How to make a Voltage Controlled Oscillator

There are simpler, but this works out well enough, and it's a familiar circuit.

Tim

-- Deep Friar: a very philosophical monk. Website:

Hello Tim,

Either you have forgotten to paste the link, or somehow my newsreader does not show it.

Cheers, Cem

The VCO in a 4046. See figures 14B and D on page 11

The Analog Devices AD654 is simpler

and Farnell still stocks it, so I guess it is also off-the-shelf.

-- Bill Sloman, Nijmegen

Derp derp...

It's one thing to forget an attachment, but a link, that's just embarassing :)

Tim

-- Deep Friar: a very philos>

On a sunny day (26 Oct 2010 02:16:16 GMT) it happened Cem Uzunoglu wrote in :

On a sunny day (Tue, 26 Oct 2010 01:03:23 -0500) it happened "Tim Williams" wrote in :

You can make it linear by replacing the 100k resistors by PNP current sources.

Tim Williams:

It may be a byproduct of bad quoting.

Wow! A modified RC flip flop. Never thought of that. Thank you very much Tim!

Jan: Will this thing work logarithmically this way? That would be must more suited to my application than linear then? I'm going to use it to generate tones.

On Oct 26, 4:16=A0am, Cem Uzunoglu

I like this one : TL084 + 2N3819 both triangle and rectangle waveforms output :

On a sunny day (26 Oct 2010 11:16:03 GMT) it happened Cem Uzunoglu wrote in :

I knew somebody who designed a dB VU meter using RC constants, You need a minimum starting voltage here though, depending on the Vbe. For very high values of the resistors, and high values the control voltage relative to the supply voltage, it will also approach linear.

Maybe to make it log better use a linear version and add a log amplifier in front. or a log potentiometer perhaps?

Looks cool. There are no resistor values on the schematics though. Are they important?

Damn! I knew that, and thought that I'd dug out a 4000 series data sheet! I must have failed to copy the URL and pasted an earlier reference.

Sorry.

-- Bill Sloman, Nijmegen

Resistors will give you RC time constants, which have an exponential sort of thing. For voltages close to Vbe, time goes to infinity (or the gain is too low to turn on the transistors anymore and it stops completely). For large voltages, it looks linear (in which case, the resistors approximate current sources).

If you require linearity, the good old hysteretic comparator with current sources is hard to beat.

Tim

A 74HC05, capacitor and three resistors:

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

Pffbt. Too many resistors, and way too many pins:

If you'll accept a control current, instead of a control voltage, it can be done in two components :)

Worked fine in the toob days. Back then, they had a hundred volts to play with, so the required headroom wasn't a problem. It is quite drifty and noisy, and should be kept away from external light or radiation.

Tim

On a sunny day (Tue, 26 Oct 2010 10:45:49 -0500) it happened "Tim Williams" wrote in :

Same idea works with an UJT too:

Uc +12 V | | [ ] R1 |-- |---------- | UJT | \|-- === |------- pulse out | C1 [ ] R2 | | /// ///

I used a PNP current source for R1 to make it controllable.

Fun, I'll have to bread board it.

George H.

Been there, done that. That NE schematic sucks unless you use a thyratron designed to work that way; or add a thyristor so the NE lamp latches it into the open state.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant