When I searched, I ran across various estimates of what the power draw is of the Raspberry Pi Zero W under load. This recently became a subject of interest, in selecting a (hopefully) not too big (heavy, bulky) and not too small (run times) rechargeable battery for one.
The estimates I was seeing on-line were from .36 A (360 mA) to .4a (400 mA) under maximum load. After a little experimentation I thivink that’s a little high, but I wasn’t doing huge processor loads, so maybe it might hit that briefly. I was running some fairly intensive processes but in command-line mode over SSH with Wifi running, so no video at all, and the maximum I was seeing was .2 A (200 mA), so I’m guessing that’s a good conservative estimate of an average load with Wifi running.
Interestingly, when it was idling but connected to Wifi I was seeing just half of that, about .1 A or 100 mA. Idling with Wifi operating is what say, a PirateBox installation over Raspbian could be expected to be doing much of the time, between client connections.
So, say the average load for such a setup, Wifi enabled but idling half the time might be about .15 A or 150 mA. That means that disregarding inefficiencies, in theory a 2,500 mAH cell should power it in use for about 16 hours. Even at a solid 200 mA it would be 12.5 hours.
That’s a huge difference. Once upon a time I spent almost a week running a similar setup on a Raspberry Pi 2, with a USB flash drive and a USB Wifi dongle (sequentially) on a pair of the biggest rechargeable battery packs I could find, supposedly With very little real use load I could not get 12 hours out of either one, which tended to confirm my guess that the whole rig was drawing about an amp. Here we’re talking 15-20% of that, and that was just a few years ago. On that same battery pack the Raspberry Pi Zero might have run for two days or more.
For perspective, 200 mA at 5 volts is 1 watt. When’s the last time you worried about a 1-watt light being left on? I’m going to use these things for more 24/7 tasks…
Just FYI, in “halt” mode (shutdown from the command line) the Raspberry Pi Zero W seems to pull about .02 A (20 mA) with a good short USB cord, and closer to .03 A with a not-so-great longer, thinner cord.
One important point here
is that the standard for the old, standard, plain-vanilla USB jack is that it supply a good, steady 500 mA of power. These days many supply more. I’m told the original spec was for it to initially supply 100 mA until handshake, and then 500 mA, but I’m not sure I”ve ever seen one do that, because… why?
What this means,aside from battery life, is that pretty much any USB jack will supply plenty of power to run a Raspberry Pi Zero or a Raspberry Pi Zero W, at full load, even with Wifi and Bluetooth blazing away, and a good safety margin. Since the USB jacks on the Raspberry Pi Zero and Raspberry Pi Zero W are of the OTG variety, the same cable that powers it can be used for data communication as well.
– Robert the Wombat
A good article but some corrections are in order.
1. You have to consider the difference between the battery/cell voltage and the final voltage your boost converter supplies to the Pi when calculating how much current the battery is sourcing. For example, a Pi consuming 150ma at 5V (=0.75W) must draw AT LEAST the same amount of power from the battery. It will actually draw more because your boost converter is not 100% efficient. So if your Pi is drawing 0.75W at 5V, your battery must supply at least 0.75W at 3.7V which equates to 203ma. This is the number you need to use when calculating maH from the battery. Add in an extra 10% or so for losses in the boost converter and you have 223ma being delivered from the battery.
Therefore, a closer approximation (see below for more factors) is 2500maH / 223ma = 11.2 hours. This is 5 hours less than the original 16 hour estimate.
2. The more current you draw from a battery, the fewer maH you get out of it. The rightbattery.com (no affiliation) website lists discharge results for different batteries at different loads. So for example, a 2500maH NiMH Eneloop Pro battery (which is considered top of the line by some) can drop to 2472maH at 200ma and 2322maH at 2A. So the more you drive that battery the less you get out of it. Alkalines are notoriously bad for this.
And unless you are using top rated batteries with know controlled testing results, you should consider the capacity rating as being only for very small current draws (50ma or less). Therefore, drawing 223ma from your battery will drop the capacity rating slightly (much less than if you were trying to drive 1A or more from it).
3. The capacity ratings on the rightbattery.com web site assume that a battery or cell is “dead” when a the output voltage for a 1.2V cell reaches 1.0V. A battery pack with 3 x 1.2V cells will be considered “dead” when it reaches 3.0V. But it is possible that your boost converter can cut out at levels higher or lower than this. For example, if your boost converter drops out at 3.2 volts, you may only get 75% of the rated capacity.
4. As batteries get older and are charged over and over again, they lose the ability to hold charge. It is important to factor this into the design.
With all of that said, I tend to cut the rated amount of time in half. So instead of a 16hour theoretical capacity, assume you will get at least 8hours and maybe more. But bank on the fact that it will be significantly less than the theoretical maximum.