10-12 years ago, you could buy a USB battery for 10-20 euros, and use it to power a Yoctopuce experiment without asking yourself too many questions. Unfortunately, today it's very hard to find a USB battery that can power a Yocto-Wireless or YoctoHub-GSM-based experiment. So we decided to invest some time, and money, in trying to find a good one.
The problem with current USB batteries
The Yocto-Wireless and YoctoHub-GSM sleep function
The Yocto-Wireless and YoctoHub-GSM both have an interesting feature. They can go to sleep and wake up at regular intervals. When they wake up, they usually contact a server, do what they have to do, such as communicating to the server the values of all the sensors attached to them, and then go back to sleep. When they are asleep, these hubs consume only 15-20µA, compared with 60 to 200mA when they are working. This sleep mechanism therefore minimizes the average power consumption of a battery-powered Yoctopuce experiment. At first sight, the most practical type of battery for this kind of use is USB, but....
USB batteries
The most common USB batteries suffer from two problems. They are generally designed for a specific purpose: recharging portable devices such as mobile phones, tablets, and computers. Moreover, they contain electronic components that consume some current, even when the battery is doing nothing. So, if you wait long enough, a USB battery will discharge completely. To ensure that their customers don't end up with an empty battery when they need it most, battery manufacturers have developed a little trick. If the battery notices that little or no current is being drawn from it, it shuts down completely, thus preventing it from draining itself.
Of course, this doesn't suit Yoctopuce customers who use Yocto-Wireless and YoctoHub-GSM: once the hub has gone into sleep mode, the battery switches off completely, and the hub can't wake up for lack of power. 10-15 years ago, you could find entry-level batteries without this protection feature, but now it's mission impossible.
An old battery without the protection in question. Had we known, we would have bought dozens of them ten years ago.
So we decided to embark on a quest for the perfect battery for battery-powered Yoctopuce experiments. Typically, we're looking for a battery system with a USB output and an explicit ON / OFF function, rather than an automatic switch-off system. And this week, we present two candidates.
The UPS-15-POE router from IPS
We've been looking at mini-UPSes for a while, because since a UPS is supposed to supply power no matter what, there's little chance of it deciding to switch off without a compelling reason. We found the UPS-15-POE model from Intelligent Power Sources. It comes in a 16x11x3cm box, has a USB-A output, a 5 and 12V configurable output, an ON/OFF button, and a 100-240V AC input.
The IPS UPS-15-POE
Note that we suspect that IPS is not the real manufacturer of this device, because you can find a host of other devices on the Internet that look suspiciously like it, but are sold under other brand names. Nevertheless, the UPS-15-POE claims a capacity of 8800mAh. So we set up a little test with a Yocto-Meteo connected to a YoctoHub-Wireless-n, which wakes up every 10 minutes to send its measures to a VirtualHub for Web hosting the Yocto-Visualization for Web application, enabling us to monitor the experiment. In this configuration, when awake, the YoctoHub-Wireless-n and its Yocto-Meteo consume around 90mA.
_1.svg "The UPS was able to power a YoctoHub-Wireless-n and a Yocto-Meteo during three and a half days, at a rate of one wake up call every 10 minutes")
The UPS was able to power a YoctoHub-Wireless-n and a Yocto-Meteo during three and a half days, at a rate of one wake up call every 10 minutes
The experiment ran for 3 and a half days before the UPS ran out of batteries and shut down. We recharged it, and tried a slightly different experiment: monitoring with a Yocto-Volt the output of the UPS under no load. The experiment lasted 4 days.
-EN_1.svg "Without any load, the UPS only lasts 4 days anyway")
Without any load, the UPS only lasts 4 days anyway
This UPS-15-POE model meets the basic requirement of behaving like a battery with a USB output that doesn't switch off automatically, but its high internal power consumption means it can't last more than four days. Of course, the fact that it plugs directly into the mains means it's best suited to indoor applications that need to keep running even in the event of a power failure.
Omnicharge 20C+
On paper, and despite its price tag of 300 euros, the 20C+ model from Omnicharge has everything it takes to handle our case study with flying colors:
- An auto power-off which you can disable,
- Two USB-A outputs,
- 20000mAh capacity,
- and the option of recharging with a USB charger or a solar panel.
So we rushed out to buy one.
The Omnicharge 20C+ (fingerprints are virtually impossible to clean off)
It comes in a rather heavy box, measuring around 13x12x2.5cm. It has a very nice OLED display that shows, among other things, the battery charge percentage, input power and output power, bearing in mind that the output power displayed includes the internal consumption of the box. We checked that the battery can be charged using a solar panel. We disabled all the options we didn't need to limit the battery's internal consumption, and, full of hope, ran the same experiment with a YoctoHub-Wireless-n and a Yocto-Meteo. We soon noticed that the battery was draining at an alarming rate, with a displayed consumption slowly oscillating between 1 and 2 W. It wasn't long before the Omnicharge 20C+ shut down, out of battery, after 2 1/2 days. Imagine our disappointment.
_2.svg "Our experience with the Omnicharge 20C+ lasted just over 2.5 days.")
Our experience with the Omnicharge 20C+ lasted just over 2.5 days.
With a no-load consumption of 1W, and assuming that the internal electronics operate at 5V, this implies a no-load current of around 200mA, which is frankly excessive. In the best-case scenario, the battery would last no more than 100 hours, or around 4 days. Note that the claimed 20000mAh corresponds to the battery's nominal cell voltage of 3.7V, not 5V. Actually, The battery's real capacity which is claimed to be 71Wh.
We contacted Omnicharge customer service, which confirmed that the 20C+ was designed as a travel charger that must be switched off between uses, and was never intended as a long-term power supply. We wish the propaganda on their website had been clearer on this point.
That said, if you need to run an experiment in the middle of nowhere, the Omnicharge 20C+ may be a solution if you connect it to a solar panel able of supplying on average more energy than the battery and experiment consume, i.e. 2-3W, which is completely feasible.
Note that we've noticed a small bug: as long as the battery is switched on, it's possible to recharge it with any USB source. But if the battery level drops so low that the Omnicharge 20C+ decides to switch itself off, then it would seem that only a charger with the "Power Delivery" function is able of waking it up. In other words, if one of your solar systems based on an Omnicharge battery dies, you'd better go to the site with a spare battery that's been pre-charged.
Conclusion
For a first round of tests, it's a little disappointing: as much as we didn't expect miracles from the IPS model, we didn't imagine that the Omnicharge 20C+ would do even worse. Omnicharge has a 40C+ model in its catalog with a claimed capacity of 40000mAh, which, based on the performance of the 20C+ model, should be able of lasting 5-6 days.
Despite this inauspicious start, we're still looking for the ideal battery to power experiments based on Yoctopuce modules. If you have any suggestions for models to test, please let us know in the comments.
