Simplest Transistor Oscillator and Breadboard Crystal Oscillator Actually Working!

Print Friendly, PDF & Email

To have a more complete understanding of today’s networked tech world, you have to link some key electronic fundamentals to the OSI layer 1 area of IT networking that deals with the actual voltages that generate binary values 1 and 0 that have to be regulated/counted with reference to time of an accurate reference clock of some sort e.g. an oscillator circuit, to then be counted, grouped then streamed or stored in various ways according to some set rules or protocols like I2C, ADSL, Ethernet etc. See my Post on encapsulation for more info:

https://stevepedwards.com/DebianAdmin/encapsulation-from-bits-to-gigabytes-in-200-years-2/

These clocks can come in many forms from different natural or man made sources – pendulum, tuning fork, atomic decay, light wave frequency, piezo crystals or other electronic components that oscillate periodically, and hopefully consistently over “time” itself!

The more consistent, the more accurate and useful they are – see the history of the Longtitude Clock prize for navigation for perspective!:

https://en.wikipedia.org/wiki/John_Harrison

The simplest powered oscillator circuit I’ve found is below:

A “reversed” NPN transistor (I used a 2n2222) with floating base and reversed voltages breaks down at about 12V or more, causing oscillation. Try different V, R and C values to alter the Time Constant, T=RC. Be mindful of minimum resistor values to not fry the tranny or LED (<20mA)

For accurate electronic time for very cheap, the Quartz Crystal oscillator is used in almost every circuit on most devices – ever!

They can be really difficult to get going on a breadboard due to high frequency noise and bad connections causing capacitance issues – and they won’t necessarily run at their stated frequency, but at a harmonic vibration above their fundamental tone. It took 5 different attempts over 3 days with different circuits to get this 16 MHz crystal stable and working – but at 23.3xxx MHz not 16!

The next step would be to feed this into circuitry that would generate, say 5V logic square waves instead of the sine wave then divide the rate down to suit your desired bus speed…get the general idea?