2
\$\begingroup\$

I want to make a timer that counts down from 15 minutes, in 5 minute intervals.

It should use 3 LEDs, and behavious should be such that when it is switched on, all 3 LEDs are lit. Then after 5 mins one of the LEDs goes out. Then after another 5 mins another LED goes out, and finally after 5 more minutes the final LED goes out.

This is my first stab at circuit design, so I wondered if someone could give this the once over. I suspect I've missed some key components, so any corrections or advice would be much appreciated.

Here's the circuit, based on a 4060B to do the timing, and a 4017 decade chip to keep count. There's also a 4049 inverter to ensure the LEDs go off as the pins go high, and a 74AC11032 chip to facilitate the "or" logic.

enter image description here

\$\endgroup\$
  • \$\begingroup\$ You can be brutal - I won't be offended... \$\endgroup\$ – Ben Nov 17 '11 at 20:19
  • \$\begingroup\$ Are my calculations out, or does that R/C combination give a 36 or so minute clock period? Or is C1 not in µF but in something else? \$\endgroup\$ – Majenko Nov 17 '11 at 21:29
3
\$\begingroup\$

I would suggest looking at a 74HC164 shift register as an alternative to the 4017. Every time the clock goes high, the "and" of the A and B inputs will be copied to the first data output at the same time as each of the other data outputs grabs the state of the one before. The chip also has a "clear" wire which, when driven low, forces all the outputs low regardless of the state of the clock.

If you use such a chip with A and B strapped high, then on power-up or when /CLEAR is hit, all outputs will go low. Then when the first clock pulse arrives, the first output will go high. The second clock pulse will cause the second output to go high, etc. up to the eighth. If you need more than eight outputs, you can add any number of additional chips; tie the A and B inputs of each additional chips to the last output of the previous one.

For example, using this approach you could if desired use two such chips and hit them with a clock once per minute, thus providing a 15-minute readout in one-minute increments. Or, if you like, you could use four chips and get the readout in 30-second increments.

\$\endgroup\$
1
\$\begingroup\$

It looks a reasonable circuit to me.

I'm no expert though ;)

I can't see anything to be brutal about.

I take it that when the count reaches 0 it stays there until the next time you power it up? It would be a simple enough task to add a reset button so that you can reset the timer with a simple button press.

If you really want me to be brutal I could say "You have gone about it all wrong! Discrete logic is so last year! You should be using a microcontroller to do it all!"

Yes, a microcontroller would simplify the circuit somewhat, but would that be overkill for just a 15 minute counter? Well, you could get it to do so much more with one - like flash all the lights when the time us up, etc.

Here's a little example circuit using a microcontroller. The one I chose is one I just happened to have a symbol for in my schematic program, and there are much smaller ones that would be more suited to the job, but you get the idea:

enter image description here

I can't recall if that microcontroller has the built-in oscillator (I think it does), so no external crystal or RC network is needed.

Of course, you need some way to program the microcontroller, and they do take some of the fun out of low-level circuit design...

\$\endgroup\$
  • \$\begingroup\$ A micro isn't really over kill means you can add to the project such as 3 minute intervals and varying time length. \$\endgroup\$ – Dean Nov 18 '11 at 0:10
  • \$\begingroup\$ I'm kind of veering towards microcontrollers instead. Is there a good open source simulator I could prove the concept with? Ideally I'd like to write it in C as i have experience of that already. \$\endgroup\$ – Ben Nov 18 '11 at 10:16
  • \$\begingroup\$ Most of the microcontroller manufacturers provide their tools free of charge (some charge for "professional" versions) and they usually include a simulator. As for open-source, some of the manufacturer ones are, or there are some gnu tools that are OK. I usually operate direct on a chip myself as they're cheap enough and can be re-used thousands of times. If you're just getting into it then you may want to look at the Arduino or the Launchpad as a good introduction - the Launchpad is only around $5. \$\endgroup\$ – Majenko Nov 18 '11 at 10:26
1
\$\begingroup\$

I can't see any glaring errors popping out at me.
The overall concept seems correct, although I'm not sure about the timing if you want 5 minute intervals (see below)
Also, the gates look like they could be AND gates, whereas they were OR gates in your initial circuit which was correct. You mention OR just above the schematic though, and the code for them brings up an OR gate, so I'm figuring that's what you are planning on using whatever the symbol looks like to me.

As Majenko mentioned, the timing cap has no indication of whether it's pico/nano/micro farads.

Taking 2.3 for the multiplier (the other datasheet I saw gave 2.2) and assuming the 47 is either nanofarads or microfarads I get:

\$ 256 \cdot 2.3 \cdot 220k\Omega \cdot 47nF = 6.08 \ seconds \$

If it's microfarads:

\$ 256 \cdot 2.3 \cdot 220k\Omega \cdot 47\mu F = 6088 \ seconds \$

Either way is a long way off 300 seconds. For a \$47\mu F \$ cap I get a ~10.8k resistor is needed.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.