Variable Sine Wave Generator Design

[Frosty_47]

Hello,

I was wondering if it is possible to design a sinewave oscilator that can be tunned from 200Khz - 2Mhz by simply adjusting a potentiometer. I am completly tired of lame circuits that use dozen switches to select arrays of capacitors in order to change the range. Also, I have no intereset in "premade" IC based function generators such as MAX038. Further I refuse to use any IC based PLL or VCO. There must be a solution out there, please help. And please no Varactors either.

-Thanks

 

[Sceadwian]

Great way to shoot yourself in the foot Frosty, you just said 'no you can't use that' to most sane approaches... If there were such a simple solution don't you think a schematic for it would be on the internet already?

 

[Mr RB]

The guy holding an AK47 shot himself in the foot??? That's gotta hurt!

Now just to get back on topic... Actually I don't think 200kHz -> 2MHz is impossible. Since it's only a 1:10 freq range you might be able to use some type of phase shift oscillator that makes a sine even though it is reduced amplitude at higher frequencies, then use a simple automatic gain controller to keep a constant output amplitude regardless of frequency. I bet you could do it with one pot and a handful of transistors.

 

[Sceadwian]

With a low distortion output? You really think you can get a decent sine wave out with that kind of range? My interest is peaked, I wouldn't mind building something like that.

 

[MrAl]

Hi there,

There are a couple chips that go from 0.001 Hz to 1 MHz, but there is one chip
that goes from low Hertz to 20 MHz too. Cant remember the part number but i'll
find it if no one else can. They put out square/triangle/sine.
These are all sine synth chips that use analog circuitry to shape the sine output.
They start with a square wave, convert to triangle, then to sine.
The sine output isnt bad for some chips.

Maybe the MAX038 ?

 

[Frosty_47]

I found a solution...
I will use a variable four quadrant gyrator to adjust the frequency of oscilation.
This way, I only have to adjusty a single resistor to change the oscilation.
Gyrator + quality op-amp from Burr-Brown = very little noise
I will post results when I get the circuit to work...

 

[Frosty_47]

Hi there,

There are a couple chips that go from 0.001 Hz to 1 MHz, but there is one chip
that goes from low Hertz to 20 MHz too. Cant remember the part number but i'll
find it if no one else can. They put out square/triangle/sine.
These are all sine synth chips that use analog circuitry to shape the sine output.
They start with a square wave, convert to triangle, then to sine.
The sine output isnt bad for some chips.

Maybe the MAX038 ?

Did you just post that to piss me off?
Or did you not read my first post above?

 

[Frosty_47]

Great way to shoot yourself in the foot Frosty, you just said 'no you can't use that' to most sane approaches... If there were such a simple solution don't you think a schematic for it would be on the internet already?

No i don't. because a simple working solution would probably never make it to the textbooks and especialy internet... It would most likely be patentted by some ashole that will ask for lots of money for publishing his work.

 

[Sceadwian]

Frosty, you obviously have no comprehension of what you're asking...There are 855 transistors in the Max038, you want miracles from a handful of components? Applications like this are the entire reason why IC's exist in the first place! What you want you can't get from what you're providing as tools for the solution. Solution, use better tools...

 

[BrownOut]

Here is the one I built. It produces a square, triangle, and low-distortion sine wave. It will run from a few hertz to nearly 1MHZ. If you replace the 555 with a fast comparator circuit, you might make 2MHZ.

 

[Mr RB]

With a low distortion output? You really think you can get a decent sine wave out with that kind of range? My interest is peaked, I wouldn't mind building something like that.

Well his spec never said "low distortion"...

Brownout- Nice circuit! Can you give a bit of explanation on how the sinewave is generated from integrating the triangle wave or whater it does?

 

[Speakerguy]

Any SPI capable PIC interfaced to an AD9833 (direct digital synthesis chip). Use a pot into an ADC on the PIC to set the frequency. Problem solved.

 

[MrAl]

(With reference to the MAX038 sine synthesis chip)

Did you just post that to piss me off?
Or did you not read my first post above?

Hi there Frosty,


Oh gee, ha ha, i read your post too fast i guess and didnt see that
you were not interested in that chip. I guess i could ask then:
Why are you not interested in that chip or chips like it?

Hope i didnt heat you up too much there Mr Frosty, wouldnt want
you to melt all over the sidewalk :smile:

Good luck in your search.

 

[dark]

MAX038 or whatever stand alone sinusoid generator ICs I came across in the past are all on EOL finished . Is there any chip that is still in Full production ?

BrownOut - can you please explain your circuit a bit . It looks like it has to be on a 4-layer board andwith good layout . Can one replace 555time with some other stable square wave generator probably a XTAL controlled one? or it would be an overkill.

Thanks

 

[dark]

Did you just post that to piss me off?
Or did you not read my first post above?

Please cut profianty.

-

 

[MrAl]

Hi,

They are still available in some places, but wow they are over 40 dollars USD.

 

[BrownOut]

Well his spec never said "low distortion"...

Brownout- Nice circuit! Can you give a bit of explanation on how the sinewave is generated from integrating the triangle wave or whater it does?

The sine wave generator comprises Q5 and Q6. It amounts to a emitter coupled pair that is driven with a very linear triangle wave centered at 0V. The current of Q5 is given by:

IE(Q5)= I/(1 + e^VB(Q5)). Given sufficient base voltage, the equation approximates a sine wave to a very close tolerance. My circuit produces a slight deveation from the equation because I bias the transistors with a resistor rather than a current source (I). I've attached a graph that shows a sine wave for the interval -pi/2 to pi/2 and a plot of the given equation appropriately scaled so that the two plots superimpose. The circuit is operated back and forth between -pi/2 and pi/2 to produce the sine wave.

The circuit is still going through some refinement. I need to test the stability of the sine wave generator over the expected operating temperature range.

Q1, Q3 and C1 comprise the triangle wave generator, while Q2 and Q4 buffers and inverts it. Q7 and Q8 shift the level so that the triangle wave is centered at 0V. R30 makes a fine level adjustment. R22 and R19 sets the level of the triangle wave to produce the best sinewave from the last stage. R2/R5 are set to compensate for the 555's asymmetrical output, crude compensation, but it works. A180 degree sine may be taken from Q6.

 

[hdc090360]

Frosty,

Is this the sort of simple thing you are looking for?

**broken link removed**

There is no mention of bandwidth limits, but that is a matter of careful component selection.

 

[Sceadwian]

Although I'm not a particularly good analog person I don't think something with such a low component count could possible be very linear through that wide a range. There would have to be some kind of variable capacitance to keep the distortion sane from low to high frequencies, you're talking either a varactor or ganged knobs.

 

[MrAl]

Hi again,


There is another way to shape the sine wave from the triangle wave. It requires
something like 5 pnp and 5 npn transistors. If not anything else it is interesting
but im not sure if it appears on data sheets any longer. It was originally in
the data sheet for the 8038 chip i think, a long time ago. The 5 transistors
(5 for positive half cycle, 5 for negative half) act as 5 breakpoints giving
quite a bit of control over the output wave shape.