Low frequency analog VCO with sinewave output?

On a sunny day (Wed, 24 Aug 2011 02:22:29 -0700 (PDT)) it happened davew wrote in :

There was this 22?? sort of chip that does triangle and sine. Used, it, if I only could remember the number... :-)

EXAR XR2209

If you are restricted to one chip, this is probably as good as it gets. It distorts a triangular wave to a tolerable approximation to a sine wave.

You can do better by feeding a square wave through a shift register clock at some convenient multiple of the square wave frequency, tying a series of resistors to the outputs of the shift register and summing the currents through the resistors into a virtual earth to produce a staircase approximation to a sine wave, but it takes quite a long shift register and quite a few close tolerance resistors to do better than the 8038, and it's still a three chip solution - basically a programmable logic device to provide the shift register and the dividers to generate the square wave, the VCO to to generate the clock at some fairly high multiple of the output frequency, and an op amp to sum the currents from the resistors, plus you have to find board space for the resistors.

The Analog Devices DDS chips do provide a very good one chip solution

- the more expensive chips include 14-bit DACs - but it isn't an analog solution.

-- Bill Sloman, Nijmegen

--- Clue:

"Best I've found so far might be the old 8038 function generator chip."

:-)

-- JF

"davew"

Sorry, not a single chip solution. But, if the system where this will be used already has a microcontroller, why not build your own? Because of the huge consumer electronics market there are many good and inexpensive audio codecs and DACs. (AK4425A for example) Feed it with the right data and wou can generate any waveform. A two-chip solution is probably doable - a micro with a built in ADC and an audio DAC.

-- Roberto Waltman

Hi Roberto, thanks for the idea. As it happens I want to stick to analogue in this application. Bit of a dinosaur project in mind. ;-)

XR2206

Duh, check the data sheet......

The 2207 doesn't but the 2206 does ;-)

You can find a few app notes with triangle-to-sine converters (using diodes or JFETs, typically). If you don't need a single-chip solution, that's another way, but the parts count will be relatively high.

Getting relatively low distortion (

Yup, makes a nice phase-coherent FSK modem transmitter. I'm using them now, spares for a 25-year old distributed automation system. Story, eventually.

John

-- Many thanks,

Don Lancaster voice phone: (928)428-4073 Synergetics 3860 West First Street Box 809 Thatcher, AZ 85552 rss:

Please visit my GURU's LAIR web site at

How about a CD4040 and a bunch of resistors? (I think I mean the CD4040

-- the 4020, 4040, and 4060 are all big-ass counters, one of which has an oscillator built in).

It's a ripple counter, but if you're going slow enough the clock glitches shouldn't get you too bad.

Thanks but I'm sticking with the analogue method in this case. Frequency stability isn't so much of a problem as I'm controlling frequency as part of a discrete PLL system.

A PIC with a PWM output, a resistor, a cap, and a blindfold.

The blindfold is to make it "analog".

But why do you need it to be analog? In particular, if the 8038 is good enough, why can't a _really cruddy_ digital solution work just fine? Those old "triangle into sine" generators had really nasty glitches at the peaks of the sine waves, and I'm not even going to guess at the overall harmonic distortion (or temperature dependence thereof).

Which leads to the questions:

• What frequency range does it need to run in?

• How good does it have to be? * How quiet does it need to be? * How linear does the command voltage vs. frequency need to be? * How close to a sine does it need to be?

• Temperature range?

• what other questions did I leave out?