measure mains frequency

[jaspal]

what is the best method for measuring the mains frequency from a household mains power supply as accurately as possible and then to report this on a pc?

 

[flatulent]

First, safety requires that you use a transformer that reduces the mains to a few volt level. Then you need a frequency or period counter.

 

[prinsloo]

ossiloscope2.51,down load from web. (for your comp). and a sound card + a ac transformer with an output of less than 2 volt. this program allows you to store waves etc. on comp

 

[FoxyRick]

If you do PIC's (or other MCU's) that's the easiest, probably cheapest, and accurate enough (0.1% enough? that's easy to get) method.

Just use the timer function to count a single period (detect the zero-crossing with a comparator to start/stop the timer).

Then send this via an RS232 level convertor to the PC.

Signal source from a transformer of course (essential for isolation, not just lowering the volts), maybe plus resistive divider to get the volts down for your comparator.

FoxyRick.

 

[echo47]

Depending on what jaspal means by "as accurately as possible", oscilloscopes and MCUs could be poor choices because they usually have much lower long-term accuracy than an AC mains that (in many countries) is disciplined to an atomic standard.

 

[IanP]

Instead of using transformer you can employ optocoupler; its output can be of TTL level with a pull-up resistor connected to +5V.
This can be connected to one of data lines of the PC's parallel port.
You will need short program, written in any language you know, which will calculate frequency with accuracy set by PC internal clock..

 

[FoxyRick]

Depending on what jaspal means by "as accurately as possible", oscilloscopes and MCUs could be poor choices because they usually have much lower long-term accuracy than an AC mains that (in many countries) is disciplined to an atomic standard.

Agreed. However, its instantaneous frequency can vary quite a bit, before being 'corrected' by running a bit off in the opposite direction for a while.

For a really accurate instantaneous measure, use a fast counter and a TCXO or OCXO and count more periods before calculating.

I have a feeling this may be over the top for the (guessed) intended application.

Jaspal, what accuracy/precision did you have in mind? Is this a nice addition to a computer mod'ers displays, or a serious test instrument?

IanP's idea of an optocoupler is a nice one (why didn't I thnk of that?). If doing it that way, assuming it conducts on only on half-cycle, count the period between two successive on's (or successive off's) rather then just on-to-off. That will more-or-less eliminate the voltage drop required for it to light up and any hysteresys.

FoxyRick.

 

[Element_115]

You can buy a Multimeter and manually enter the data. But if you really want
good data you should get an Oscope, GPIB card, and Labview (or MatLab).
This way you can run a program that will get as much data as you want.
But it's alot of investment.

Cheers

 

[aNdreiBuuu]

do a google search on AVR frequency meter...

**broken link removed**

for mains isolation use an optocoupler (0 detector circuit...)

hope this help,
//a

 

[anggoro_yn]

I agree with element_115.

Get some multimeter, especially one with print-out capabilities. Then check the frequency of your outlet, print it out and send the result to the power company. Complain to the them when the frequency is out of spec or give the compliment when the frequency is as they promised

This is simpler, especially when our main purpose is to make complain to the power company. If we build a D-I-Y devices ..... they will say: "it's your divices problem .... not ours" (at least this is what will happen in my place :-(

aaNg

 

[Borber]

Any counter with standard gate time 1s or 10s has low resolution so the high precision frequency measurement can not be acheived. Better results can be acheived by measuring period duration with high frequency clock. Period duration must be then converted to frequency.