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.

Active HF phase shifter?

Status
Not open for further replies.

neazoi

Advanced Member level 6
Joined
Jan 5, 2008
Messages
4,118
Helped
13
Reputation
26
Reaction score
15
Trophy points
1,318
Location
Greece
Activity points
36,907
This circuit comes from a national semiconductor application note (AN-32).

Is it suitable for HF usage (1-30MHz)?
If not as it is, what components values should I change? (I will use a J310 instead)
 

Attachments

  • Capture.PNG
    Capture.PNG
    49.3 KB · Views: 144

It will work but the limitations of the components, particularly the potentiometer will seriously degrade it as the frequency increases. The principle is to use the first and second stages as phase splitters, the drain and source signals will be (theoretically) inverted. One is then fed through a capacitor to provide fixed further shift and the other through a variable path, the variable resistance. The point where they meet will sum a waveform from the difference of the signals. The main complication is the poor passage of HF through the potentiometer into the load capacitance. You probably need to scale the coupling capacitor down to a much lower value, maybe a few 10's of pF and the potentiometer down to 1K or less to make it effective. It will also null the frequency when R=XC so expect the output level to be very variable as the control is turned.

The circuit will work best at low frequencies and is often used to produce 'hollow' or spacial sound effects in music. For RF it will still work but not very well.

Brian.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
It will work but the limitations of the components, particularly the potentiometer will seriously degrade it as the frequency increases. The principle is to use the first and second stages as phase splitters, the drain and source signals will be (theoretically) inverted. One is then fed through a capacitor to provide fixed further shift and the other through a variable path, the variable resistance. The point where they meet will sum a waveform from the difference of the signals. The main complication is the poor passage of HF through the potentiometer into the load capacitance. You probably need to scale the coupling capacitor down to a much lower value, maybe a few 10's of pF and the potentiometer down to 1K or less to make it effective. It will also null the frequency when R=XC so expect the output level to be very variable as the control is turned.

The circuit will work best at low frequencies and is often used to produce 'hollow' or spacial sound effects in music. For RF it will still work but not very well.

Brian.

Thanks. I will give it a try. I thought the potentiometer and capacitor values were calculated for the required phase shift. Ok it is nice to hear that it can work with other values too.
It will be tested to see how effective it will be in a 1-30MHz phasing receiver, as the LO phaser. In that aspect:

1. Assuming a 360 degrees phase shift, could I avoid the USB/LSB switch (phase pair select) in the audio path of the receiver and just tune the LO phase in one of the mixers, so that it lags or perceeds the other mixer LO? Will this provide the same USB/LSB selection or not?

2. Since I thought also that there will be amplitude variations when phase is adjusted, I thought to include some kind of detection (envelope) on both LO parts (phased and not phased) and manually, or somehow automatically adjust both LOparts to be of equal amplitude. Any ideas would be greatly helpful.

3. Since only 90 degrees of difference are needed, I could probably remove the second stage?
 

adjust both LOparts to be of equal amplitude.

By adding an inductor to the capacitor, attenuation is not as severe. Because you don't have a resistor involved. For 1-30 MHz range an inductor consists of a few turns around a pencil.

3. Since only 90 degrees of difference are needed, I could probably remove the second stage?

A tunable inductor and/or capacitor could make things simpler.

There is more than one way to arrange the LC network (depending on whether you want phase advance or delay), and a potentiometer is still an option. It requires experimentation with a few configurations.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
By adding an inductor to the capacitor, attenuation is not as severe. Because you don't have a resistor involved. For 1-30 MHz range an inductor consists of a few turns around a pencil.



A tunable inductor and/or capacitor could make things simpler.

There is more than one way to arrange the LC network (depending on whether you want phase advance or delay), and a potentiometer is still an option. It requires experimentation with a few configurations.

You mean replacing the resistor with an inductor and the capavitor with a variable one?

How about leaving the resistor as it is and replacing the capacitor with a variable one? This won't degrade the performance due to the potentiometer as Brian mentioned.
I think that the LC instead of the RC will have the benefit of keeping the amplitude more stable. I have read that for PI LC based phase shift networks compared to RC. But the simplicity of the RC is great.
But I am not sure if fixed values of C and one value of potentiometer, will be suitable for the whole HF or if I need to change these depend on the band.
Unfortunately, until I fix my scope I cannot do any measurements, so I am blind...
 

The all-pass circuit implemented in post #1 does ideally not involve magnitude variations, as longs as R >> Rs, Rd.

If your application is fixed 90° phase shift over 1 - 30 MHz frequency range, I would prefer a solution that must be not tuned to frequency.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
The all-pass circuit implemented in post #1 does ideally not involve magnitude variations, as longs as R >> Rs, Rd.

What does it mean FvM? That with the values shown it won't have amplitude variations? Wave I understood that right?
 

You mean replacing the resistor with an inductor and the capavitor with a variable one?

It's an idea worth some experimentation.
Or to have a potentiometer choose how much comes from a capacitor and how much from an inductor.
This simulation selects a range between phase advance or retard.
The potentiometer was dialed from left to right during the run. The scope trace tells the story. Notice at the beginning, output phase leads the source. By the end, output lags.

pot select L C phase advance or retard 5 MHz.png

These values seemed suitable for 5 MHz. They were chosen to attenuate the source 5V amplitude about half. To select values your own project 1-30 MHz may require some trade-offs in performance.

There is another possibility: make a series LC. Ground one end, apply your signal to the other end. Tap for output between the L & C.
Reverse the LC, to obtain different behavior.
 
  • Like
Reactions: neazoi

    neazoi

    Points: 2
    Helpful Answer Positive Rating
It's an idea worth some experimentation.
Or to have a potentiometer choose how much comes from a capacitor and how much from an inductor.
This simulation selects a range between phase advance or retard.
The potentiometer was dialed from left to right during the run. The scope trace tells the story. Notice at the beginning, output phase leads the source. By the end, output lags.

View attachment 155050

These values seemed suitable for 5 MHz. They were chosen to attenuate the source 5V amplitude about half. To select values your own project 1-30 MHz may require some trade-offs in performance.

There is another possibility: make a series LC. Ground one end, apply your signal to the other end. Tap for output between the L & C.
Reverse the LC, to obtain different behavior.

Thanks very much for the interesting simulation and ideas!
I think I will stick to the RC circuit and try to correct any amplitude difference on later stages by ACG amplifiers or by a level comparator with envelope detector.
If I would use an LC I would go for a simple power divider, followed by a PI LC filter at one end. This wives good performance but it is not broadband.

What Intrigues me in the circuit in post #1 is to determine if it is broadband (1-30MHz) accepting amplitude variations as a tradeoff and try to correct them later on.
 

The phase shift in your original post depends on the reactance of the capacitor so it varies with frequency. If you manage to find suitable values it will still need 'tuning' to each frequency to get the required shift. Note that AGC will help to stabilize the output level but there is theoretical zero in the output when the resistor is adjusted to some point and you won't be able to compensate for that.

Brian.
 

The phase shift in your original post depends on the reactance of the capacitor so it varies with frequency. If you manage to find suitable values it will still need 'tuning' to each frequency to get the required shift. Note that AGC will help to stabilize the output level but there is theoretical zero in the output when the resistor is adjusted to some point and you won't be able to compensate for that.

Brian.

I do not mind tuning for this application. But having to change the capacitor for each band is undesirable. Also having to tune both the phase AND the amplitude is undesirable, as it will be too difficult on-the-fly without any test gear.
Even with an AGC amplifier some amplitude tuning may be needed, I do not know how accurate this will be. However just a little would be required. Without any amplitude tuning I should be able to reject some sideband bu just tuning the phase. Then give it a fine amplitude tuning to reject that for good.
That is the idea.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top