Discussion in 'Electronic Design' started by Don A. Gilmore, Jan 3, 2004.

1. ### Don A. GilmoreGuest

Hi kids:

I've got a signal from a sensor that I convert to a simple, clean square
wave at an audio frequency (anywhere between about 20 and 5000 Hz). I would
like to be able to shift the phase of this wave by variable amounts. In
other words, if we consider the conditioned square wave as starting at 0
deg. is there a way to set up a circuit so that I can shift the wave
infinitely up to +/- 90 deg. (relative to the input wave) by turning a knob
(like a pot)?

Any responses are greatly appreciated.

Don

2. ### Rene TschaggelarGuest

Such as an allpass filter ?

Rene

3. ### Jim ThompsonGuest

You could make a variable *delay*, but anticipators are in short
supply ;-)

...Jim Thompson

4. ### Fred BloggsGuest

Since you are working with square waves, one approach would be a PLL
with a phase-shifting DC bias summed with the PD output driving a
PI-loop filter- this will force the VCO to advance/retard its phase
relative to your input reference to compensate for the bias- pot controlled.

5. ### Klaus Vestergaard KragelundGuest

Since you don't know the frequency of the signal, you can't adjust the phase
accordingly. If I had to do what you seem to wnat, I would use a
microcontroller. One part of the code determines the frequency semi real
time and adjusts the phase accordingly. May I ask why you need such a delay?

Cheers

Klaus

6. ### GenomeGuest

Perhaps you should learn the delicate art of sucking cock first. Having
mastered that one, from your own perspective, you might graduate to a bit of
the oral on your girl/boy friend or wife.

Hint, there are other parts that are part of the experience. Try shoving

Now you've learnt how to ask nicely you might try having a look at one of
those phase locked loop things, unless them's been discontinued, 4046 or
74HC4046.

Once you have figured out how to make one lock and be stable you might move
on and think what will happen if you dribble a DC offset into the input of
the VCO.

Make sense?

No?

Well, shouldn't have fucked your cluster on the first try.

DNA

7. ### Spehro PefhanyGuest

It would be easy enough to do digitally, but of course your estimate
of required edge timing would be, at best, as good as the most
measurement of the period and position of the signal edge. This is an
unusual requirement. What are you trying to do?

Best regards,
Spehro Pefhany

8. ### Jim ThompsonGuest

For square waves ?

...Jim Thompson

9. ### John WoodgateGuest

I read in sci.electronics.design that Don A. Gilmore
You can't phase-shift a square wave without using heroic means. What you
can do is to introduce a time-delay. A precision monostable with
adjustable period is one way of doing it.

10. ### Fred BloggsGuest

No- he has to keep DC away from the VCO input for that 250:1 range- he
has to apply the offset bias to an error integrator, and then to the
VCO- the VCO phase output is the integration of the control input
voltage offset from the DC value required to settle on the reference
frequency-and you don't know what that is....

11. ### Jim ThompsonGuest

On the S.E.D/Schematics page of my website, see:

CrudeDelay.pdf
EqualRiseFallEdgeDealy.pdf/P-N-Edge-OneShot
ExcessPhase-AllPass.pdf
SyncRingOsc.pdf

The SyncRingOsc could be embellished to tap outputs from the shift
register.

...Jim Thompson

12. ### GenomeGuest

No, no, no, no, no.

Mr PD goes so many volts per degree. Mr DC offset...... offsets it. The rest
of the bucket dunt matter..... caveat, as long as it's stable including the
method of offsetting.

I will not be having your babies!

DNA

13. ### Don A. GilmoreGuest

Hi John:

Hey, I like this idea. It's simple. I'm guessing that I can use a pot with
the timer to control the delay and thus the phase shift.

Is there a particular timer that you recommend? I take it an ordinary 555

Don

14. ### Fred BloggsGuest

It's simple- but not very good- the PLL makes the phase adjust frequency
independent- the timer does not. What kind of sensor is this anyway?-and
why do you need adjustable phase shift?

15. ### Fred BloggsGuest

Well- still working on the laser project- you condescendingly posted:

"Hi kids:

I'm trying to sense fairly fast (about 5000 Hz) pulses from a laser beam and
keep it cheap for a comsumer-type application."

Now for the +/- phase shift- not talking about your big secret design

16. ### John WoodgateGuest

I read in sci.electronics.design that Don A. Gilmore
For the frequencies you cite, a 555 may be OK, but to be sure a 4047 or
more modern device would be better.

17. ### James MeyerGuest

I would start with the original signal. If it isn't close to a sine
wave, I'd band-pass filter it to make it so. Then I'd pass the sine wave
through another filter that had an adjustable phase delay associated with it.

THEN I'd convert the signal to a square wave.

Jim

18. ### Mark ZenierGuest

is set by measuring the duration of the previous cycle.

(Hey, whatever the customer wants. ;-) )

Mark Zenier Washington State resident

19. ### BaphometGuest

I'm not sure you have provided sufficient information to effect a realistic
solution. For example, is the frequency of the input square wave
continuously changing? If so, how fast? Do you really need infinite
resolution of the phase shift (time delay) or if not, what resolution would
be acceptable? How much phase jitter can your circuit tolerate? It would
really be nice to know the end application because not knowing is like
shooting in the dark without night vision. Then again, perhaps I'm just not
understanding the problem as stated.

20. ### MikeGuest

I knew you were going to say that.

-- Mike --