I need the schematic of a very simple sinusoidal signal generator

[jpm83]

Thanks for answering I've been planning to do a function generator of some sort, and now I have decided to make this for square/sine waves. http://sound.westhost.com/project86.htm And this for saw waves. http://www.electronics-lab.com/projects/test/010/

Janne

[gez]

Sorry, slip of the brain.  C is in Farads.  So 0.47uF (for example) would be entered into your calculator as .47 X 10 to the minus six.

[jpm83]

Did I get this right? Rt 47k, R1 100R, R2 100k, C 0,47µf =f=1131,73 Hz
http://www.play-hookey.com/analog/triangle_waveform_generator.html

Janne

[gez]

Did I get this right? Rt 47k, R1 100R, R2 100k, C 0,47µf =f=1131,73 Hz
http://www.play-hookey.com/analog/triangle_waveform_generator.html

Janne

Assuming that "0,47uf" = 0.47uF, then yes, the frequency would be around the 11kHz mark.  However, the amplitude of the waveform would be minuscule with only a 100R resistor for R1.

The way to set up these oscillators is to decide upon the desired amplitude first, then calculate the values of R1 and R2 to achieve this.  Rearanging the formula for amplitude you get:

R1 = (Amplitude X R2)/Vcc

It's also better to use values well above 100R for Rt, and make the cap smaller instead.  You'd be better off just adapting a schematic that is already in use.  Just make the cap smaller (the required value can be calculated by rearranging the formula for frequency).

[jpm83]

Would this be worth breadboarding?

By jpm83 at 2008-03-03

Janne

[jpm83]

And this was the starting point. http://www.electronics-lab.com/projects/test/010/

Janne

[gez]

Would this be worth breadboarding?

By jpm83 at 2008-03-03

Janne

Sure, but it could do with a stop resistor in series with the 10k pot.  Making the pot larger will give you a greater range, although everything will bunch up at one end if you make it too large.  There are fixes for this (see Flind's book - it's dirt cheap - for examples), but the circuit takes on a greater level of complexity.  For a higher frequency out, make the .47u cap smaller.

[jpm83]

Thanks for the replies. I will play around with this in next week, when my holiday ends and I get back to my flat.

Janne

[gigimarga]

thx all...i will try this weekend!!

[ErikMiller]

Here's one I built years ago, and still use in my shop. Jameco sells the main IC for under $5. Works great from a 9V battery.

[jpm83]

Would this be worth breadboarding?

By jpm83 at 2008-03-03

Janne

Well I played around with this today and 4,7nf C1 & 250k R1 gave me this at 2530,00 Hz.


Janne

[jpm83]

Forgot the question part. So what do you think is this usable?

Janne

[JDoyle]

When I tune a fuzz, I prefer to use a triangle wave.

My thoughts exactly! AND I think it is more representative of a guitar signal BECAUSE it has the peaks. A sine is representative of a sine and a sine alone. Very very few things output a natural sine wave and in the end - a sine wave sounds boring.

And if you do make a triangle generator - you can run it to half a LM13700 OTA arranged as a simple amplifier & if you overdrive it by just the right amount, you get (very close to) a sine! Certainly close enough for rock and roll tweaking.

I believe the 'right amount' is 78mVp-p if I am remembering right. This is the way that almost every write up I have read makes a sine wave. Though it is the front end differential amplifier that does the overdriving, it just so happens that with a 13700 or 3080 it is already built and ready for you...

Of course, in the final analysis, a musician & a guitar is what you need. And ears.

Amen to that!!!!!!

[gez]

Forgot the question part. So what do you think is this usable?

Looks odd.  What op-amp are you using?  Try reducing the 47k resistor (RT) untill things look cleaner.  This will reduce amplitude, but by the look of things you need to do this.  Increase the cap value to compensate if necessary.

[Paul Perry (Frostwave)]

Janne, is that a fairly low samplerate software scope or USB cro? I can't see why the circuit would have given an output with shape varying from cycle to cycle.

[gez]

Janne, is that a fairly low samplerate software scope or USB cro? I can't see why the circuit would have given an output with shape varying from cycle to cycle.

It could be that the amp's input range is being exceeded on the Schmidt and it's causing some weirdness in the whole amp?  Or there's slight clipping in the output of the integrator - though it would latch if that were the case, so I doubt it.  The way to be certain would be to reduce the 47K to reduce amplitude and see if things clean up.  If not, then yeah, it must be the scope.

[jpm83]

The opamp is HA17458 and program is visual analyzer my hardware was GEOFEX audio probe to Alto S-6 mixing console to M-Audio Delta Audiophile 2496. I Will have to try reducing the amplitude and see what happens.

Janne

[gez]

OK, page 6 of the data sheet (output swing vs supply voltage):

http://www.ortodoxism.ro/datasheets/HitachiSemiconductor/mXwvwss.pdf

At 9V supply the available output swing is close to the peak-to-peak amplitude shown by your scope.  The distortion is probably that caused by the amp saturating as it approaches its limits.  Presumably there's still enough travel when saturated to be able to trigger the Schmidt and prevent latching.

Definitely reduce the 47k and hopefully things will clean up.

[Paul Perry (Frostwave)]

What worries me about the screenshot, is that the 'distortion' - which may either be a genuine representation, or else an artifact of the measuring equipment - is different, from one cycle to the next.
That makes me pretty suspicious

[gez]

What worries me about the screenshot, is that the 'distortion' - which may either be a genuine representation, or else an artifact of the measuring equipment - is different, from one cycle to the next.
That makes me pretty suspicious

This could well explain it, Paul:

Well I played around with this today and 4,7nf C1 & 250k R1 gave me this at 2530,00 Hz

(my highlight).

About 10X (easily) the frequency you'd want, unless that comma is supposed to be full stop?  Would need to see the other resistor values used to do the calculations.