90 degree 1

[sparkyyyyy]

Hi,

I have an incoming frequency that is set to 60 Hz. I will call this my sync signal. I also have a 60 Hz signal that varies in phase and amplitude, which I will call my modulated signal. This signal is from the same source as my sync signal. Although the sync signal is not modulated.

I need to be able to demodulate the modulated signal. I already have an in phase component which is my sync signal. I need to be able to generate a signal that is a square or sine wave that is 90 degree out of phase from my sync signal.

If it helps to solve the problem, I can also change the phase of my sync signal. What is important is that I have two signals that have a 90-degree phase relationship and are phase locked to the sync signal. I will use these two signals to demodulate my modulated signal.

In addition, I would like to have this circuit work if I change my frequency from 60 Hz to 50 Hz.

Thank you for any help you can give me.

 

[fjrg76]

Ther is a thing with such a name is "quadrature-encoder" It can help you

 

[sparkyyyyy]

Thank you fjrg76

I am not sure how to apply the quadrature encoder to my problem. I don't generate the 60 Hz signal. I just need to synchronize to it. Then generate both in phase and quadrature signals that are synchronized to the incoming signal.

Thanks,

 

[cbarn24050]

Hi, at these low frequecies the simplest way would be to use a micro (a pic,or avr) to get your phase shift.

 

[nbucska]

Convert the input to squarewave with AC coupled comparator
with AC hysteresys. Phaselock to it a 120Hz PLL and
reclock with it the sqw twice.


<nbucska@pcperipherals.com>

 

[rodar]

If you do not need hair splitting accuracy
try a RC filter with a cutoff of 5 hz. This
will give a phase shift of around 85 deg.
Then feed this to a AC coupled comparator to
recover the full strengh phase shifted signal.

Otherwise nbucska has a good solution.
Generate a 120 Hz squarewave with a voltage controled
oscillator then divide it by two and feed a
phase lock loop to sink it with the 60 hz reference.

And cbarn24050 is correct that it is a simple program
for a small CPU. If your circuit allready has a CPU
this is the way.

Rodar

 

[sparkyyyyy]

Hello,

Thank you all for your suggestions. I will need a 90 +/- 1 degree quadrature signal. I do not have a CPU in my circuit. I must also admit that I am not sure how to use a CPU to accomplish a 90 degree phase from my sync signal.

I used the phase locked loop concept, I think my 90 degree signal was noisy even though it was phase locked and looked like it should work. I believe this because when I used an RC filter to get close to my 90 degree phase difference, I was able to demodulate my modulated signal without noise showing up in it. However I did not have the 90 +/- 1 degree accuracy I needed. When I used the phase locked loop concept, I had a lot of noise in my demodulated signal that I did not see when I used the RC filter. I know I need to do some more work on this. I am not sure why I am getting this noise in my phase lock loop circuit. Any additional suggestions are appreciated. Thanks again to all of you who offered suggestions.

sparkyyyyy

 

[Heydave]

You can use a 4046 type PLL and a few flipflops to multiply the frequency of the sync signal. The various outputs of the flipflop counters can be used to create whatever phase shifted clocks you want. You should be able to create a stable lock over a 2 to 1 frequency range easily. You could add A/D, fifo, and D/A to make a phase delayed analog signal.

 

[cbarn24050]

Hi,sparkyy, is this a &quot;1 off&quot; or do you intend production?

 

[sparkyyyyy]

Hi cbarn24050,

This product will go into production.

 

[cbarn24050]

Hi sparky, in that case you could just ask someone local to sort out a program for a 8 pin PIC. It's a trivial program and the part is very cheap.

 

[sparkyyyyy]

Hi cbarn24050,
Thank you for your input. I will try to see if I can find someone to program a PIC for me. I have not done this myself. Thanks again.

 

[sparkyyyyy]

In a previous email I indicated that the incoming frequency could be 50 or 60 Hz. When I think of using a PIC, I think of using a counter to delay my output after so many clock cycles of the PIC. For example, I would delay a certain number of clock cycles(PIC Clock =2 MHz or some other value). When I have a 60 Hz input in order to generate my 90 +/-1 degree signal, I would clock in the rising edge of my 60 Hz signal and delay it by 1/(60x4) seconds and then change my output from low to high. I would also do this for the falling edge of my 60 Hz signal and change my output from high to low. Am I correct so far in this approach? Would the PIC sense if a 50 or 60 Hz source was applied to the circuit by counting the time from rising edge to falling edge? Thanks

 

[cbarn24050]

Hi sparky, yes thats about it. Use your input to cause an interrupt which starts a timer. When the timer expires it reads the input and outputs what it &quot;sees&quot;. You can also run a counter to determine the frequency of your input and so calculate the required timer interval. It would then not need an accurate crystal clock and would work with any frequency. You could even use an analouge input with a pot to enable you to adjust the the phase shift, &quot;fine tuning&quot; if you wanted.

 

[sparkyyyyy]

Hi cbarn24050, I appreciate all of your help. This is the approach that I am going to take.

 

[sparkyyyyy]

Hello to everyone who responded to my post. Thank you for taking the time to offer suggestions to my problem.