How are people powering battery-driven projects?

Exactly Oric...(4) AA Alkalines start out at like 6.75 volts, and it is enough at first for the 05 regulator, but will drop out somewhere too soon before the batteries are depleted....If I wire the batteries directly to the 5v bus it exceeds the processor's voltage spec of around 5.5v. I have a simple project, but the lcd display has a large drain led backlight. Most cases don't come with capacity for (6) AA or even AAA...so I think it is 9v for me at this time...I need to let a supplier know for prototyping samples. I suspect a 9v won;t last long.

Coding, I said almost too low, because I have bench tested 4 batteries through the vreg. It works fine bespite being below the 7v spec. But will drop out before we want it to....

If this were a 3.3v design, all would be fine, just use (2) AAs and be done with it. But 5v is a bugger.

I use cell phone battery chargers that I hook up the to the USB port. At the moment, I'm using the EZOPower 5000 maH charger that provides two separate USB ports. In the past, I let my UNO R3 run with the blink sketch and it lasted for 4 days or so using no special techniques for saving power. I picked mine up from Amazon: http://www.amazon.com/EZOPower-Universal-External-Rechargeable-Generation/dp/B004I0KY68/ref=sr_1_14?ie=UTF8&qid=1355430888&sr=8-14&keywords=ezopower+5000maH+battery.

Yes interesting...I may end up just offering a DC input jack and let people power it how they see fit...like you suggest or otherwise...thanks

Yes interesting...I may end up just offering a DC input jack and let people power it how they see fit...like you suggest or otherwise...thanks

To answer two questions, if you let them "power it how they see fit" and don't include a v.reg,
be certain they'll blow the whole thing. Risk goes up by 10X if you have a power jack on the
board. They can connect anything.

Batteries direct to microcontroller may be ok in a very carefully controlled situation and where
you're not using A/D converter, but risky for a general case. I think it's better to use a v.reg
and have a good overhead margin in voltage so it keeps running as the batteries discharge.
With 6 NiMH AA-cells and an LDO v.reg, the cells are depleted when down to about 1V, which
is where the LDO v.reg also drops out. 4 alkaline cells are also a usable choice.

NiMH AA-cells have up to 2500 mA-hr, the measly 9V batteries maybe 1/10 that. I might use
a 9V battery with a microcontroller operating with a 32-Khz xtal.

Batteries direct to microcontroller may be ok in a very carefully controlled situation and where you're not using A/D converter,

So using an AVR's A/D converter when powered directly by batteries is a bad idea. Why?

but risky for a general case.

Why?

OP,

If you worry about wasting power with 1.55=7.5V arrangement or 1.56 (easy to find such housing), you can use a switched-mode dc-dc converter to converter instead of linear dc-dc converter such as 7805. Switch mode converters have higher efficiencies so you are not losing power due to a voltage drop from say 7.5V to 5V. Look it up.

Yeah, for that matter, Pololu has some nice "boost" regulators that will run off 3 or 4 rechargeable
cells, and boost the voltage to 5V and higher, and continue to operate as the batteries run down.
[disregard the step-down devices].

Oric, maybe lost in translation, my statement meant I would provide a jack to power the device through the built-in 78M05.
I will be using Nano's, they are cheap in compatible version, why reinvent the wheel? Of course that could change if I find another cheaper option just as good. It'll be nice to keep the USB port for updates in the field.

There are certainly lots of options and boost and buck converters are great, but when producing a mass produced piece, you have to consider all the costs and complexity as well of course. Like the power switch discussion...is it worth it to eliminate an on-off switch, if it means adding another button anyhow, along with a couple transistors and a bunch of other glue parts?

I want to keep this as low cost as possible.

cornwallav8r:
It's a quandary...a 4-cell AA battery is too high voltage for the 5v input, but almost too low for the 78M05 regulator input, adn will drop out fairly quickly despite battery capacity left. A 9v doesn;t have the endurance of multiple batteries.
I do see however there is a 5v switching arrangement on the Nano at least, to switch, from the USB 5v to the 5v main bus? Not sure that is an alternative...

So what are people doing in this regard?

So nobody can blame me, I won't suggest that anyone do this, but I personally have run an Uno R3 board on 4 AA lithium cells in series. As you may or may not know, a brand new lithium AA cell is around 1.80 volts DC, which is 7.2 volts for 4 of them.

The 4 AA cells in series power the board's "5 volt" supply directly. It does not go through the regulator.

I've made several projects powered like this and they are running fine.

Krupski:

cornwallav8r:
It's a quandary...a 4-cell AA battery is too high voltage for the 5v input, but almost too low for the 78M05 regulator input, adn will drop out fairly quickly despite battery capacity left. A 9v doesn;t have the endurance of multiple batteries.
I do see however there is a 5v switching arrangement on the Nano at least, to switch, from the USB 5v to the 5v main bus? Not sure that is an alternative...

So what are people doing in this regard?

So nobody can blame me, I won't suggest that anyone do this, but I personally have run an Uno R3 board on 4 AA lithium cells in series. As you may or may not know, a brand new lithium AA cell is around 1.80 volts DC, which is 7.2 volts for 4 of them.

The 4 AA cells in series power the board's "5 volt" supply directly. It does not go through the regulator.

I've made several projects powered like this and they are running fine.

Well that is well above the absolute maximum Vcc voltage limit of 6.0 per the AVR datasheet. How do you account for not having a bunch of burned out boards? And it's magic is not a proper answer. :wink:

Lefty

Krupski... I kinda expected that answer :slight_smile:
Yeah, it doesn't surprise me that it'll work over spec.
But the important thing here, is I want to make a whole bunch of these...and they need to work on every single item, not just "that one I tried on the bench". Not surprised at all it'll work for awhile like that in testing.

@liudr:
Just curious: What would be some DISadvantages of switched-mode dc-dc converters? Aside from cost, it appears they are better than LDOs in most situations, am I wrong?
E.g., I am considering powering my circuit with a Li-ion battery and using the TPS63000/TPS63002 Buck-boost regulators; they seem to have very good efficiency in both buck and boost modes.

I've run uCs off 3 AA batteries, ~4.5V, even a '2560 based board @ 16 MHz.

I've also run 3.3V Promini @ 8 MHz from a single LiPo battery.

I don' think I saw what the application is here. Haven't seen what the need is for boost converters unless the goal is to run from a single AAA or AA size battery.

The single biggest issue with battery projects is how long do you want the batteries to last
before they have to be replaced.
That effectively defines what sort of batteries are needed, and hence what voltage batteries are needed.
Id stay away from any kind of arranagement that needs voltage converters of any kind, because although they are more
efficent that linear regulators, at very low power levels , they can consume more power than the micro.

CrossRoads:
I don' think I saw what the application is here. Haven't seen what the need is for boost converters unless the goal is to run from a single AAA or AA size battery.

Well I'm sure there are other, more practical applications but...

I've been designing and selling APRS Trackers (for the non-ham, think 'LoJack') for high-altitude balloon flights, where the goal is to be small and light-weight.

My latest version is powered via a TI TPS61202 (so 2xAA batteries) that drives the on-board 5v Radiometrix 300mW transmitter (HX1) as well as a 3.3V LDO regulator that powers the ATmega328P at 8Mhz (and the ublox GPS).

For a look at the prototype (and get a copy of the design files), check out Dropbox - ProjectSB - Simplify your life

Running off 2xAA batteries cuts weight by at least 30 grams and so far, my latest test has logged 13hrs (and is still running) off 2xAA (L91's).

(Plus I've had people interested in using the tracker w/ LiPo's for terrestrial use - things like parades and SAR work).

Regards,

Brad
KF7FER

PS If you're interested in watching paint dry, you can see the prototype running a battery test at Login – aprs.fi – live APRS map

I'm looking to expand my experience with boost and buck converters both, the TPS61202 looks interesting. Are you hand assembling? How tough is it to solder that package, and are you soldering the thermal pad?

CrossRoads:
Haven't seen what the need is for boost converters unless the goal is to run from a single AAA or AA size battery.

A little out of context, but I just bought a Pololu Zumo robot, and on their Zumo shield they use
4 AA-cells powering a boost converter outputting 7.45V, rated about 1A.

The 4 AA-cells also directly power an h-bridge to control the motors [beautiful little gearmotors],
and the boost regulator provides power to a standard Arduino UNO/etc, board that plugs in on top.
You end up with a neat tracked mini-sumo that's less than 2" tall with everything running off 4 NiMH
AA-cells.

Jack-

When I got this set of PCBs back from the fab, I found every excuse for the first three days to not start. I honestly was terrified at soldering the TPS61202.
I always hand-assemble my boards, though I have used a hot-plate before.

But I ended up using an Aoyue hot-air rework station and solder paste to handle just the TPS61202, then I hand-assembled the rest of the board.

It worked the first time even. Ok, the boost converter did. I got the wrong LDO and had to swap it, but otherwise the board was fine - especially for "Revision A".

I probably should add that the guys in the UK and Australia doing HAB work have been using boost converters for some time (see New boards pAVA Rev6 | Ava High Altitude Balloon Project) so I got a lot of help from them (thanks M0UPU and VK5QI!)

I hope this helps,

Brad
KF7FER

PS I have to double-check the test data but I believe we are at 22hrs and counting (battery voltage is still at 2.59v). Those "ultimate lithium" batteries are expensive but when compared to generic "big-box" store batteries they run much longer (I got 7hrs off of 2xAA "Kirkland" batteries) and weigh almost 1/2 as much (15g vs 25g).