delay using right channel as reference. The ~5mS of processing delay renders the phase info (using right channel as reference) useless.
Another user has asked for the ability to manually specify a path delay. That is a possibility to help with digital filters. And the QA40x can accurately measure path delays. In the plot below, you can see the path delay of the biquad filter processing using the MISC Path Delay Automated Test:
Zooming in a bit, we can see the traces in a bit more detail. The green is the output signal from the analyzer.
Zooming in again, that is a single cycle 1 kHz sine (green):
And zooming in again on the input to the analyzer (from the DUT), we see the response to the stimuli (red):
The blue plot a few traces back was the cross-correlation energy. The peak of that indicates the processing delay, and that’s about 4.833 mS. Now, if we get rid of the biquad gain (make it flat response), and run the test again, we get 4.7708 mS of delay. This is probably the truer measure of path delay. But this highlights the challenge of picking your reference. It’s hard to do on digital filter.
Zooming in again, we can see the analyzer input and output much more clearly when the DUT response is flat:
But if you dont’ have the ability to adjust the digital filter, it’s hard to learn the path delay accurately.
In any case, back to active/passive (non-digital) filters, the phase technique described above in the prior should give you meaningful results. And for digital filters with processing delays, the frequency response doesn’t require a careful reference estimate. And usually you can look at a frequency response and see where the poles and zeros fall, and from that you know how the phase is being manipulated.