Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Very simple phase detector

Status
Not open for further replies.

neazoi

Advanced Member level 6
Joined
Jan 5, 2008
Messages
4,119
Helped
13
Reputation
26
Reaction score
15
Trophy points
1,318
Location
Greece
Activity points
36,918

Attachments

  • Simple_Phase_Detector.gif
    Simple_Phase_Detector.gif
    79.6 KB · Views: 277

when the phases of VCO and Ref signal are same the effective voltage across base-emitter of transistor is nil.
A voltage exists when there is a phase difference.
This gets reflected at collector as error voltage.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
when the phases of VCO and Ref signal are same the effective voltage across base-emitter of transistor is nil.
A voltage exists when there is a phase difference.
This gets reflected at collector as error voltage.

Thanks a lot for the description!
So this depends two identical frequency signals phases to work (correct me if I am wrong).
I wonder if I feed in a comb generator instead of the reference oscillator, what will happen? Will the VCO be locked to each harmonic of the comb generator as you tune close to it, then as you tune far away locked to the second harmonic and so on ?
Can it really be accomplished by such a simple circuit?
 
Last edited:

that is what is called a "phase sampler". The reference clock momentarily turns on the transistor. When the transistor is turned ON, it charges the 0.01 uF cap, and then the transistor turns off. The voltage "stored" on the cap is an indication of the phase difference.

These type of phase samplers work better when schottky diodes are used to do the sampling, since they turn ON/OFF much faster than a bipolar transistor will, and give a more accurate detection output.

If you use this in a PLL circuit, you will find that the 0.01 uF capacitor causes a POLE, and the 2.2 uF/47 ohm combination causes a ZERO. In the open loop transfer function. Since the VCO also has a POLE, you are looking at a control loop with two poles and one zero...which should be fundamentally stable unless there are other parasitics, integrator op amps, or delays you are not showing
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
that is what is called a "phase sampler". The reference clock momentarily turns on the transistor. When the transistor is turned ON, it charges the 0.01 uF cap, and then the transistor turns off. The voltage "stored" on the cap is an indication of the phase difference.

These type of phase samplers work better when schottky diodes are used to do the sampling, since they turn ON/OFF much faster than a bipolar transistor will, and give a more accurate detection output.

If you use this in a PLL circuit, you will find that the 0.01 uF capacitor causes a POLE, and the 2.2 uF/47 ohm combination causes a ZERO. In the open loop transfer function. Since the VCO also has a POLE, you are looking at a control loop with two poles and one zero...which should be fundamentally stable unless there are other parasitics, integrator op amps, or delays you are not showing

All right,
So I guess the comb generator idea won't work, because all harmonics will try to switch on/off the transistor simultaneously?
 

a comb generator MAY be acceptable. If all those harmonics are in the right phase, they would form one big impulse voltage spike, which would turn the transistor on very nicely.

Put the comb generator output thru a 10 dB attenuator, and then look at it on an oscilloscope. Is it a voltage spike?
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
a comb generator MAY be acceptable. If all those harmonics are in the right phase, they would form one big impulse voltage spike, which would turn the transistor on very nicely.

Put the comb generator output thru a 10 dB attenuator, and then look at it on an oscilloscope. Is it a voltage spike?
But you want to compare the phase of a single reference signal with the phase of the vco. If the reference signal is actually a comb of frequencies, which one of them would you compare to? If it is the one closer to the vco at the current time, it makes sense. However Since this phase detector depends on the switching of the transistor, I really do not see how the pase of the harmonics can achieve that.
It would be helpful to know the mechanism behind this, what do you mean by voltage spike on the scope, you mean a square wave?
 

If the reference signal is actually a comb of frequencies, which one of them would you compare to?
The frequency which falls in the detector lock range. Because the lock range is very small (< 100 Hz with the present component values), it's probably difficult to get it locked at all.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
The frequency which falls in the detector lock range. Because the lock range is very small (< 100 Hz with the present component values), it's probably difficult to get it locked at all.
Why the lock range is only 100Hz or so in that circuit FvM?
 

Lock range is set by the operation principle of the phase detector and the loop filter bandwidth.

Why not evaluate the circuit behavior in a simulation or hardware test?
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
Lock range is set by the operation principle of the phase detector and the loop filter bandwidth.

Why not evaluate the circuit behavior in a simulation or hardware test?

I am not good at simulations and probably my results won't me valuable :(

It is the simplicity of this detector that drives me to try it in simple VFOs to see how well will it work or if it will work at all. The comb generator idea reminded me one of the parts of the wadley loop so I was thinking it could be done.
 

But you want to compare the phase of a single reference signal with the phase of the vco. If the reference signal is actually a comb of frequencies, which one of them would you compare to? If it is the one closer to the vco at the current time, it makes sense. However Since this phase detector depends on the switching of the transistor, I really do not see how the pase of the harmonics can achieve that.
It would be helpful to know the mechanism behind this, what do you mean by voltage spike on the scope, you mean a square wave?

your mind seems stuck thinking in the frequency domain. Try to think in the time domain instead!

try this paper:
 

Attachments

  • diode sampler.pdf
    142.2 KB · Views: 149
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
Lock range is set by the operation principle of the phase detector and the loop filter bandwidth.

Why not evaluate the circuit behavior in a simulation or hardware test?

Does this http://www.qrp4u.de/docs/en/frequenzanzeige/index.htm work in the same principle? In the sense that it compares the incoming signal with the reference (local oscillator) and produces a difference voltage.
 

The NE612 uses a multiplier principle rather than coincidence so the operation isn't quite the same.

Perhaps it might be easier to visualize it in the time domain by thinking of the signal in logic terms than frequency. There is lots of documentation on the operation of an XOR logic gate as a phase detector and also on the CD4046 operation which is similar. Thinking of signals as 'on' or 'off' may be easier to understand than the more abstract thinking needed to consider the interaction of sine waves.

Brian.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
The NE612 uses a multiplier principle rather than coincidence so the operation isn't quite the same.

Perhaps it might be easier to visualize it in the time domain by thinking of the signal in logic terms than frequency. There is lots of documentation on the operation of an XOR logic gate as a phase detector and also on the CD4046 operation which is similar. Thinking of signals as 'on' or 'off' may be easier to understand than the more abstract thinking needed to consider the interaction of sine waves.

Brian.

Thanks, I probably won't change anything in the phase detector but I will try it to see how will it work with the "harmonics" principle I thought.

A little bit off-topic but do you think I could replace the ne612 mixer in the analogue frequency meter posted above, with a simple bjt mixer followed by another stage of discrete sine-to-square converter? I worry about the LO suppression, which won't happen in the simple bjt mixer.

Or perhaps a variation of this one https://circuit-diagramz.com/discrete-frequency-voltage-converter/ is more straight forward?
Or the attached one?
 

Attachments

  • frerqconvert.JPG
    frerqconvert.JPG
    52.7 KB · Views: 243
Last edited:

Both those circuits will only work if the feed impedance is very low, at low frequency (< 10KHz) and with constant level square waves. Even variation in low/high ratio will change the result so they are not really practical for RF use.

As you correctly pointed out, any unbalanced mixer will produce the original frequencies, their sum and their differences simultaneously at the output so are not suitable for measuring with V to F converters. I have not checked for references but maybe searching for pulse counting detectors will help. They are a kind of FM discriminator using the rate of charge on a capacitor to change frequency to voltage, essentially a high frequency charge pump. Some old simple FM receivers used them but as they have almost no AM rejection their performance was generally poor. They do however have an output frequency response down to DC so you might be able to adapt as a medium range frequency meter.

Brian.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
I have not checked for references but maybe searching for pulse counting detectors will help. They are a kind of FM discriminator using the rate of charge on a capacitor to change frequency to voltage, essentially a high frequency charge pump. Some old simple FM receivers used them but as they have almost no AM rejection their performance was generally poor. They do however have an output frequency response down to DC so you might be able to adapt as a medium range frequency meter.

Brian.
Interesting! see these. I am not sure how they work and I cannot find info, but They could be used as frequency meters but also as vfo stabilizers maybe?
 

Attachments

  • 460khz Pulse Count Discriminator from RSGB  VHF UHF MANUAL 3rd Ed 1978.jpg
    460khz Pulse Count Discriminator from RSGB VHF UHF MANUAL 3rd Ed 1978.jpg
    70.4 KB · Views: 150
  • pulsecountingcircuitdiagramoffmdemodulator.jpg
    pulsecountingcircuitdiagramoffmdemodulator.jpg
    86.7 KB · Views: 231
  • Disc i.jpg
    Disc i.jpg
    30.7 KB · Views: 226

With some modification to remove the AC coupling at the output (these are really intended for audio recovery) they can be used as analog frequency meters but to be useful as a frequency stabilizer there has to be a reference to compare against. As pulse counting detectors work best at low frequencies, you could down-mix a higher frequency using a stable LO then use one of those circuits to produce an error voltage. To be honest, it wouldn't perform any better than a conventional PLL but it would be quite a lot more complicated.

For interest, when used in VHF FM receivers a zero (or near zero) IF was normally used so the station was shifted down to no more than a few KHz IF before being fed to the detector. Doing that made the detector easier to build and as down-mixing preserves the frequency deviation, it made the shift very large in relation to the background carrier.

Brian.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
With some modification to remove the AC coupling at the output (these are really intended for audio recovery) they can be used as analog frequency meters but to be useful as a frequency stabilizer there has to be a reference to compare against. As pulse counting detectors work best at low frequencies, you could down-mix a higher frequency using a stable LO then use one of those circuits to produce an error voltage. To be honest, it wouldn't perform any better than a conventional PLL but it would be quite a lot more complicated.

For interest, when used in VHF FM receivers a zero (or near zero) IF was normally used so the station was shifted down to no more than a few KHz IF before being fed to the detector. Doing that made the detector easier to build and as down-mixing preserves the frequency deviation, it made the shift very large in relation to the background carrier.

Brian.

The problem with down mixing is that you need a local oscillator. This means that in a multi-band RX you would need many local oscillators one for each band.
Now, I am not sure how do they work, but I guess they produce an AC (Amplidude variations?) when an FM signal is present in. This means that the dead carrier is used as a "reference" and any deviation is used as an "error" voltage.
I am sorry I am taking the conversation too far, but it would be interesting to investigate if they could be used as simple vfo stabilizers or frequency meters based on the voltage error.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top