Slow but precise blinking LED

[JBOUN3]

First day here. First breadboard project:

LED blinking at precise hertz, 4, 10, and 40. Note, very slow not MHz nor KHz, very slow.

The 555 IC version worked rather well but not reliable enough by way of clock-steady. Currently using 4060 IC using RC better, but also not reliable enough. When I added a crystal it turned out that the only effect of swapping the 4.000 with 4.700 - same with 7.500 and 8.000 crystal seems to have nothing to do with the resonance of the crystal. In fact, the LED continued to blink close to the same rate without the crystal! No doubt the crystal was not singing at all, but simply changing the RC resonance like a rather dumb resistor.

I'm looking for precision. 40.000 Hz if possible (and why not?) or even more precise.

Full disclosure: project involves optigenetics. (I'm no neurologist, - retired computer programmer looking to enhance my (and everybody else's) brainwave synchrony, particularly along the gamma band, 40 Hz. - for more info, please look up Dr. Tsai at the Picower institute at MIT and check out the effect of 40 Hz on Altzheimer's- model mice. Great news, in my supposedly humble opinion!)

So would you please enlighten me as to what need I do to make my crystal comply (other than study EE for a few years, which looks like an option in any case). Tightening up proximity on the breadboard resulted in a 17% change in frequency, thus proving that (any) b-board stage is but a wind-up to "real" prototyping on a real circuit board.

Thank you in advance.

JB

 

[KlausST]

Hi,

an xtal is the way to go.
You may use the 4060 IC or you may use any microcontroller.
The benefit of the microcontroller is, that you don't need a soldering iron to change blink frequency.

Now to your problem:
Obviously the Xtal doesn't oscillate at the correct frequency.
I assume this has to do with your wiring. A breadboard is not suitable for this high frequencies.
Simply connecting devices according the schematic does not work in this case, because your wiring introduces stray capacitance and stray inductance. Such a breadboard circuit may work or not.

Without a solid, low impedance GND plane it's difficult to debug.
Best is to decide that the connections close to the 4060 GND pin is your GND star point.
* Connect the 4060_VCC_capacitor (ceramics) directely - without additional wires - to this GND point.
(For sure: short and no additional wire to the VCC pin also)
* Connect the oscillator_out_capacitor (ceramics, COG, NP0) the same way to GND
* Connect the oscillator_in_capacitor (ceramics, COG, NP0) the same way to GND
* connect the xtal as close as possible to the oscillator pins of the 4066.

If it still doesn't work, then show us a picture of your breadboard.

Klaus

 

[FvM]

I've used 4060 with watch crystal at times when no other low part count timing solutions have been available, these days a tiny (e.g. 6 or 8 pin) microcontroller would be my first choice. Depending on the accuracy requirements either internal RC oscillator (1 -2 % initial accuracy) or an external crystal can be used.

 

[schmitt trigger]

If you are capable of programming, a microcontroller with a crystal oscillator is again the best and cheapest choice.

Otherwise a 32 Khz crystal oscillator made with an unbuffered inverter, followed by a CD4059 programmable divide-by-N counter, and a CD4013 flip flop to obtain the 50% duty cycle.