Page 1 of 1

Bode Analyzer seems quite good

Posted: Sun Nov 30, 2014 8:01 pm
by Ghisluca
I would like to share my positive impressions about the Bode Analyzer.
After unzipping the latest version, I have copied the command-line version, bode, to the root/bin directory of the SD card. The reason why I prefer the command-line version is that I can then copy the measurement data into Excel, and format the plot as it best pleases me. The terminal program I used is PuTTY.
I have then plotted the transfer function of a simple RC pole, 1000 ohms and about 1.2 uF (MKT) ==> about 133 Hz, sweeping logaritmically from 10 Hz to 1000 Hz , averaging 10 samples at each one of the 200 frequency values:
bode 1 0.1 0 10 200 10 2000 1
even with just 100mV amplitude stimulus the plot is quite clean, with just a bit of noise showing up when the ratio Out/In approaches -40 dB (that is, when Out approaches just few mV ... I am favourably impressed ! ). Also, at -3 dB the measured phase is -42.5°, very closed to the -45° theoretically expected value. So, overall: good A to D and related processing performances. Bravo!
I wonder why the blog states that the minimum frequency is 1000 Hz, as it appears to work just fine also from as low 10 Hz ... true, you have to be patient for quite a while ( > 1 hour ...), as it seems to take much longer when starting from such low frequency. But, as far as the pattern of lit LEDs keeps changing (well, now and then) you should not worry that it might have freezzzzz ...
Side question:
might it be that the reason it becomes so slow at low frequencies could be due to a fixed sampling rate, and which would thus result in many more samples needed to cover one complete period at low frequencies ? if so, then the next logical improvement might consists in adapting the sampling rate to the particular frequency value being processed at each measurement point.

Re: Bode Analyzer seems quite good

Posted: Sun Nov 30, 2014 8:24 pm
by Ghisluca
sorry, the plot's 3rd decade was not completed.
Replace with this.

Re: Bode Analyzer seems quite good

Posted: Mon Dec 08, 2014 5:31 pm
by redpitaya
Hello Luca,

Thank you for your positive impressions, your question and for your side question! We love side questions :).

Well, long measurement at low frequencies occurs, because the bode code for amplitude calculation (lock -in method) always acquire the same number of periods of signals. The time needed to acquire for example, 10 periods of 10Hz signal in respect to 10 periods of 1MHz signal is much longer.

Minimum frequency is 10 Hz.

Thank you, and I hope that helped.

Feel free to contact us if you have any more questions.