Variable Frequency and Amplitude Oscillator Circuit

[strahd_von_zarovich]

Hi Everyone,
I am looking for an oscillator topology or IC which can produce sinusoidal signals at frequencies between 100kHz - 2MHz, and amplitude of up to 2Vpp. Current capability of circuit and amplitude of signals are not big problems, since I can solve these problems with amplifiers.

I have very limited space so I can't use lots of components and I can't use digital ICs which need to be programmed. If I can adjust frequency with couple of resistors, capacitors etc., it would be very good solution for me.

I tried Wien Bridge and Phase Shift Oscillator topologies but I didn't like them very much. They are not very stable and as far as I can see they are not very suitable for generating signals at MHz.

My circuit works with Wien Bridge oscillator, but it degrades the performance. So, probably any IC or topology which can produce good frequency and amplitude stability would be enough. I am not looking for PPM stability.

Could you please suggest me a topology or IC(like ICL8038) or any solution for this purpose.

Thanks in advance.

 

[d123]

Hi,

No doubt you've seen this version in design note, but just in case.

Last line of design note, read it the other day too elsewhere - possibly in other pdf with surname of engineer as title, an unfortunate namesake of Donald... slyt701 is pretty good - succinct and clear about issues and solutions.

A couple of years ago I breadboarded same/similar whilst looking for good sine wave - results were ugly (skewed) and my implementation tended to die when changing frequency. In retrospect, besides less experience and ingormation than now, I suspect I wasn't using an OA with x10 bandwidth of sine wave frequency.

OPA322 (RRIO, 1.8 - 5.5V, input bias cancellation) allegedly good to go at G = 1 up to 10M, according to datasheet graphs, recollect G starts to droop after 1MHz.

Joke: Use any suitable clock, three CD4017s with diode + resistor ladders, smoothed steps to curve by C.

I'm curious as to why my OA sine wave experience sounds very similar to yours.

 

[danadakk]

Questions :

1) What controls the freq, a V input or pot or V controlled freq or rotary encoder ?
2) What is freq resolution desired ?
3) What is freq accuracy desired ?
4) What is stability over T and V needed ?


Regards, Dana.

 

[BradtheRad]

Sine-like wave obtained from shift register consisting of several D flip-flops. Resistors at the outputs are weighted to create desired volt levels.

(Spun-off from circuit found in menu of Falstad's animated interactive simulator.)falstad.com/circuit

A negative supply pulls the output down if you wish bipolar AC output. (Or use a series capacitor.)

To smooth the output you can add capacitive smoothing...
Or attach additional flip-flop stages for better resolution.

 

[strahd_von_zarovich]

Questions :

1) What controls the freq, a V input or pot or V controlled freq or rotary encoder ?
2) What is freq resolution desired ?
3) What is freq accuracy desired ?
4) What is stability over T and V needed ?


Regards, Dana.

Dear Dana,
Thanks for your answer.

1. If I can set the frequency by replacing components, then it doesn't matter. I don't want a signal generator circuit which I can set it's frequency and amplitude whenever I want. I will set the frequency to a value then I won't change it.

2. 1kHz is enough, but 10kHz even 100kHz is highly acceptable.

3. Frequency accuracy is not very important. If I want 500kHz sinusoidal like signal, and the real frequency is 490kHz It doesn't really matter.

4. This is the most important aspect for me. I just want a signal with constant frequency and constant amplitude. Some kind of VCO which I can set it's frequency with components would be excellent solution for me. But I couldn't find a suitable one, since their frequency range is pretty high.

 

[danadakk]

Bradthe Rads approach needs a minor change, on power up reset
all flops need to be set to 0. Note his clock must be changed to set
freq, and it follows :

freq = 10 x Freqdesired

freqmax = 10 x 2 Mhz = 20 Mhz.

Also using a LPF with a simple RC means that amplitude varies widely
over the freq range unless that is also made to track freq. And simple
LPF leaves sine distortion fairly high unless the number bits increased,
which leads to higher clk freqs as well.

One approach I see is to use two chips, a DDS and a xtal and a simple
part, like ATTINY85, to control it and interface to you and settings of
amplitude and freq, controlled by a pot or V you input to ATTINY85
for setting freq and amplitude.

If you do not need changeable V, then there is a DDS that takes a
parallel input, from some dip switches. AD9910, but its pretty pricey.
AD9850 also can be operated in parallel mode, lower in cost, but
still not cheap.....

The DDS approach is very stable (driven by use of a xtal for clock),
and very low sine distortion..

For a discrete solution one could think of using a LC tank, as basic
freq determining component, and using a JFET in series with one
of the components to effect its "value" with JFET Rds. But thats very
non linear, and circuit would need some form of AGC to keep amplitude
constant.

LC Oscillator: Circuit Working, Types, and Applications

In this article, we will focus on an LC oscillator ,which is a type of oscillator. It is also known as an LC tuned circuit or an LC resonant circuit.

www.ourpcb.com

Regards, Dana.

--- Updated Aug 11, 2022 ---

Possibly this part and some filtering....

https://www.ti.com/lit/ds/symlink/sn74lv4046a.pdf?ts=1660220223180

Regards, Dana.

 

[danadakk]

Specs to consider :

 

[d123]

Hi,

It's the invisible man again.

What's wrong with the very well-known and archaic XR2209 + simple RC integrator at output of triangle wave to turn into sine wave? It's a VCO... It goes from 0.01 Hz to 1 MHz..., 20ppm drift, needs +-4V, ~8mA to operate.

EXAR XR2209 datasheet

 

[danadakk]

Hi,

It's the invisible man again.

What's wrong with the very well-known and archaic XR2209 + simple RC integrator at output of triangle wave to turn into sine wave? It's a VCO... It goes from 0.01 Hz to 1 MHz..., 20ppm drift, needs +-4V, ~8mA to operate.

EXAR XR2209 datasheet

His freq range is to 2 Mhz....


Regards, Dana.

 

[strahd_von_zarovich]

Hi,

It's the invisible man again.

What's wrong with the very well-known and archaic XR2209 + simple RC integrator at output of triangle wave to turn into sine wave? It's a VCO... It goes from 0.01 Hz to 1 MHz..., 20ppm drift, needs +-4V, ~8mA to operate.

EXAR XR2209 datasheet

Thank you very much for you answers. Texas Instrument's "A quick Sine Wave Generator" paper worked pretty well. I used 1MHz bandwidth opamp with small cap value on the first stage to create triangle waves, and signals look pretty well on oscilloscope. Even with this configuration performance improvement is fairly high.

 

[d123]

His freq range is to 2 Mhz....


Regards, Dana.

You genuinely take pleasure in jumping on me don't you? I've been watching a pattern emerge. It must be love. Whatever. I don't mind/care. For a while now, I observe that this forum is being let down by several members who are ill-tempered in many of their replies, it will alienate new and old members alike - counter-productive and unnecessary. I thought forums, especially one about electronics, were a space of camaraderie and friendship, positivity, not playground stuff and dysfunctional behaviours such as ignoring a member on purpose and shouted posts at people. I am very calm at this exact moment, btw.

Okay, so he can use the CD4017 well-known sine wave generator circuit I mentioned in post #2, or spend more time learning to make one of the op amp versions work, or contemplate awhile and maybe come up with his own design, buy a specific part, or he could use the XR2209 and a frequency doubling circuit which can be achieved with a cheap logic IC. Using integration and/or differentiation to his advantage with the cheapest components possible (Rs and Cs) without sacrificing quality of output signal. Or the part you suggested.

--- Updated Aug 11, 2022 ---

Thank you very much for you answers. Texas Instrument's "A quick Sine Wave Generator" paper worked pretty well. I used 1MHz bandwidth opamp with small cap value on the first stage to create triangle waves, and signals look pretty well on oscilloscope. Even with this configuration performance improvement is fairly high.

I'm so glad to hear that, truly. Great, now for the hard part of polishing the circuit to perfection, good luck! Glad you mentioned that, I found the design note by chance and wondered what sort of sine wave it could muster, thanks for letting everyone know.

 

[danadakk]

You genuinely take pleasure in jumping on me don't you? I've been watching a pattern emerge. It must be love. Whatever. I don't mind/care. For a while now, I observe that this forum is being let down by several members who are ill-tempered in many of their replies, it will alienate new and old members alike - counter-productive and unnecessary. I thought forums, especially one about electronics, were a space of camaraderie and friendship, positivity, not playground stuff and dysfunctional behaviours such as ignoring a member on purpose and shouted posts at people. I am very calm at this exact moment, btw.

Okay, so he can use the CD4017 well-known sine wave generator circuit I mentioned in post #2, or spend more time learning to make one of the op amp versions work, or contemplate awhile and maybe come up with his own design, buy a specific part, or he could use the XR2209 and a frequency doubling circuit which can be achieved with a cheap logic IC. Using integration and/or differentiation to his advantage with the cheapest components possible (Rs and Cs) without sacrificing quality of output signal. Or the part you suggested.

--- Updated Aug 11, 2022 ---

I'm so glad to hear that, truly. Great, now for the hard part of polishing the circuit to perfection, good luck! Glad you mentioned that, I found the design note by chance and wondered what sort of sine wave it could muster, thanks for letting everyone know.

A CD4017 cannot be clocked fast enough to do anything resembling a sine at 2 Mhz. But with freq
multiplication thats a possibility. Partsy though.

ANY and ALL errors I make on forum I fully expect the other seniors here, or for that matter
anyone, to correct me. Clearly our goal is to deliver to the OP ACCURATE information. Otherwise
they would be smart to never return.

That is camaraderie, friendship, positivity, and respect.

NO ON THE LOVE, not even sure why you brought that up, not looking for that here.


Regards, Dana.

--- Updated Aug 11, 2022 ---

For design purposes here is a table of distortion versus sample size :

Sinewave Distortion vs Table Size

Update from original post, this update is 06/15/2021 Done using DAC fed by DMA Table of various size., No post filtering applied. Spur at 12.4 Khz still trying to ID. But basic performance should still stand. Note Third Order ignores the non harmonic components. I am still trying to ID the...

www.electro-tech-online.com

Regards, Dana.

 

[d123]

Hi,

Yeah, you're probably right about that in some way, actually, sorry.

CD4017 datasheet begs to differ: it claims 2.5MHz @ 5V, at 325ns tplh/tphl and worst case of 625ns, that's just over 3MHz or just over 1.5MHz. But no doubt if I read the small print might I ask if at that speed it would be a square wave or a -3dB shark's fin?!

IT WAS SARCASM. Me neither.

 

[danadakk]

If the CD4017 is used, to make a sine, it has to be clocked
N times to generate one cycle of the sine. N being a f() of
how many outputs are used.

So max CD4017 Sine freq = fmax CD4017 / N = 2.5 Mhz / 10 = 250 Khz using all 10
outputs.

Digital Step Sine wave Generator Circuit – Homemade Circuit Projects

www.homemade-circuits.com

Regards, Dana.

 

[d123]

If the CD4017 is used, to make a sine, it has to be clocked
N times to generate one cycle of the sine. N being a f() of
how many outputs are used.

So max CD4017 Sine freq = fmax CD4017 / N = 2.5 Mhz / 10 = 250 Khz using all 10
outputs.

Digital Step Sine wave Generator Circuit – Homemade Circuit Projects

www.homemade-circuits.com

Regards, Dana.

Hi Dana,

You make a good point, thanks.

In a last ditch, desperate, failed attempt to prove 2 MHz can be dragged out of a CD4017, I worked out best case and clocking max specs case for Vdd 15V, and the best you get is about 11MHz or 15MHz, which /10 is still not 2MHz - unless you use five outputs for positive sine and the other five for negative period of sine, a circuit I saw about PWM ages ago, that would give 2.2 to 3MHz, and be very blocky, chunky steps, maybe. But it shows it can be done, but only at limits so would be exercise in stubborn, pompous pedantry rather than your very good point. Thank goodness Strahd plans to use OAs instead of the CD4017 we keep churning over and over . What about HC and HCT, etc. - tens of nanosecond switching instead of hundreds of nanoseconds?

 

[danadakk]

Did not realize there was a Johnson counter in HC -

https://assets.nexperia.com/documents/data-sheet/74HC_HCT4017.pdf

That will clock fast enough.


Regards, Dana.

 

[BigBoss]

I believe ICL8038 will serve you well.
Datasheet says that IC can be used up to 300 kHz by changing R or C value.

https://www.mit.edu/~6.331/icl8038data.pdf

I guess it's still available.

Or something like that with a simple RC Clock circuit.