Using the PCS10 for the recording of battery-charching I found low-frequent (0,02-0,03 Hz) oscillations between maximum voltage of the charger and zero!. These oscillations I couldn’t find with a normal oscilloscope or with a multimeter. What I found was a 50 Hz oscillation on the charger current with an amplitude of ca. 10% of the charging voltage.
What may be the reason for this low frequent oscillation and how can it be removed?
[quote=“biber”]Using the PCS10 for the recording of battery-charching I found low-frequent (0,02-0,03 Hz) oscillations between maximum voltage of the charger and zero!. These oscillations I couldn’t find with a normal oscilloscope or with a multimeter. What I found was a 50 Hz oscillation on the charger current with an amplitude of ca. 10% of the charging voltage.
What may be the reason for this low frequent oscillation and how can it be removed?[/quote]
With what instrument did you observe the 50Hz oscillation?
I ask because the PCS10 has a maximum sample rate of 100S/s. Using it to observe the the 50Hz component places that signal right at the Nyquist frequency, where even a slight differential between the 50Hz component and the PCS10 sampling frequency would appear as aliased signals at the beat (difference) frequency.
This could explain the low-frequency (0.02-0.03Hz) excursions seen in the digital monitoring, that do not appear using the analog tools because they are not really there…
Based on the explanation in the previous mail, most probably the “oscillation” was the result of aliasing. The input frequency, about 50Hz, was aliased to lower frequency.
I made a test:
Using the function generator I generated a signal having a frequency of 50.01Hz.
Then I fed this signal to the input of the PCS10 and set the sampling period to 10 samples per second (10Hz).
The resulting waveform seen on the PCS10 screen was like this:
This is absolutely OK and normal effect when the sampling rate is lower than the Nyquist frequency.
Here is the same waveform seen on the oscilloscope screen:
to cliffyk:
the instrument I used was an old analog 5 MHz oscilloscope.
The explanations of the aliasing problem seem to be resonable. I will make some further tests and post the results.
Thanks to you both.
[quote=“biber”]to cliffyk:
the instrument I used was an old analog 5 MHz oscilloscope.
The explanations of the aliasing problem seem to be resonable. I will make some further tests and post the results.
Thanks to you both.[/quote]
That explains it. You must live in an place that uses 50Hz mains power and the 'scope was seeing that frequency riding on the battery charger’s poorly filtered output (this is not unusual, here in the 'Sates it would be a 60Hz signal).
The PCS10 was sampling this charger output at a frequency that placed the Nyquist frequency (½ the sample rate) near 50Hz or some multiple thereof. The result being the aliased (phantom) 0.02 to 0.03Hz “signal” you observed.
Here’s a good illustration of why aliasing occurs:
To put some numbers to it, here a “real” 90Hz wave is being sampled at 100S/s; resulting in the sampling instrument’s “seeing” a 10Hz aliased signal.
[color=#0000BF]This is a straightforward description of the “aliasing problem”[/color].
In practice, for any serious work, the Nyquist frequency should be at least four to five times higher than that of the signals to be observed.