[time-nuts] Re: measuring the mains frequency

dschuecker Detlef.Schuecker at dschuecker.de
Tue Feb 28 10:34:08 UTC 2023


Hi,

me too. I'm tracking ENTSO-e frequency, the western european grid. I 
sample the down-transformed voltage at 160ksamples/s, apply a high-Q 
digital bandpass filtering and fit a sine wave.  I do this for every 
single sine. The system is clocked with a 1E-8 TCXO. The resolution is 
1E-4Hz. I store the results.

There are some interesting patterns in the frequency. In the first 
quarter of the morning/afternoon hours there is a systematic rise/fall 
in frequency. This has to do with EEX ( European Energy Exchange) 
trading periods and over/underestimation of production/demand.  Maybe 
one could turn this knowledge into money, but unfortunately Im 
interested in the wrong thing (electronics) instead of the real 
important stuff (stock trading).

Cheers

Detlef

Am 27.02.2023 um 20:23 schrieb folkert via time-nuts:
> Hi,
>
> Has anyone experimented with measuring the net frequency? (50Hz in
> Europe, 60Hz in the USA)
>
> Me and a friend are trying to do this as accurate and precise as
> possible. His solution can do that with only a 0.0006Hz error, mine
> sofar does +/- 2Hz.
> Deviations of max. 0.150Hz are allowed (according to
> https://www.mainsfrequency.com/ ) so 2Hz is useless.
>
> I tried the following:
>
> - use the hardware comparator of an AVR328 (Arduino Nano) together with
>    a 10kHz clock-source and then count the number of external clock-ticks
>    between 2 falling edges.
>
> - I used something like
>    https://www.botnroll.com/en/outros/4301-230vac-voltage-sensor-for-arduino.html
>    to make the 230V into a 5V (DC) sine wave and connected that to D6 of
>    the Arduino. D7 is connected to a voltage divider to get a 2.5V
>    voltage reference (should be the zero-crossing voltage of the 230V
>    input).
>
> - connected a PicDiv from Tom (a PD03) to an OXCO and the other end to
>    pin D2 (ext interrupt) of the Arduino.
>
> +/- 2Hz (that's 80PPk right?) is way too much. So i'm curious if someone
> has a suggestion on how to improve this?
>
>
>
> Regards,
>
>
> Folkert van Heusden.
>
> p.s. the source code is at https://vanheusden.com/permshare/50hz.ino
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe send an email to time-nuts-leave at lists.febo.com
>




More information about the Time-nuts_lists.febo.com mailing list