CMOS Timer Project?

[Newtronics]

Hi. I am new to electronics as you will guess. What I want to do (need) is to somehow put together a timer which will close a circuit momentarilly at random, time wise. so that the project will operate at different times of the day, in no given order. The project when finished would have to be small, powered by 4.5 to 9 volts. This is where I thought CMOS would be the best. Can anyone give me some pointers for such a random timer/switch device? Any help appreciated. Thanks.

 

[Roff]

More info is needed. What does "momentarily" mean in your project? How much current does the circuit draw that you want to provide power to? Does it run off the same power supply as the "timer"? How many times a day (or alternately, how frequently), on average, do you want this "random" event to occur?

 

[Newtronics]

CMOS Timer Project

Hi Ron. Many thanks for coming back on this. Don't laugh please, but my nearest and dearest has a mechanical Song Bird which sings beautifully when you get up and press the button. Works off three AA dry cells. She would like it to sing for her often (randomly) without having to get up and keep pressing the button. There is room inside the base for a small circuit. So I thought, if I could build something that would connect the wires from the push button at varying intervals, this would do the job. It would have to be a small project. Maybe taking power from the 4.5 volts of the batteries, which would be better. Maybe a solid state relay to close the contacts. Momentarilly because the button only needs to be pressed briefly to operate the Song Bird. Silly? Of course. But it could bring so much happiness to someone I love. I could try and build a Synthesiser or even even a Super Robot, but they would not please my wife like this would. Thanks Ron. Hope this helps. Walt.

 

[Roff]

Newtronics, that was useful and interesting info, but you didn't answer most of my questions. The answers are important to coming up with a solution. Reread my questions carefully and try to answer them.

 

[john1]

Hi Newtronics, Hi Ron,

How about one of those key rings instead.
One of those key rings that you whistle for if you lose it.

They take only a tiny amount of current on standby,
and i feel sure you could get it to drive a reed-switch
through a transistor, instead of the answering beep.

It also might be amusing because the bird would appear
to answer the whistle.

I take it that the button only has to have a brief press,
then the bird carries on for a bit.

John

 

[mogur]

Random output from logic gates is sorta against the rules. You can fake it with pseudo-random (eventually repetitious) output by recursive jamming of shift-registers. [See below.] True randomness is really hard even in micro circuits where commonly they have to seed the start of psuedo-random generator subroutines with realtime clock values (which are fairly independent of the program execution timing). I haven't tried it, but my first thought for an almost true random device would be to store repetitive binary samplings of contact noise, to be clocked out of memory on demand. This would be over-kill for this project, however, and not very battery efficient. A whistle sounds good to me. :? :wink:

Here is the circuit I tried to suggest earlier (a dual 555 probably needed to stretch the timing out to hours)-

The circuit makes use of inexpensive TTL (Transistor-Transistor-Logic) chips to produce a pseudo random sequence of on-off states for the LED. The circuit is actually called a 127-state pseudo-random sequence generator, typically used to generate "white noise" or random numbers. A 555 Timer Chip is used as a self-oscillating clock circuit which "cycles" a 74164 8-bit shift register chip. The 6th and 7th register outputs are combined using a 7486 Exclusive-OR logic chip and the results fed back into the 1st register. This results in a continuing cycling of the shift register output so that 127 different states are repeated. In other words, the shift register output states are not repeated until 127 "clicks" of the clock. Sampling the output of any register will yield an apparently random length of on-off cycles. This is used to drive a LED, lighting the LED when the output is a 0 and turning the LED off when the output is 1. Since the 555 timer circuit is set to run at about 90 hertz (cycles/sec), the output sequence would repeat at about every 1.5 seconds (frequency of 1/127 of 90). This makes the LED on/off cycling appear random.

 

[john1]

Thank you mogur!

you obviously recognise those little whistle key rings ...

 

[Newtronics]

CMOS Timer Project

Great to have so many coming in with replies. The humour is great too. This Singing Bird though. The original circuit need not come into it really. The 3 1.5 volt dry cells are hardly touched as the circuit draws only minimal current, hardly any more than the inbuilt digital clock that runs alongside it.

All that is needed is (it has to be very small) an independant circuit, taking it's power from the dry cells, able to operate a sub miniature relay or even a reed switch, to close the contacts (the press button contacts) that cause the Song Bird to sing. The contacts need to be closed very briefly then opened again. If we could get this Song Bird to operate on around 4 or 5 times each day, that would be great.

I personally do not have that knowledge. But I do know that there must be many of you gentlemen that could think of a semiconductor device that would operate the relay randomely in this fashion. Can I put it to you again? If I have answered imperfectly, then I apologise for my failing. Thanks again. Walt.

 

[bogdanfirst]

how about a PIR or a presence sensor, so that it sings when you go close to it?

 

[john1]

Hi Newtronics,

you haven't been offered a random device yet because such a
device is actually very difficult to make.

I would be interested to see if someone comes up with a
suitable device that would operate four or five times per day
on a random basis.

john

 

[mogur]

Oops, I meant to quote my earlier post, but must have hit the edit button instead. Oh well, check above for my 2¢.

[earlier post used to be]-

If she truely insists on random chirping, then clock an eight bit shift register (74164) with the 555, Exclusive-OR (7486) the last two registers, and then feed that output back into the first shift register. The output of any of the registers will then be a 127-state pseudo-random, repeating pattern.