I need a simple (preferably one CMOS chip) timer/counter which puts out a logic high for about 1 hour out of every 24 hours for a battery charger circuit. Does not need crystal accuracy; a R-C oscillator is good enough. Does not need to be settable or resettable; it will run continuously and I don't care what time of day/night the output will be active.
I have an old Motorala CMOS manual which shows a MC14451 which would work for this, but this part appears to be obsolete, and is only available as NOS. I could cobble something with CMOS hex inverter used as an oscillator, a string of counters, and some and gates to create the output period, but I'm hoping that someone has a single chip solution for me?
There are 86400 seconds in a day so what if you used a 555 timer that had a .86 second cycle that fed into 5 divide by ten IC's in series. for a total of a divide by 100000 count. Then use the 1 in 100000 pulse output to trigger another one shot 555 timer that does a roughly 1 hour time out.
All analog and crude but still a interesting design challenge.
Starting from 32kHz, it would take a 22 bit counter to get to 24 hours. The Capture counter in a PIC is 16 bits, so I could use software to count overflows from the 16 bit counter which is counting the 32kHz crystal. Which small PIC has the counter hardware?
The CD4060 is a 14-bit binary counter with a built in oscillator circuit that can be connected as an RC (or crystal) type oscillator. A 5.27Hz oscillator design would give you a 24-hour cycle. You would need another chip enabled by the 24-hour cycle to give you the 1 hour period.
Thanks Carl. That one looks like it might work. If I AND Q14, Q13, Q12, Q10 (unfortunately they didn't make Q11 available) I would get two 45min periods in a 24hour cycle, 45 min apart? Q14 would be high for 12 hours, Q13 for 6h; ... Q10 for 0.75h (45 min).
Thanks Carl. That one looks like it might work. If I AND Q14, Q13, Q12, Q10 (unfortunately they didn't make Q11 available) I would get two 45min periods in a 24hour cycle, 45 min apart? Q14 would be high for 12 hours, Q13 for 6h; ... Q10 for 0.75h (45 min).
This is the sort of thing I see here from time to time, and think I might want to come back to it. Then, the thread eventually rolls off the page, and I can't find it when I want it. Is there a bank where these can be stored?
This is the sort of thing I see here from time to time, and think I might want to come back to it. Then, the thread eventually rolls off the page, and I can't find it when I want it. Is there a bank where these can be stored?
This is a single-IC solution, uses a PIC12F508. Uses mains frequency as timebase. Has an output that goes on and off the same time(s) every 24 hrs. The original application was for an applicance that learned a daily routine, in real time. You program the required on and off times by being there at the right time and pressing the button.
It's my code I'd wrote, I pass it into the public domain with no warranty or support.
It's an "electronic version" of this common **broken link removed**mechanical timer, complete with the Pegs (you know, the motor thingy with the pegs round the wheel that goes round each day)
A simple way to get a programmable 24 hour timer is to take an old digital alarm clock radio and tie a relay into the circuit that powered the radio part. Use it to turn on the device in question.
I did that to one in high school for an experiment in order to get my boom box to work off of a timer. I also set up one that way for a friend of mine years ago so he could use it as a engine heater timer in the winter.
Most alarm clocks automatically shut off the alarm after an hour or so.
I don't have a simulator program, but on occasion have had to due long time periods. I set the divider to where I think it should be and then use a faster clock than needed, and then measure the output time on my frequency counter that also measures period. If the divide is correct then I use the intended clock. I have used the 4541,4060, and 4040 many times.