Phase Detection

shocks · Joined: 10 Aug 2006 Posts: 5

Hi Guys
Im stuck on a project .
I have a 3 phase square wave input, 0 to 5 volts . Each square wave is 50 hertz and the phase difference btween phases is 120 degrees.
It basically mimics a 3 phase utilty system set at 50hz. (but in a square wave format.

I need to be able to verify on 3 pic inputs that I have 3 phase voltages present. It sounds straight forward but the voltage goes to zero for 50% of the time on each input, giving a wrong indication.

I would appreciate any ideas.

Im using a 16F877a

Cheers Wink

kender · Joined: 09 Aug 2004 Posts: 768 Location: Silicon Valley

Storic · Joined: 03 Dec 2005 Posts: 182 Location: Australia SA

It is just a thought, I had produced a dimmer circuit where I applied a full wave rectification on 240V phase to nutral, a couple of limiting resistors to drive a 5V opto that at zero crossing I got a 50 hz pulse. I was measuring a 5V pulse.

How about having all 3 Phase pulses the same instead of a square wave, I expect you are using some sort of 1/2 wave circuit that give you a 50 hz square wave.

Andrew
_________________
What has been learnt if you make the same mistake? Wink

treitmey · Joined: 23 Jan 2004 Posts: 1094 Location: Appleton,WI USA

Could a schmitt trigger set with an output pulse longer then the "off" time on all three inputs. or feed them into NAND/AND

Humberto · Posted: Wed Aug 16, 2006 10:47 am

SherpaDoug · Joined: 07 Sep 2003 Posts: 1640 Location: Cape Cod Mass USA

How about:
1) Look for a verified (debounced) rising edge of phase 1 and verify that phase 2 is low and phase 3 is high.
2) Wait for a verified rising edge of phase 2 and verify that 3 is low and 1 is still high.
3) Wait for 3 to rise and verify that 1 is low and 2 is high.
4) Check that the whole process took about the right amount of time (2/3 cycle + debounce time).

This will check that all three phases are there, in the right order, and the right frequency. What more do you need to know?
_________________
The search for better is endless. Instead simply find very good and get the job done.

shocks · Joined: 10 Aug 2006 Posts: 5

Guys
You have all been a great help.
The ideas are first class.
I have established the actual testing for phase rotation, by waiting for a debouced rising edge on 2 phases at a time, and performing this operation twice on a good result to verify correct rotation.
In reality this may be OTT but in if this works out it will be implemented in a Medium voltage format at 10kv and 20kv. At these voltages the actual test points are electrostatically coupled to the phases. I therefore have no source of current , huge impedance and a possibility of a 'Ghost Phase'.This is the reason for not using the ZERO crossing method. The 10 KV system also uses an earth fault protection system tying the faulted phase directly to earth at the station and increacing the other 2 phase voltage accordingly by adjustment of the star point.
All of these facts requires a need for absolute certainty of the presence of 3 phases before giving an indication of phase rotation, as im sure you know a grounded phase can be highly misleading.

I will post progress and any other glitches.

Much Thanks