low frequency tuned oscillator

[michiel]

Recently i started a project: trying to build my own function generator.
The most important reason for me doing this project is learning more about oscillators and their limits and possibilities.
So far i found a few designs such as the colpitts oscillator and the wien bridge oscillator.
The problem i have with both is their low tuning range due to the low values of availible variable capacitors(100pF max on farnell).
I'm aiming for a sine with a frequency from 10Hz to 500kHz though i could realise this with multiple oscillators.
I would appreciate your input in this.

 

[MikeMl]

Look up the historical Hewlett-Packard model 200. It used a Wein bridge with a dual-section variable capacitor out of an AM desk-top radio. AFAICR, that oscillator had several ranges, the lowest of which went down to 20Hz.

Maybe you can find one and "transistorize" it...

Of particular interest is how Drs H & P stabilized the oscillator so that it has low harmonic distortion.

 

[JimB]

I'm aiming for a sine with a frequency from 10Hz to 500kHz though i could realise this with multiple oscillators.

Using "simple" oscillators, it is usual to limit the frequency ratio to no more than 3:1 for LC oscillators (often only 2:1 in more exacting applications), for Wein bridge types using a variable R or a variable C to tuning ratio is usually no more than 10:1.

You are trying for a 50,000:1 ratio. Using simple oscillators you are not going to achieve that in one range.

It could be done by mixing two oscillators, one fixed and one variable.
You could have a fixed 10Mhz oscillator, and a variable 10 to 10.5 Mhz oscillator (nice easy 1.05:1 tuning ratio), feed these two oscillators into mixer and lowpass filter the output to get 0 to 500khz.
But this is getting away from "simple".

Another way of obtaining a wide tuning range is to use a Direct Digital Synthesiser (DDS).
Modules for these types are cheaply available:
https://www.ebay.com/bhp/ad9850-module
But then this is not simple, but all digital and complex.

JimB

 

[michiel]

It could be done by mixing two oscillators, one fixed and one variable.
You could have a fixed 10Mhz oscillator, and a variable 10 to 10.5 Mhz oscillator (nice easy 1.05:1 tuning ratio), feed these two oscillators into mixer and lowpass filter the output to get 0 to 500khz.
But this is getting away from "simple".
JimB

thanks for the response.
this seems feasible to me. i have an AD633(multiplier) ic laying around from a FM project.
i would use a crystal for a fixed frequency and a varicap diode for the variable lc/rc oscillator.
the AD633 has a bandwidth of 5MHz which might be due to the op amp buffer.
I'll test the multiplier with 10MHz after the weekend and post the results for anyone interested.

 

[unclejed613]

it doesn't "HAVE" to be 10Mhz. you said the multiplier has a bandwidth of 5Mhz, you could so it with a 5Mhz fixed source and a 4.5-5Mhz variable oscillator. you will get the same results, with the only difference being the direction of the tuning. using a mixer to get difference frequencies works whether your variable oscillator is above or below the fixed source.