I need something to give me a pulse.

I am finalizing a circuit to power my Arduino board, which I'll eventually share once I get everything working. My only problem is that I am missing one component that provides a short duration logic "high" pulse. This pulse will be used to turn on the power to my Arduino (so the pulse cannot come from the Arduino itself).

Is there something such as a very low power 555 timer chip that can provide a ~1 second pulse about every 5 minutes? The accuracy of either of these values is not critical. If it provides a 0.8 second pulse every 6 minutes, that's fine. The only constraint is that the device should be as simple as possible and consume VERY low power (I don't have much to spare). Is there anything better than a low power version of the 555 chip?

Any suggestions?

By "low power" I assume you mean the oscillator draws little current. A search of "low current oscillator" on google resulted in this: http://www.discovercircuits.com/O/o-astable.htm A whole bunch of astable oscillators.

This one looked promising:

http://www.discovercircuits.com/DJ-Circuits/3NAOSC1.htm

Once you're dead, you're dead man...

/get the paddles...CLEAR!

;D

IMHO the issue is not "low power" but "slow". The time constant of 5 minutes is hard(ly) to realize by a pure analog circuit. Most likely a multistage counter will be needed, too. 75HC4060 can devide by 16k so 20ms ticks will do... It will draw around 8 uA.

An Ultra Low Power micro controller as the MSP430 uses 0.7uA in standby! Atmels picoPower variants @1.8 V use the same 650nA in "power save mode", so the datasheet says. And yes, they are Arduinos!

Once you're dead, you're dead man...

Bah! I was about to post a clever comment :P

taking a quick look at the datasheet, it seems most arduino microcontrollers can enter a power down mode, and then be woken up by a watchdog timer interrupt. the WDT won't be for as long as you want, but you could just count how many times it has been woken up and do what it needs to do on the nth time.

Once you're dead, you're dead man...

/get the paddles...CLEAR!

Exactly what I was thinking, I chuckled at the topic title. I decided to actually be helpful though. ;D

chip that can provide a ~1 second pulse about every 5 minutes

An RTC with alarm capabilities (e.g. DS1337) can do that for you. It is also extremely low power (microamps).

Okay, everything I’ve tried seems to be overkill for what I want. I need a simple circuit that gives me a logic low pulse about every 20 minutes. The duration needs to be <1 second. It doesn’t need to be a perfect pulse though, just something that swings to ground every xx minutes or so (where xx is preferably 10-60 minutes). It should stay high the rest of the time.

Is there something I can “hack apart” in my house that might do this? An alarm clock only goes “off” once per day . . .

you can sort of do it with a 555 555 and 556 Timer Circuits it won’t be reliable at the length of time you need, but as you say, you don’t need it to be reliable.

you can sort of do it with a 555

What on earth makes you think this?? The OP now talkes of "something between 10 and 60 minutes".

I kindly ask you to set up a 6555 or any CMOS-Version of the 555 and try 10 MOhms and 100uF. I am not really sure it will not work, but I have some doubts...

I solved the problem. I used a low voltage version of the 555, with a 470 uF capacitor and a 1 M Ohm and 22 k Ohm resistors. I have it running in an astable mode and I get the logic low pulse I need every 6 minutes. It's a little more frequent than I'd like, but it will work.

Oh, and the low power 555 draws just 0.1 mA when running. That's perfect!

I think this approach was in your first posting already. So what had been the problem in the first place? Not knowing that CMOS versions of the 555 existed?

In fact as you might have seen, I also do not like something containing a 470uF :slight_smile:

I also do not like something containing a 470uF

why not? I would have looked at using 1kohm and 0.1F myself, high capacity low voltage gold caps are quite small and cheap.

I have not exactly looked into the values. The background of my concerns is, that electrolytes have noticable leakage: There is internal current discharging them, similar to cheap rechargable batteries.
When you charge a high capacity cap via 2MOhms @ 2V= 1uA, this could just be enough to uphold its charge level, if ever :slight_smile:

Edit:

I looked it up now. Your goldcap (5.5V 0.1F) might work, its leakage is < 30uA; charging through a 100k @ 4 volts will give 40uA.

However the full charge time will have nothing to do with your expectations,und will heavily depend on environment temperature…

O.k. 1k can provide 4mA, which is definiitely fine. 1k*0.1F = 100 s = around 1.5 minutes.
A .1F goldcap will be around 3 USD.

Maybe 10k and a .1F goldcap will be compromise…