Crystal square wave oscillator discrete components

[neazoi]

hello I need a schematic of a crystal controlled square wave oscillator at about 10-14MHz using discrete components only (not IC)

All the oscillator schematics I find are sine wave...

 

[FvM]

You can e.g. build CMOS gates from a number of discrete N- and PMOS transistors. Or add a limiting amplifier to a sine wave oscillator.

 

[neazoi]

The limiter amplifier I think will produce a waveform like this **broken link removed** if I am right?
Also building the gate from ttl or mosfet is not too convenient on 12MHz...

Is there any way of doing this using some kind of diode rectification, or any other simple way?
I have found this **broken link removed** and works ok but only for 500KHz...

 

[FvM]

Your frequency requirement is the basic problem, because it involves low resistance values and high supply currents with conventional transistor circuits. But I don't see a principle limitation in this regard. You can e.g. use 5 GHz transistors at medium currents, build schmitt-trigger circuits or whatever you like. If I remember right, there has been a classical circuit, that adds a crystal to a cross-coupled two transistor multivibrator.

The discrete CMOS suggestion was meant seriously, but you probably have problems to get suitable transistors.

I must confess, that I have limited motivation to reinvent transistor circuits just for fun, that have been appropriate a half century ago, before TTL IC came up. I'm designing a lot of discrete transistor circuits, but mainly for applications, where no suitable integrated solution exists.

 

[neazoi]

Thank you, It seems impossible to find a crystal controlled multivibrator on the web I would appreciate if you could post any sources

Unfortunatelly when you are building a device that you want to be able to be built after many many years, you need to implement circuits using discrete components..

 

[dick_freebird]

You're likely going to be stuck with a sine wave core,
and gain it up until it clips and looks digital.

And it's not true that you need to use discretes. Look
how long things like CD4000 logic have been around.
And it doesn't look like they're going anywhere.

In fact some of the "unbuffered inverter" gates are
useful for this sort of thing. Or for a capacitor-
coupled, self-biased inverter gain chain as you might
want to append to whatever crystal sine oscillator
you do end up with.

But you have to ask yourself whether this is all worth
it, and whether wanting many years of production life
is really valuable. Because a CMOS digital clock-in-a-can
exactly as you describe, is pretty damn cheap and
plentiful. And standard.

 

[FvM]

There may be always reasons to create a discrete circuit, and it's up to you to decide if it's reasonable. But if you decide for it, you should also acquire the tools of the trade to design such circuits from the scratch.

 

[neazoi]

Thank you all for your replies.
As fvm said, there are reasons why someone would want to use discrete components where could, I am just not the person who likes to buy ready made things, If so I would buy a crystal oscillator module (but these are not all square wave indeed)
Anyway, I have built this circuit **broken link removed** and here are the results for 500KHz

but it would not oscillate at all at 12MHz. Also the voltage is critical in maintaining a 50% duty cycle.

The idea of a clipping amplifier seem attractive but you need more transistors to overdrive it or you need to increase the voltage too much. Then you need to attenuate the output signal to about 5V which is the TTL standard.

The idea of a crystal controlled astable multivibrator seems excellent but the only crystal controlled circuit I have found is this, which does not have any components values on it

 

[erikl]

The idea of a crystal controlled astable multivibrator seems excellent but the only crystal controlled circuit I have found is this, which does not have any components values on it

That's standard for patent descriptions ;-) . Here's another suggestion.

 

[neazoi]

Ok... and then i could use something like this
**broken link removed**
or this
**broken link removed**
or this
**broken link removed**

to transform it on discrete components? I am actually very susceptible if it will oscillate if I transform it to the other types of logic. One way is try and see, but I would like some comments..

 

[erikl]

For your 12MHz oscillator I'd suggest the TTL version. Simulate or just try it!

 

[neazoi]

Hello I made both ttl and dtl versions. As expected, it did not work...

 

[FvM]

How does your TTL crystal oscillator look like? A usual TTL crystal oscillator is build from two AC coupled gates. But the discrete "TTL" inverter from your link has a higher output resistance than 74xx parts, because it misses the output stage. It may work anyway, but possibly require a different dimensioning of the oscillator circuit.

 

[neazoi]

I tried a sinewave oscillator and then a pair of diodes ti clip the whole thing. It worked up to an extend but no clean signal..
I do not know what do I have to do to bring this **broken link removed** into higher frequency, any suggestions?