Continue to Site

# how to design a phase generator

Status
Not open for further replies.

#### saif haider

##### Full Member level 1
i need to design a phase generator that could gave a phase shift of 90degree lead or lag any suggestions

#### Humungus

##### Full Member level 6
I do not remember the structure by heart, but there are some all-pass filters that allows you to change the phase of the output signal by changing the value of just one component. Make a search on them.

#### saif haider

##### Full Member level 1
Humungus said:
I do not remember the structure by heart, but there are some all-pass filters that allows you to change the phase of the output signal by changing the value of just one component. Make a search on them.
thanks for ur suggestion but already gone through but their is no use plz help me i really need to make it for my final year project

#### on1aag

##### Full Member level 3
Hi Saif Haider,

If you want to design a 90 degrees phase shift oscillator you gonna
need a circuit that can do that, an integrator ( or a differentiator)
will perform that task. But you also have to shift the phase even
further in order to get your signal back to 0 or 180 degrees.
So you need two of those and an inverting amplifier to get the ac
signal back to 0 or a multiple of 360 degrees. The trick is to get the
ac phase shift to 0 degrees and the dc "phase shift" to 180 degrees to
keep the circuit stable for dc.
Here's an example from the 1974 Nonlinear Circuits Handbook from
Analog Devices.

You can build this circuit with modern fet input opamps like the TL084.
Forget both four quadrant multipliers, they are only needed if you want
the frequency to be controllable by a dc voltage, they only act as voltage
controlled amplifiers. Both 1N935's are temperature compensated 9 volts
zenerdiodes but they should be connected in series, not in parallel.
If you want the frequency to be variable, replace both 10k resistors of
the integrators by a 1k resistor in series with a 10K (stereo)potentiometer
and you could switch the 15 nF capacitors to get different ranges.

My simulator says it's OK.

Vert. 5V/div Hor. 100µS/div

on1aag.

### saif haider

Points: 2

#### saif haider

##### Full Member level 1
on1aag said:
Hi Saif Haider,

If you want to design a 90 degrees phase shift oscillator you gonna
need a circuit that can do that, an integrator ( or a differentiator)
will perform that task. But you also have to shift the phase even
further in order to get your signal back to 0 or 180 degrees.
So you need two of those and an inverting amplifier to get the ac
signal back to 0 or a multiple of 360 degrees. The trick is to get the
ac phase shift to 0 degrees and the dc "phase shift" to 180 degrees to
keep the circuit stable for dc.
Here's an example from the 1974 Nonlinear Circuits Handbook from
Analog Devices.

You can build this circuit with modern fet input opamps like the TL084.
Forget both four quadrant multipliers, they are only needed if you want
the frequency to be controllable by a dc voltage, they only act as voltage
controlled amplifiers. Both 1N935's are temperature compensated 9 volts
zenerdiodes but they should be connected in series, not in parallel.
If you want the frequency to be variable, replace both 10k resistors of
the integrators by a 1k resistor in series with a 10K (stereo)potentiometer
and you could switch the 15 nF capacitors to get different ranges.

My simulator says it's OK.

Vert. 5V/div Hor. 100µS/div

on1aag.
could this circuit also be able to gave phase lag of 90 i mean -90degree

Last edited by a moderator:

#### on1aag

##### Full Member level 3
Hi Saif Haider,

Phase doesn't mean anything if it isn't referred to something else.
If there's a phase lag or lead will depend on what you assume to be
the reference.

on1aag.

#### saif haider

##### Full Member level 1
on1aag said:
Hi Saif Haider,

Phase doesn't mean anything if it isn't referred to something else.
If there's a phase lag or lead will depend on what you assume to be
the reference.

on1aag.
i know that phase shift means that you gave a lead or lag to current with refference to time with voltage but the general question is how this is possible, how to perform -90degree lag i know we us sin to cosine converter but how it perform in case of -90degree

#### xxargs

##### Full Member level 4
saif haider said:
on1aag said:
Hi Saif Haider,

on1aag.
could this circuit also be able to gave phase lag of 90 i mean -90degree

use 180 degree unit-amplifier (inverter) on 90 degree output and you make -90 degree output.

### saif haider

Points: 2

#### saif haider

##### Full Member level 1
xxargs said:
saif haider said:
on1aag said:
Hi Saif Haider,

on1aag.
could this circuit also be able to gave phase lag of 90 i mean -90degree

use 180 degree unit-amplifier (inverter) on 90 degree output and you make -90 degree output.

i want to know that either this circuit works on 418MHz or not if not then tell what shall be the best design for desired freq
with best regards

#### xxargs

##### Full Member level 4
saif haider said:
xxargs said:
saif haider said:
on1aag said:
Hi Saif Haider,

on1aag.
could this circuit also be able to gave phase lag of 90 i mean -90degree

use 180 degree unit-amplifier (inverter) on 90 degree output and you make -90 degree output.

i want to know that either this circuit works on 418MHz or not if not then tell what shall be the best design for desired freq
with best regards

You not talk about high frequency before, is hard to find OP-amp working well in this frequecy range and frequency depend componet going to very small value.

For rigth answer, you must give right information on your problem - and frequency range is very importent info to make right answer ie. totaly different solution if we talk low (< 1 MHz ) or high frequency (> 10 MHz)... if no frequency information given - answer is near always related to low frequency solution.

---

Other way in high frequency range is using single oscillator and split output to two way with 90-degree hybrid[1] and/or using different path length of transmissions line. For 418 MHz need only 179.4 mm (air way length) or around 122 mm RG223-cable different path-length to make 90 degree differense between output.

[1] see Anaren or Mini-circurit for 90-degree splitters/hybrids, you can also make splitters, coupler structures and delay-strip on PCB via microstrip etc.

### saif haider

Points: 2

#### saif haider

##### Full Member level 1
xxargs said:
saif haider said:
xxargs said:
saif haider said:
on1aag said:
Hi Saif Haider,

on1aag.
could this circuit also be able to gave phase lag of 90 i mean -90degree

use 180 degree unit-amplifier (inverter) on 90 degree output and you make -90 degree output.

i want to know that either this circuit works on 418MHz or not if not then tell what shall be the best design for desired freq
with best regards

You not talk about high frequency before, is hard to find OP-amp working well in this frequecy range and frequency depend componet going to very small value.

For rigth answer, you must give right information on your problem - and frequency range is very importent info to make right answer ie. totaly different solution if we talk low (< 1 MHz ) or high frequency (> 10 MHz)... if no frequency information given - answer is near always related to low frequency solution.

---

Other way in high frequency range is using single oscillator and split output to two way with 90-degree hybrid[1] and/or using different path length of transmissions line. For 418 MHz need only 179.4 mm (air way length) or around 122 mm RG223-cable different path-length to make 90 degree differense between output.

[1] see Anaren or Mini-circurit for 90-degree splitters/hybrids, you can also make splitters, coupler structures and delay-strip on PCB via microstrip etc.

i admit my mistake that i should mention my frequency first but what i want tell u is that i new to this field, i just have experience in books with block diagrams and nothing.
my project demands in-phase RF transmission of single and out of phase -90 and 90 degree phase shift i just know that it could be possible if i use inductor and capacitor in parrall, then i shall be able to gave that shift plz help me

bye the way thanks alot

#### saif haider

##### Full Member level 1
on1aag said:
Hi Saif Haider,

Phase doesn't mean anything if it isn't referred to something else.
If there's a phase lag or lead will depend on what you assume to be
the reference.

on1aag.

i want to know that either this circuit works on 418MHz or not if not then tell what shall be the best design for desired freq
with best regards

#### saif haider

##### Full Member level 1
on1aag said:
Hi Saif Haider,

If you want to design a 90 degrees phase shift oscillator you gonna
need a circuit that can do that, an integrator ( or a differentiator)
will perform that task. But you also have to shift the phase even
further in order to get your signal back to 0 or 180 degrees.
So you need two of those and an inverting amplifier to get the ac
signal back to 0 or a multiple of 360 degrees. The trick is to get the
ac phase shift to 0 degrees and the dc "phase shift" to 180 degrees to
keep the circuit stable for dc.
Here's an example from the 1974 Nonlinear Circuits Handbook from
Analog Devices.

You can build this circuit with modern fet input opamps like the TL084.
Forget both four quadrant multipliers, they are only needed if you want
the frequency to be controllable by a dc voltage, they only act as voltage
controlled amplifiers. Both 1N935's are temperature compensated 9 volts
zenerdiodes but they should be connected in series, not in parallel.
If you want the frequency to be variable, replace both 10k resistors of
the integrators by a 1k resistor in series with a 10K (stereo)potentiometer
and you could switch the 15 nF capacitors to get different ranges.

My simulator says it's OK.

Vert. 5V/div Hor. 100µS/div

on1aag.
hi plz do consider this it important
i am final year student and i have a project on smart antenna(adaptive phased array),i using frequency of 418MHz, the antenna is two dipoles of spacing of 3/8.

now the problem i am facing is in phase generator, i want to gave a phase shift of 90 degree, all i know is i have to use LC oscillator and then path length transmission line. i am new to this field having less idea about this, all these thing are quiet confusing. plz let me have ur views and suggestion
with regards
saif

#### bigpop

##### Junior Member level 3
use DLL which can generate any phase shift you wanted

#### saif haider

##### Full Member level 1
bigpop said:
use DLL which can generate any phase shift you wanted
let me know either it will work on 418mhz freq

#### bigpop

##### Junior Member level 3
yes, it can. fro example. you can design a delay cell with about 75ps delay time. use 32 delay cells and use DLL loop control whole delay time of 1/418Mhz. Then you can get 32 phase clocks

Status
Not open for further replies.