out of phase signals

[cyprio7]

hi guys.

is it possible to detect how much two identical signals are out of phase by each other and if so the degree of accuracy at which you can detect this in terms of degrees?.. and also, whether i can use ready made chip/s to detect this error of phase difference or would i have to build a circuit for this myself. iv had a look but phase difference detection keeps coming up as PLL on google which isnt quite the info im after

thanks

 

[cyprio7]

..

the signal is RF btw in the MHz region. sorry i forgot to mention that.

 

[audioguru]

A CD4046 PLL IC has a few phase detectors that work up to about 4MHz.
A 74HC4046 has the same plus another phase detector and they work up to about 30MHz.

 

[G8RPI]

Hi,
You don't say what you want to do with the phase information, e.g display on a meter, input to a computer etc, or the level and waveform of the signals.
The simplest device for measuring phase shift (particuarly on sine waves at moderate levels) is to use a double balanced mixer ( DBM ). Feed with two signals of the same frequency and level it will give a DC output proportional to the phase of the inputs.
Mini-Circuits make a range of DBM's the classic being the SBL-1 see
< **broken link removed** >
Some are designed for phase detector use < **broken link removed** > but any will work. As you can see from the data sheet, they are just 4 diodes (Not a bridge rectifier, look closely) and two small transformers.
You can also use either of the detectors from one of the 4046 PLL devices, just ignore the VCO. See http://www.fairchildsemi.com/ds/CD/CD4046BC.pdf Signals in on pins 3 & 14, output for phase detector 1 (EX-OR) on pin 2, Detector 2 on pin 13
For digitial signals use an Exclusive-OR gate (7486) this will give a output pulse with a width proportional to the phase of two inputs, This can be integrated (low pass filtered) with a resistor and capacitor to give a DC voltage.
HTH,
G8RPI.

 

[cyprio7]

i would possibly like to input the phase information into a PIC microcontroller.

thanks for the advice guys, but this is the problem. those phase detectors you showed me, yes they detect phase, but, i need to know more than just the magnitude of the difference of phase. i need to know, which signal is leading and which one is lagging, and by how much.

eg, f1 and f2 are identical frequencies. i detect that there is a 20 degree phase difference. but which one is leading by 20deg and which one is lagging by 20degrees. i need the info on which one is infront and behind. are their any chips which can do this do you think? or if its possible to build something manually to perform this function?

it needs to work in the MHz range, although the higher frequencies at which it can work the better.

 

[audioguru]

The phase detector in a CD4046 and in a 74HC4046 PLL gives a polarity output to indicate which input is leading or lagging. The amplitude of the DC voltage of the phase detector indicates the amount of phase shift.

 

[cyprio7]

thanks for the info. mayb i could get one that works at 433MHz.

If im recieving a data RF signal at an antenna... and say the area has a lot of hard surfaces which cause reflection of RF signal... the antenna will recieve two or more of the same signal wouldnt it because of the reflections, but slightly out of phase and attenuated slightly? is there a way i can seperate this so i can just "take in" the strongest direct signal because i think the reflections will cause a problem?

thanks

 

[audioguru]

VHF Wireless microphone systems for performers and high price car radios use "diversity" which uses two antennas and two RF tuners and a circuit that selects the one that has the strongest signal. Frequently a reflection is the best signal.

 

[raybo]

at one time i had the need for such a device what i did is i biased two very fast gates at treshold and fed the signal thru caps to trigger them it worked pretty good. Ho i forgot you must criss cross the input of both to the outputs like a flip flop. it was fast and cheap. it may not suffice for you because you need square wave input but it is just an idea. the karnaugh map makes those flip flop with unstable states but since it is continiously polled it works.

 

[cyprio7]

audio guru, you know those phase comparators you were telling me about. are they capable of comparing phases to the accuracy of 1 degree or less?

 

[audioguru]

audio guru, you know those phase comparators you were telling me about. are they capable of comparing phases to the accuracy of 1 degree or less?

The datasheet doesn't say.
I suppose you could amplify the phase-error signal if its voltage is too small.
Try it and see.

 

[Russlk]

This is a way to detect lead and lag and phase, how to do it at 433 mHz is the challange.

 

[cyprio7]

thanks for the schematic. so using this schematic, the greater the length of the "phase" signal pulse means the greater the lag? but how do you know actually which one is lagging which if you were to feed it into a PIC for example. i tried drawing the signals but i dont understand.

if i want to measure minute differences in phase, presumably id want to be measuring the lag pulse length, which means measuring timing, which means having an extremely fast clock? unless i transmit at a lower frequency because i currently wanted to transmit at 433MHz because maplin sell some sexy 433MHz antennas that i think would recieve better than a piece of shabby wire lol.

iv been looking at some of the ASK RF transmitting and recieving modules that RF solutions have to offer. Do you think these reciever modules have built in circuitry to deal with multipath propogation or do i have to worry about that seperately, the data sheet doesnt say.

i am receiving two identical RF signals, at two seperate antennas/recievers and then comparing their phases in relation to each other. This is assuming each antenna/reciever is recieving a single copy of that signal. if there are reflections going on, each antenna would be recieving multiple out of phase signals wouldnt it which would screw up my measurements? really i just want each of the two antennas to recieve the *one* signal thats coming direct from the tx..and not a reflected signal.. or at least, for it to know which of the load of signals is the direct one and which is the reflected one.

could i possibly make it "time" which one arrives first and then make it select and take only that signal into consideration instead of the reflected versions?

remember i want to compare the direct signal which is coming from the same tx to two seperate receivers.

yes audio guru i could amplify it thats a good idea but if im after precision, i think errors in the amplifying circuit could filter down into error of measuring the phase difference.

 

[Russlk]

It appears that you want to transmit two signals of the same frequency from one antenna. That is not possible. The signal will be only one with phase a composite of the two. You will need a reference signal of a different frequency in order to seperate them. The reference frequency is commonly 1/2 or double the carrier.

 

[Russlk]

May I suggest that you use subcarriers that are modulated with the phase info? You can seperate the subcarriers easily and use a modulation frequency that can be handled by standard logic. You should have no trouble resolving 0.1 degree. Your clock frequency has to be above 3600 X modulation frequency.