if i had an astable timer which can time around 1hr, how can i make the output LED stay on for 10 seconds rather than the full hour when the capacitor discharges? what would the circuit look like?

 

An astable for an hour would be pretty unlikely to work well (although what you want is actually an un-symmetrical astable). A more practical solution would be a micro-controller, it looks tailor made for an eight pin PIC.

 

If you use a counter-timer like the CD4521BE, an accurate 1 hour time can be obtained. Follow that with a 555 timer for 10 second time.

__________________
see my website: www.geocities.com/russlk

 

Hi qyc88,

you can of course build a simple capacitor/resistor
charge/discharge circuit using only a few discrete
components, but the accuracy is poor.

If accuracy is not an issue, merely that the unit
operates at "around 1 hr" as you put it then an
fet based circuit similar to the one shown below
should suffice.

I made this a few weeks ago, its drawn mostly from
memory, i made very few notes about it, this version
gave me about 30 mins as i recall.

to increase that to an hour, try doubling the cap,
the 1.5 k is the component that will mainly affect
the lit time of the led.

I used a BF244 fet transistor, but any normal fet
should be ok, but check the gate control direction
with this one, lowering gate volts below source will
progressively close the device, that is normal.

With some types, raising gate volts above the source
will progressively open the device, not so normal
but thats ok, just a different type, and wont do for
this circuit. (there are other variations too)

The SCR can be almost any small type.

The charge curve is a straight line, which means its
better than you would think at consistency, but no
where near as good as oscillators with counters.

Best of luck with it, John

Attached Images

Timer.jpg (11.4 KB, 288 views)

 

John, there is no path to charge the capacitor in that circuit.

Edit:
A thousand pardons. I just discovered that BF244 is a JFET. My bad.

 

Thats quite ok Ron,
thanks for having a good look.
I fiddled around for ages to get that to work,
i was working along the lines that i had seen a
similar circuit years ago using a triode.

As you probably realise, this type of fet has
characteristics which make it virtually a triode.

The circuit is so blindingly simple that it is a
surprise when it actually works!

the charging voltage for the cap is only marginally
above the rising voltage on the cap, which gives the
long period of charge.
I did find that leaky caps wouldn't do very well, i
don't know why not, i assume the leakiness might vary
a little with voltage, causing the charge to 'stall'.

It was adapted from a circuit used in an old
oscilloscope that i worked on many years ago. I don't
remember its name, i vaguely recollect it was a split
beam with a blue trace, and a fitting for photosensitive
paper across the front.

I was surprised at the time to find the very long
trace times from such a simple circuit, especially that
the components used then were not exceptionally large
capacitors, thats why it stuck in my mind.

Cheers, John

 

It's called a bootstrap ramp generator, because the resistor that charges the capacitor is bootstrapped such that (approximately) constant voltage appears across the resistor, making it a constant current source. Below is a block diagram, and one other possible implementation.

Attached Images

bootstrap.GIF (8.1 KB, 244 views)

 

hi, i will recommend you to use a PIC, so your accuracy will increse a lot. You can use a retroalimentation into one pin to detect the transition from low to high level and viceversa, and part of your program must include that when you detect the transition, another pin goes high for 10 seconds, always most accuracly than a 555.