What's consuming all the power on Mega board?

The digital blink example is loaded on my Mega board.

The board is powered by barrel jack, Vin measuring 12.1v using wall wart, regulated DC style.

My meter (in-line at the jack) shows consumption of the Arduino Mega board to be 60mA when the LED is off and about 65mA when on. It's not a fast meter... so I might not be catching the extremes well...

No matter, my question is: what is sucking the power?

The 1280 shouldn't be... not at 16MHz full tilt. The power LED has a 1K in series so cannot be more than 5mA contributor. Similar for pin13 LED. Is the USB chip that hungry? The 358? Or do the dual voltage regulators really consume that much current... (I thought they might contribute 1 mA or so of actual current draw).

I was hoping to embed this Arduino board into a car such that it would be powered all the time. 60mA is about three times higher than I'd like to see constant draw for this device.

If I were to cut a trace to lighten the load, and really only need the MCU running in the target... and maybe less support circuitry, where do I best go hunting?

forget linear voltage regulators when you want to go in a car ;)

you have 14.4V in a car when the engine is running, so the linear regulator kills about 9.4V to get the 5V. this will let it get hot and your efficiency is somewhere...

take a DC/DC converter and there you go with 14V and nearly no current, but 5V and some current on the other side :)

have something here: http://code.google.com/p/multidisplay/wiki/Powersupply

Your mega board has other consumers of current. The +5vdc voltage regulator has some 'overhead' current consumption, the FTDI USB serial converter chip consumes some current, etc. Your reading sounds about right to me.


have something here: http://code.google.com/p/multidisplay/wiki/Powersupply

is a linear regulator... with some surrounding robustness.

Usually, I put a 150 Ohm resistor (wattage sized to load) in place of that choke, then run in series through a 1n4001 then into the regulator, caps either side. LM7805 is robust enough to handle the big dumps experienced in the car operations. No zener shunt required, so long as inrush current is limited - which the 150 resistor does.

I've used lotsa linears in automotive applications like this. LM2931 is a better choice for auto app's tho. I might have some kicking around, but if I replace the regulator on the Mega board, is the consumption going to drop?

Just wondering what on the Arduino board is so hungry.

It's not so much tolerating the 14.4 charging voltage I'm concerned about, its the off-state 12.6 battery sitting there being drained I'm concerned about.

A basic arduino w/328 is 30-35 ma. I have to think there's something odd driving your current up.

If there's really nothing else connected you can reduce the clock speed using the pre-scaler to cut power use at the expense of mips.

I think the point trying to be made above is because of the large volt drop you need, a linear regulator may not be the most efficient solution.

Wondering out loud here, what if I put a series of a few 1n4001 diodes ahead of the Vreg input. Each one would drop a volt or so... and is one less volt the regulator would have to blow off as heat. I don't think the diodes would heat at all with these currents.

The large voltage drop may cause you to draw more power, but it shouldn't directly affect the current draw. It dissipates extra watts ( (12-5)*0.065 for the 65mA you're drawing. Nearly half a watt, which is a lot compared to the (5*0.065) = 1/3 W that the board as a whole is using, but doesn't particularly change the 65mA number... (adding some diodes doesn't help the wasted power, it just distributes the wasted power across more components (although MAYBE the regulator power would be less for a smaller differential. Hmm.))

(now, a switching regulator would change the equation. With any sort of linear regulator (including your string of diodes) you're using 12V*0.065 mA, or about .8 W to power your 0.3W-consuming Arduino. If you were to find an 80% efficient switching cell-phone regulator that outputs 7V, you could cut this down to about .6W total, which is a pretty easy fix for 25% improvement.)

Checking a couple datasheets, it looks like each of the 1280, FTDI, and voltage regulator chips consume "about" 20mA. Add a couple mA for the op amp and power LED, and you come out pretty close to the 65mA you're seeing. (except that at least some of those are "maximum" current draws, and you'd expect/hope that a random actual case would be better.)

The Arduino boards aren't really designed for low power consumption; you could choose a regulator with lower quiescent power, make sure the FTDI is turned off when not connected (it might be programmable to do that anyway?), and add some software support for assorted low power modes. (now THAT would be an interesting variation/clone for someone to produce!)

make sure the FTDI is turned off when not connected

The Seeeduino mega does this by powering the FTDI chip only, directly from the USB power pins, regardless of how the rest of the board is powered. Makes sense, if you are not connected to the USB why power the FTDI from external power.

The Seeeduino design doesn't have a auto-power selector circuit, rather it has a user manual switch to allow switching from USB/external voltage for the rest of the board, which can also acts as a power on/off switch if you only have one voltage source connected.


I like the ideas.

find an 80% efficient switching cell-phone regulator that outputs 7V

Further yet, find a 12 to FIVE volt switcher (I think I've got one kicking around for some old gadget / cell charger) and just connect its output to the Vcc rail directly. Going around the whole on-board linear section.

And, cut the FTDI supply and connect directly to USB.. Brilliant. 20mA saved during run mode. And still able to upload new code with the target installed in the car.

Why wouldn't Arduino brand think of that? Had I know Seeduino does this, I'd have picked it over Arduino. Now I just have to get surgical.

Chiseling off the USB chip and the regulator gets me down to just the MEGA consuming 20mA at full clock... which is acceptable for me to leave connected to car battery all the time as a constant draw.

Plus... the load of my peripherals, but I can manage these ... keeping them switched off under the control of Arduino.

Chances are that your Arduino need not run all the time --> look at the sleep modes. This can bring down power consumption a lot. With regard to Arduino vs. Seeduino: the Seeduino clearly has a much superior design. Especially it gets you rid of the strange non standard connector spacing.

Cheers, Udo

Why wouldn't Arduino brand think of that?

I think it's a historical thing. It's always easier to improve on an existing design rather then a initial design where there are not too many examples to copy from. I think the time line goes something like this for Arduino/Seeduino designs. 1. First Arduino's used mega8 chips and RS-232 serial link. 2. next generation Arduino's used USB serial link and a two way jumper pin to select USB or external power option. 3. Current Arduinos use auto-voltage selector circuit. 4. Seeeduino uses manual switch to select power, and powers USB chip via USB power only, at least on their Mega board (not sure about their 328 board). It also adds a option to solder a addition connecter that is on .1" grid, while still retaining the Arduino 'standard' connector spacing.

This may be oversimplification, but I think it gets the basic evolution correct. There are many manufactures of Arduino clones or redesigns these days and there are many variations for powering the boards and what kind of serial link if any are built into the board/module.


It's not so much tolerating the 14.4 charging voltage I'm concerned about, its the off-state 12.6 battery sitting there being drained I'm concerned about.

I make it 3 weeks to a flat battery for a 33Ah one... Not good considering automotive batteries are not designed to be stored other than at fully charged state and lose capacity over their lifetime... However if you use the vehicle every day its not as issue. A few watts of solar panel could finesse the problem ;)

That must be a typo 33Ah. !?

And yes, the wacky non-on-center 0.1" spacing of one side of the board is just stoooooopid. That just p*sses a guy off.

If you check around the online electronics shops you'll find a regulator that will viciously filter automotive power to +5volt using a switching regulator that will work much better. Automotive power can kick out huge voltage spikes and vary wildly at times.

Also did you put all the non used pins in a defined stated, read somewhere that also accounts for a bit off power consumption.