sine ouput from opamp

[featherheart9]

edited-
my ques is dat i am giving x as input (it can be ac nd dc both) to the circuit and i want sin(x) as output .
and i have to make dis circuit using opamps
i read sumwhere dat quadrature oscilator can give such output. but i am not able to understnd
so plz xplain and if their is anyother way then do tell me

(sorry for inconvenience)

 

[JimB]

i am confused so plz explain

So are we.

If you would like to rewrite your question in coherent English, you are more likely to receive a meaningful answer.

JimB

 

[alec_t]

What do you mean by 'sine of input'? Both amplitude and frequency have to be specified to define a sine wave.
An opamp plus phase-shift network can be used to make a sine oscillator.

 

[JimB]

Having just read alec_t's words I think that I get what the OP wants.

I think that is looking to build a "function block" where the input is a voltage representing 0 to 90 degrees angle, and the output is a voltage representing the sine of the angle.

Just an inspired guess you understand, but if the OP returns, maybe he can elucidate.

JimB

 

[unclejed613]

yeah, i'm not sure what the OP is getting at either. if he's looking for a 90 degree phase shift in an AC signal between input and output, an integrator or differentiator is what he wants. from what i can decipher from his question, he's looking at the analog math functions that an op amp can perform. it used to be, when a manufacturer printed data books, you could find everything in the data book. it often took a bit of searching, but it was in there somewhere. other great sources of information were books like "The Op Amp Cookbook" by Walt Jung, and other similar books. getting sine and cosine outputs is easy from a multistage oscillator. but if it's a "static" sine or cosine function (input a DC level, get a DC level out that is a sine or cosine result) takes a bit more complexity, usually using piecewise linear approximation. most of the functions the OP mentioned in his post are using op amps as an analog computing device. maybe if the OP is still paying attention to this thread, he can clarify his question.....

 

[MrAl]

Hello there,

Yes it appears that he wants a 'sine' lookup in analog form. Normally we would see this as a sine lookup table in memory as in a microcontroller or something like that, or just a PROM programmed with the sine values.
So yes we would see something like these examples:
In: 0, Out: 0 (0 degrees)
In: pi/6, Out: 0.5 (30 degrees)
In: pi/4, Out: 0.7071 (45 degrees)
In: pi/2, Out: 1 (90 degrees)

This could be done with some accuracy but besides an op amp it would require (for one example) circuits that could raise the input to several different powers, like 3, 5, 7, etc., and invert every other one, so that we could use a power series to generate the sine of the input. So it would take several log/antilog circuits as well as the op amp to do the summing of all the terms and provide the sine output. Not sure if it would be worth the effort or not.

Then there is also the non linear solution, where we use a non linear circuit made with transistors or diodes to taper the input down via non linear gain to the right value that would equal the sine of the input. This could be done with several transistors NPN and PNP types. It would be an approximation to the real sine value just like the power series above would be so it would not be perfect but it would work.

 

[featherheart9]

Having just read alec_t's words I think that I get what the OP wants.

I think that is looking to build a "function block" where the input is a voltage representing 0 to 90 degrees angle, and the output is a voltage representing the sine of the angle.

Just an inspired guess you understand, but if the OP returns, maybe he can elucidate.

JimB

yes this was my question
and this can be implemntd using sine in euler's notation..alhtough it will have its own practicla limit and its design will also be very complicated.
but i was looking for theoritical approach only
Thanks

 

[MrAl]

Hello again,

Wow, it has been a long time since you replied to this thread.

So you want a sine output from the input, but what exactly will the input be?
If you just want a 90 degree phase shift, that's not too hard, but if you want a sine wave from a DC input that's harder to do.

So try to specify what input you are using, and if it could be different then show the way you want to convert it into whatever you want on the output. Show as many cases as possible.

For example,
1vac input, 1vac output 90 degrees
or
1vdc input, 1vac output, then 2vdc input, 2vac output.

The more detail you provide here the more we can help.

 

[misterT]

i am giving x as input to the circuit and i want sin(x) as output.

...but i was looking for theoritical approach only

One (theoretical) approach would be to generate a sinewave, then with a delay proportional to the input signal, trigger sample-and-hold to take a "snapshot" of the sine-signals value.

Good thing about this approach is that all the individual "blocks" in the circuits are relatively simple and well known. Difficult part may be getting them to work well together to achieve "good enough" performance.. whatever that is.

Or this idea:
**broken link removed**

 

[MrAl]

Hi MisterT,

I always meant to build one of those non linear waveshapers, but never got around to it. I figured it would be faster than those ready made chips for sale like the one that is (or was) made by Maxim. One of these days.

 

[featherheart9]

Hello again,

Wow, it has been a long time since you replied to this thread.

So you want a sine output from the input, but what exactly will the input be?
If you just want a 90 degree phase shift, that's not too hard, but if you want a sine wave from a DC input that's harder to do.

So try to specify what input you are using, and if it could be different then show the way you want to convert it into whatever you want on the output. Show as many cases as possible.

For example,
1vac input, 1vac output 90 degrees
or
1vdc input, 1vac output, then 2vdc input, 2vac output.

The more detail you provide here the more we can help.

ok here is what i am thinking
let input be signal with varying voltage level..
say v(1sec) =1 v
v(2sec) =3 v
v(3sec) =5 v nd so on
then i want output as
V(1sec) =sin(1) v
V(2sec) =sin(3) v
V(3sec) =sin(5) v
let x be input
and we can design a ckt as

x ---> [inductor] --> [exponent ckt using opamp] --> A
x ---> [capacitor] --> [exponent ckt using opamp] -->B

A and B as input to | -->[subtractor using opamp] --> [inductor] ---> output

we need to adjust gain at each level..
I think this will work...

 

[featherheart9]

One (theoretical) approach would be to generate a sinewave, then with a delay proportional to the input signal, trigger sample-and-hold to take a "snapshot" of the sine-signals value.

Good thing about this approach is that all the individual "blocks" in the circuits are relatively simple and well known. Difficult part may be getting them to work well together to achieve "good enough" performance.. whatever that is.

Or this idea:
**broken link removed**

well this second one will definelty work
and about your first approach I am bit confused can you please explain it

 

[misterT]

about your first approach I am bit confused can you please explain it

I don't know if it is worth explaining. Kind of "funny" approach.. first thing that came to my mind. Maybe I draw a picture of it when I get home

 

[MrAl]

ok here is what i am thinking
let input be signal with varying voltage level..
say v(1sec) =1 v
v(2sec) =3 v
v(3sec) =5 v nd so on
then i want output as
V(1sec) =sin(1) v
V(2sec) =sin(3) v
V(3sec) =sin(5) v
let x be input
and we can design a ckt as

x ---> [inductor] --> [exponent ckt using opamp] --> A
x ---> [capacitor] --> [exponent ckt using opamp] -->B

A and B as input to | -->[subtractor using opamp] --> [inductor] ---> output

we need to adjust gain at each level..
I think this will work...

Hello again,

Oh ok so you just want a circuit that can take the sine of the input voltage level. I would guess that you want the input to be in radians because in degrees it would have to be a very high voltage.

So for an input of 1.571 volts approximately you want an output of 1 volt.
For an input of 3.14 volts approximately you want an output of 0 volts.
For an input of 0.785 volts approximately you want an output of 0.707 volts approximately.

Probably the best approach is the waveshaper or microcontroller. The max input could be 3.14159 volts, and the min 0 volts unless you want negative input levels too and then the min could be -3.14159 volts.

 

[featherheart9]

Hello again,

Oh ok so you just want a circuit that can take the sine of the input voltage level. I would guess that you want the input to be in radians because in degrees it would have to be a very high voltage.

So for an input of 1.571 volts approximately you want an output of 1 volt.
For an input of 3.14 volts approximately you want an output of 0 volts.
For an input of 0.785 volts approximately you want an output of 0.707 volts approximately.

Probably the best approach is the waveshaper or microcontroller. The max input could be 3.14159 volts, and the min 0 volts unless you want negative input levels too and then the min could be -3.14159 volts.

yes microcontroller method will be best..but th e question was implement using opamp..
Thanks a lot for your help

 

[featherheart9]

I don't know if it is worth explaining. Kind of "funny" approach.. first thing that came to my mind. Maybe I draw a picture of it when I get home

Ok I will be waiting for the diagram..and I am sure its worth sharing here
we can get more improvements in it

 

[MrAl]

yes microcontroller method will be best..but th e question was implement using opamp..
Thanks a lot for your help

Ok sure, but you have to keep in mind that the input will be limited even though the sine function argument is not. For example, sin(0)=0 but also sin(pi)=0, and sin(2*pi)=0, and sin(100*pi)=0, so an idea would be to limit the input to plus and minus pi or just +pi.

Most likely then the waveshaper is the best approach. That's a purely analog approach. There will be some error in the result, but that's life.

 

[featherheart9]

Ok sure, but you have to keep in mind that the input will be limited even though the sine function argument is not. For example, sin(0)=0 but also sin(pi)=0, and sin(2*pi)=0, and sin(100*pi)=0, so an idea would be to limit the input to plus and minus pi or just +pi.

Most likely then the waveshaper is the best approach. That's a purely analog approach. There will be some error in the result, but that's life.

yes input must be limited..
Thanx for your help

@bold nd underlined - I completely agree

 

[MrAl]

Hi,

He he, yes, he he. The error will depend on how much effort and how many parts you care to use in the wave shape circuit. I thnk the more diodes or transistors you use the better you can get the wave shape, or in this case, the calculation.
Normally the wave shape circuit takes a ramp input (triangle really) and converts it into a sine wave with some small error. We could look up some chip part numbers that use this method if you would like to look at a few.

 

[featherheart9]

Hi,

He he, yes, he he. The error will depend on how much effort and how many parts you care to use in the wave shape circuit. I thnk the more diodes or transistors you use the better you can get the wave shape, or in this case, the calculation.
Normally the wave shape circuit takes a ramp input (triangle really) and converts it into a sine wave with some small error. We could look up some chip part numbers that use this method if you would like to look at a few.

Thanks for all your help..
I think I might try the wave shapper method..will work on it and will reach you if I have any problem or unwanted result there