I want to use a power bank as battery back up for a Raspberry Pi. The power bank has to have pass through charging so that I can leave the power bank connected to the mains while the Pi is powered from the output. I gather not all power banks can be used that way.
Does anybody have any general or specific advice about how to choose the power bank?
Yes, that's the problem. I guess you could always ask before purchase. Anyway, I can confirm that RS 10 Ah unit /does/ work as a UPS. I have one of my Raspberry Pi PCs working on it right now. Unplugged the main, and the RPi carried on working. Replugged the mains and the device continues to supply power to the RPi.
The spec says, "Input power source rating: 500mA @ +5V"
I'm not sure if that limits the input current to 500 mA or if it is saying 500 mA is required from the power source. If the unit won't draw more than 500 mA it won't power the rPi at max current draw without draining the battery, no?
is currently the best bet at GBP 48.99? This charges at 350 mA, but then again it is a UPS and can run off mains, batteries or solar cells and seems to accept supply voltages from 5v - 18v.
This looks a bit more reasonable when you know that Amazon flogs 2500 mAh 'phone chargers, i.e. a pocketable 2500 mAh hard-cased Li-poly battery for anywhere between GBP 9 - GBP 25 and only the more expensive ones have chargers and a micro-USB output connector.
martin@ | Martin Gregorie
gregorie. | Essex, UK
The instructions are along the lines of: if the input source can supply
1A or more the unit will draw 1A otherwise it draws 0.5A. Not sure how it determines that. Mine has been running an RPi B with Wi-Fi and a GPS attached for 20 hours now, and as soon as I started it the "charge greater than 80% LED" was flashing, and it still is. When running, that RPi takes 0.4A. My USB ammeter is only USB-A connectors, so I can't easily measure the current into the unit, or know how it determines whether to draw 0.5A or 1A.
USB devices negotiate current draw with the host when connected. Default is 100/150mA (USB2/3) maximum 500/900mA. If the data pins are shorted the device knows it's a dumb charger and can draw 500-1500mA. ISTR there's a recent extension (USB3.1?) to this that provides 2A at 5V but I don't recall the details, and of course Apple chargers/devices have their own method.
There are (at least) 3 different schemes for the data lines to indicate it's a dedicated charging port. Shorting (usually done with a 200 ohm resistor) is one way. Another is 2.0V on one and 2.7V on the other data line to indicate 10W output (2.1A), or 5W output if the voltages on the two data lines are are swapped. Another is 1.2V on the DP line. There are USB PSU chips which detect which method the USB device is trying to use, and adapt the effective resistance and voltages on the data lines to match it (such as Texas Instruments TPS2511).
However, the mess with USB charging standards (the Chinese standard, EU standard, Apple, and USB Consortium standard) mean many devices nowadays work out the current available by steadily increasing the draw whilst monitoring the voltage drop, and back off a bit when the voltage drops below some device specific threshold (such as 4.75V), or if the supply cuts out. This can mean that a low spec charging cable can severely limit the ability to draw max current from a supply even when both the device and the supply are capable of high current use.
10W (2.1A) can be available from USB chargers to Chinese and EU standards.
The USB consortium standards allow for up to 1.5A on USB2 dedicated charging port I think, but they also have standards for up to 90W or 100W for laptops but using different connectors and higher voltages. I've never seen these in use anywhere, possibly because you have to pay a license fee to implement their standards, which I don't think is the case if you use the Chinese or EU dedicated charging port standards.
[email address is not usable -- followup in the newsgroup]
I once got a pack somewhere (I think it was included with an old HD 808 camera I ordered via ebay) that exibits this behaviour. I was more or less happy until I found out it did not recharge the battery when something is connected to the output.
Another one for the checklist: does it recharge when power comes back?
FWIW: I have an Anker 15,000 mAh unit (bought from amazon) and I powered a Pi from it last week - noticed the following morning it was down to 1 dot, so plugged it in to charge - and it started charging. However it's taken 3 days to charge fully - while powering in the Pi.
That, I suspect is the biggest issue with these units - the ability to take 2x the output capacity to enable charging in a sensible time span...
I suspect it would go off an stay off when it ran out of juice though.
My 2c is to choose the appropriate sla/wet/agm/lipoly/???? to provide for what ever duration blackout/relocation period you want and then mate it to the appropriately battery charger and USB power supply.
I looked at the manual and they say you can do this pass through charging, but they warn you that it "may" reduce the battery life in the Anker unit. I found in the manual for one of the smaller Anker units they tell you not to do it because of harm to the battery life.
As a 72 hour charge for that battery should (in theory) need just 208mA, a 2 amp PSU should have plenty of capacity for running both.
Charging a LiPO or Li-ion battery is fairly complex: a good charger provides constant current at a 1C rate (a 1000mA cell is charged at 1 amp) until the full pack voltage is reached and then constant voltage until the current falls to an arbitrary small value, when the charger reports fully charged and turns off.
My guess is that the built-in charger was designed on the assumption that there is no load on the pack when its being charged, so the running RPi is simply eating some of the 1C current, meaning that the battery is being changed at a rather lower rate than expected. That shouldn't hurt the battery as the charging circuit will just dimly sit and wait until the battery gets to the end of the constant current phase. After a power outage I imagine that the charger restarts in constant current mode.
martin@ | Martin Gregorie
gregorie. | Essex, UK