Circuit ideas

[Zig]

Hello group, I have a requirement for an oscillator.
So far I am using the humble CD4060 which clocks at somewhat higher rate and the first output in the divider chain provides me with the required 50% duty cycle signal.

I have a couple of spare CD4066 analog switches in an adjacent package.
I know its possible to get the two analog switches to oscilate and have infact got them to do so inthe old cross coupled flip flop configuration ( a dual transistor flip flop).

I was wondering if there were other configurations out there in the wild( have looked and searched the web ) but no joy.

Circuit suggestions invited.

[mcphill]

Originally Posted by Zig Hello group, I have a requirement for an oscillator. So far I am using the humble CD4060 which clocks at somewhat higher rate and the first output in the divider chain provides me with the required 50% duty cycle signal. I have a couple of spare CD4066 analog switches in an adjacent package. I know its possible to get the two analog switches to oscilate and have infact got them to do so inthe old cross coupled flip flop configuration ( a dual transistor flip flop). I was wondering if there were other configurations out there in the wild( have looked and searched the web ) but no joy. Circuit suggestions invited.

How 'bout a $2 PIC chip and just program whatever waveform you want?

[Al_The_Man]

There is the CD4541 it has an internal osc and can be used with a Xtal also.
What frequency are you aiming for ?
Al.

[amplexus]

Check out the cmos cookbook; how accurate? square wave? Single phase? frequency? duty cycle? A pic is probably not the best choice, try a cpld or lower level parts
Amplexus Ender

[Zig]

Ok the situation is as follows..
The circuit uses CD4060 already as a signal source.
It drives a bridge which is followed by a synchronous detector based on CD4066switches.
I have a couple of CD4066 switches as spares in a package.
By getting the two spare analog switches to oscilate ( 50% duty cycle) i get rid of CD4060 chip.
and by rearanging the gain stages between the bridge and detection stage i can also get rid of an inverter package used to generate biphase clock for the detector.

All around a saving of two chips and some real estate.

[steve323]

If you just need an oscillator, why not just buy a metal can oscillator? There are 7663 to pick from at Digikey. Most should have +/-100ppm or better accuracy.

If you just need something to toggle, then you can't get much simpler than a 555 timer.

Steve

[Zig]

Thank You all for the various recomendations.

Frequency in question needs to be something an order of magnitude higher than significant mains frequency harmonics and it ought not be harmonically related to amins frequency ( in my case 50 Hz).

Stability is not terribly important nor is accuracy of setting the frequency as the signal will be driving a resistive bridge. One arm of the bridge will be a PT 1000 RTD on about 2 6 feet of cable operating in an electrically noisy environment.
The output of the bridge will be synchronously detected ( which eliminates need fro precise frequency control) which will eliminate a whole bunch of noise ( process gain).

The crosscoupled gate topology provides a low frequency oscilator with biphase output.The topology however is very sensitive to chip variations and cause the devices to latch and fail. I have looked at another topology where thw two swithes are connected in series and the timing components get accessed sequentially .

So far stable performance a higher range of requencies is availble but the output is not 50% duty cycle.
Anybody witha coy of cmos cookbook ..please if you have the time peruse it with the view to seeing if there are any cd4066 oscillator ideas in it.

[Mariss Freimanis]

The answers to Zig's problem are interesting. The microcomputer guys suggest a PIC or some other processor and the Gordian Knot cutter types suggest buying an oscillator instead of building one from scratch.

This leaves out the optimal solution that Zig is after; the elegance of doing what's need with the minimal number of parts. That is a goal worth pursuing and perfecting if you mean to become a circuit designer.

The cross analog switch type oscillator will do in a pinch. The better option is if your circuit requires quad NAND gates or hex inverters, pick the Schmidt-trigger kind (CD4093 instead of CD4011, CD40106 instead of CD4069). Any CD4093 NAND gate and any CD40106 inverter can be turned into an oscillator by connecting R from output to input, C from input to GND. The oscillator frequency will be 1/RC if input hysteresis is 1/3 Vdd. It works nicely from 1MHz down to 0.1Hz or so.

Mariss

[Zig]

Maris,

The circuit already uses a hex schmidt to provide biphase clock.
By setting up cross coupled switch oscilator i obrain the two phases directly... however the oscilator is as i said marginal.

works over a limited frequency and tends to fail should the compnent choice be les than optimal ( lost three chips to experimentation).

PS.. How to attach a graphic fle

[amplexus]

can you use a simple rc relaxation oscillator to trigger your swithhes?
Amplexus Ender

[Zig]

To help the discussion along, here are a couple of ideas.

The single phase output ( monophase ) circuit is fairly well behaved. In order to generate a 50% duty cycle waveform the capacitors need to be juggled with .

The socond circuit ( biphase ) produces inphase and out of phase ( 180 degrees) outputs as is required however the circuit tends to be unstable and switches can latch up and destroy the chip.

[Zig]

The original synchronous detector idea contains an oscilator divider chip to ensure 50% duty cycle and a hex inverter to provide biphase clock signal.

Idea beiong to replace the two chip solution with the spare analog switches configured as a 50% duty cycle oscillator.