Low frequency oscillator

[epaelectronic]

Hello.

What would be the best way to get a stable 7-8Khz tunable oscillator?

Preferably a voltage controlled one.

What I have in mind currently is using a 4060 to divide down a 4Mhz crystal that is being slightly tuned with a varactor diode.

Would this kind of arrangement get me to the 1Khz frequency span I need?

[Zero999]

The 4046 is the easiest way, I know of.

https://hackaday.com/2015/08/07/logic-noise-4046-voltage-controlled-oscillator-part-one/
http://www.learningaboutelectronics.com/Articles/4046-voltage-controlled-oscillator-VCO-circuit.php

[BrianHG]

The 4046 is the easiest way, I know of.

https://hackaday.com/2015/08/07/logic-noise-4046-voltage-controlled-oscillator-part-one/
http://www.learningaboutelectronics.com/Articles/4046-voltage-controlled-oscillator-VCO-circuit.php

Yup, that's the cheapest and easiest.  Also, if you only need 5v or 3.3v, you can use the 74HC4046 if it is easier to get.  Just one word of warning, configure the chip with some headroom.  IE, if you need 7-8Khz, tune it for a range from 6Khz to 9Khz.  The IC can drift a little over time and temperature.  Also, the cap which you use to set the oscillator frequency should be an NPO variant so that the capacitor value will be immune to ambient temperature.

If you need precision, it can be complicatedly done with a cheap 1$ MCU.  Such a solution properly done would guarantee that with the voltage input at 0v, you will get dead on 7KHz, and with the voltage in at VCC, you will get exactly 8KHz.  Practically 0 drift as a reference crystal source's drift would be measured in the parts-per-million.

[Benta]

Sorry, but I'm trying to wrap my mind around the suggestions here. I've worked a lot with the 4046 (call me "4046-lover"), but for this application it's the last chip in the world I'd select.

The 4046 VCO is not very precise nor stable. It relies on the feedbacl loop in a PLL.

BrianHG introduced the idea of a $1 MCU.
A much better idea. Find one with ADC, use the timer to divide down to what you need controlled by the ADC voltage.
No continuous frequency sweep, but should do what you need.

Otherwise you'll need to provide MUCH more information on your needed spec.

[Zero999]

Well, I did say the easiest, not the best.

Experience with this kind of request tells me, they normally want to avoid an MCU.

How about the LTC6990?
https://www.mouser.co.uk/datasheet/2/609/LTC6990-1776026.pdf

[Kleinstein]

The trim range for a cystal oscillator is too small, more like 0.1% of the frequency.  It would be more an LC oscillator with varactor diode at the 4060.

[Zero999]

The trim range for a cystal oscillator is too small, more like 0.1% of the frequency.

One solution could be to use two HF oscillators, a fixed one, modulated by a VCO and the beat frequency demodulated to give the low frequency output. It's not the simplest solution though.

[Benta]

The right device would be the Philips NE566, but it's obsolete for decades (if it's a one-off, you might find some NOS somewhere).
Here's a link to the datasheet, it might give you an idea on how to roll your own with a couple of opamps.

https://html.alldatasheet.com/html-pdf/17982/PHILIPS/NE566/244/1/NE566.html

[Zero999]

How about the LM339?
The circuit is on page 10, figure 17 of the ST data sheet.
https://www.st.com/resource/en/datasheet/lm139.pdf

[Benta]

Basically a copy of the NE566. Could work.

[emece67]

.

[David Hess]

Would this kind of arrangement get me to the 1Khz frequency span I need?

Dividing down a crystal oscillator source will not allow much tuning range.  If you need crystal stability, then two crystal oscillator outputs could be mixed to produce a stable variable low frequency output.

The more modern option is a crystal oscillator driving a DDS generator which can take many forms.

[Zero999]

Basically a copy of the NE566. Could work.

Yes, I suppose it is similar to the circuitry in the NE566. It's a shame the IC was discontinued, but obvious why.