Dropping 9v to 5v @ ~18mA: Passive or active?`

I have a circuit which, intermittently, needs 9 volts and several hundred milliamps; but for nearly all of the time (> 99.9%) it will be idling, drawing around 18mA (one LED, one ATTiny, not a lot else) at 5v.

I've breadboarded etc. my circuit using a dual 5v/9v common ground PSU, but now I want to finalise it into a form where it will only be fed via a 9v PSU (most likely, a cheapo wall-wart type).

My question is, what's the best way to get my 5v/18mA power?

  • Simple voltage divider?
  • Passive regulator, i.e. 7805
  • Active buck converter?

For this particular circuit, saving a few mA is not as important as being small; however, I eventually plan to do a battery powered version, which will (in theory) only draw a few micro-Amps at idle, then between 10-20mA for about 3 seconds, then back to idle (the 9V side of the same circuit will draw its power for around 0.5-0.75 seconds at the same time the "brain" is drawing 20mA.

All thoughts appreciated, thanks.

-Simple voltage divider? Bad, voltage changes with load, will not work.

  • Passive regulator, i.e. 7805 Better, but will have to dissipate the (I * (9 -5)) watts of power. That said, I use that type regulator (LM1117) a lot for low current applications. Cheap and take little space.

  • Active buck converter? Best(??), more efficient than 7805, but maybe more expensive in cost and real estate.

Do you have any analog circuitry running at 5V? If so the lower noise from a linear regulator
would probably clinch it (lower cost too).

Since not battery powered the increased battery life for a DC-DC converter isn't very relevant.

Voltage divider is for measuring, not powering. You might consider a zener shunt
regulator (resistor/zener divider), but not a resistor divider. Zener regulation is pretty
poor though, and shunt regulation is more power wasteful anyway.

The only 5v analogue circuitry I have is a capacitive sensor, on a 1M resistor, it’s drawing so little power my multimeter couldn’t even pick it up on its uAmp setting. Based on I*(9-5) = 0.08W, I reckon the linear regulator’s barely going to get warm… as I have one here, that’s what I’ll use this time.

Thinking about the battery operated version; I’d do away with the “constantly on” LED, and I believe using level-triggered interrupts and sleep mode, it’s possible to run a Mini Pro at around 6uA (while asleep obviously… when woken it will take the usual ~18mA). However, I’m wondering if I’d be better off using a 4.5 or 6V supply (3x or 4x AA alkaline batteries), and use a buck/boost converter with a large-ish capacitor to provide the 9v…

Or, as the power draw on the logic circuit is so small… maybe have 6x AA batteries to provide the full 9v, and a “centre tap” to take logic/led power off just three cells. That’d work, wouldn’t it? And no need for any power regulation. Sure, 3 of the batteries would go flat marginally quicker than the other three, unless (over such small draws, they tend to even out).

The application, by the way, is a touch-sensor doorbell. The bell unit takes 850mA @ 9v as long as the coil is being driven, and my delay is currently 0.6s. The ATTiny85 drives it using 1 pin & a BC337 transistor, it also switches two LEDs, one a constant backlight LED, the other a “bell pressed” LED which takes over for a few seconds when the sensor fires. I’m happy with my circuit now (other than the power regulation for the 5v side), just need to finish making the hardware…

AdeV:
For this particular circuit, saving a few mA is not as important as being small; however, I eventually plan to do a battery powered version, which will (in theory) only draw a few micro-Amps at idle, then between 10-20mA for about 3 seconds, then back to idle (the 9V side of the same circuit will draw its power for around 0.5-0.75 seconds at the same time the "brain" is drawing 20mA.

The information on your 'project' that you have left out is critical to getting a good answer.

There are some 'projects' where switched mode converters are a bad idea.

For such small currents where battery life is irelevant a linear reg would be fine. And it'll only have to dissipate about 70mW of power - no heat problem.

And it's clean.

Allan

srnet:
The information on your 'project' that you have left out is critical to getting a good answer.

There are some 'projects' where switched mode converters are a bad idea.

Hmm. OK, point taken, but since I wouldn't know what information is "critical" and what information is not... I thought I'd covered the salient points (current usage, primarily, and the need for two voltages); so I guess you're thinking of something that I've not?

So rather than just suggest I've missed out critical information; if you can make any suggestions as to what information I need to include that I haven't (or didn't in my original post), that would help me learn. Just telling me I'm wrong and stopping there helps no-one.

allanhurst:
For such small currents where battery life is irelevant a linear reg would be fine. And it'll only have to dissipate about 70mW of power - no heat problem.

And it's clean.

Allan

Thanks Allan - that's the route I'm going to take for this (mains-driven) project. I should do it in SMD really, for the low height & small footprint, but I'm going to prototype it with through-hole components for now, as I have them...

The battery one I'll come to another day, as that requires a second external IC & may not even need an Arduino (or ATTiny in this case)...

Cheers!
Ade.

AdeV:
So rather than just suggest I've missed out critical information; if you can make any suggestions as to what information I need to include that I haven't (or didn't in my original post), that would help me learn

You have not described what the project is or does.

srnet:
You have not described what the project is or does.