Ravpower the ultimate Pi UPS?

RavPowerAnd now including -  the MONSTER version. Just a short while ago, I just bought this 16750maH USB battery unit (charger) with no prior knowledge of the company - on the recommendation of my pal Aidan who does our PCBs for Esp8266.

I have a specific requirement – which many of you will recognise.

1. I don’t want to have to tpush buttons to turn it on

2. I need it to work WHILE the unit itself is being charged

3. I need lots of POWER.

Some time ago I used a similar looking cheap USB charger to power a Raspberry Pi2 and it worked well but I noted it would not really charge any higher than it already was when deployed, while running the Pi. Despite this it would not lose charge either and is still doing sterling service as a UPS for one of my older PIs.

When I wrote that original article, several people wrote in to say that many such devices will NOT simultaneously charge and discharge – and true enough – there are problems with most of the alternatives I’ve tried – some work but need a button pressing to turn them on. And that means if they run out of power due to a long outage – when the power comes back on you’ll be getting nothing!!  On the web you’ll find all sorts of contraptions as alternatives for use as UPS systems – from the big monster mains uninterruptables to 12v batteries with chargers and switched down-convertors – all very, well, MESSY. And of course today the Raspberry Pi 3 for example needs peak power higher than ever before, so at least 2 amp capability per output is needed.

And so it was that I got quite excited when Aidan said he’d found an ideal candidate for the job, a heavy duty USB power pack with no less than 5 cells in it, capable of outputting up to 4 amps (one output has a maximum of 2.4amps – the other 2.1amps!!  I went to Amazon and sure enough £20 inc postage – what a bargain. But I’m in Spain and so I tried Amazon.es

Sure enough – there was no free postage on this in Spain but the OVERALL price inc. postage came out to be the same as the UK overall price – around £20. Bargain.

I ordered one and amazingly it turned up the next day here in rural Spain. And… it works! When I received the unit, I took my flat phone out with me for coffee into the village along with my sparkly new supply which came with 2 short USB leads (short is GOOD). By mid-morning (despite having just taken the unit out of the box and not charged it) my phone was full and the charger was still registering 3/4 full.

When I got back I plugged my PI3 into the unit – and waited for it to boot up - rock solid red light (which on the Pi3 indicates good power).  I tried putting the charge lead for the Ravpower unit in and out repeatedly – no  problem. Aidan had already tested this running a Pi until the pack was flat (more than a day!!!) and then plugging a charger into the pack. The PI came back up reliably.

The unit looks good – NICE finish – and has two outputs as indicated above. The button on the top allows for a gimmick white LED – and for turning on the power level indicator LEDs – there is NO on or off and the indicators go off on their own after a while. Lovely.

In short, there appears to be no downside to this unit – it is reasonably priced – costs WAY less than other alternatives and looks rock-solid. Reading the instructions it SPECIFICALLY states that the unit will give consistent output while charging – as well as short circuit and over-current protection. What more could you want!

Here is the Amazon UK link….  I’m sure you can find other pricing where you are.  Clearly you need to keep this thing charged with something able to provide what you need and more – but if it cannot handle peaks – matters not as this unit will take care of that. All-round winner it seems to me.

https://www.amazon.co.uk/RAVPower-16750mAh-Portable-External-Technology-y/dp/B00OJXVDAU/ref=sr_1_1?ie=UTF8&qid=1465388955&sr=8-1&keywords=ravpower

Summary:  The RavPower 16750MaH battery unit is working just fine – over a week down the line it is happily running a pair of Raspberry Pi2 devices and I’m confident that in the event of a power cut – it will keep them going for quite some time.

Meanwhile I got in touch with the American office of SunValleyTech to see if they’d like me to write about any of their other units. They were enthusiastic but (and I’m cutting a very long story short) when I supplied my details they could not ship to Spain – and hence they put me onto their Spanish department – and this morning, despite SEUR delivery leaving the thing in the village PUB !!! I opened up a sizeable box to find…

RavPower Monster unitAnd NOW, the 26 Ampere-hour Monster:  Yes, seriously, this is not a joke like the one you’ll see me refer to lower down in the blog – but a full 26,800MaH unit – a little too big for your pocket but it would sit nicely on a car dashboard, with 3 outputs EACH of which is capable of well over 2 amps output for a total output of over 5.5 amps and capable of being charged at 2 amps. A true MONSTER. This has no gimmicky light – no on/off button just a little test button to show the battery level. When it came out of the box it was fully charged. When you plug anything in – the light comes on momentarily to show you the state of charge and then goes out! Well done!

The unit comes in a pretty box with a nice, very short USB lead and another, longer one – maybe half a metre. Quite substantial.

When I started updating this review, I plugged my HTC ONE M8 into the new unit as the phone was just about flat. It is now 80% charged and a quick check of the lights indicated it has not made a DENT in the charger’s capacity.

I can see this larger unit sitting in a corner of my office powering two of my Raspberry Pis 24/7 – a shame to hide it away but this is going to give me some SERIOUS protection against power cuts! If it works out – you’re looking at 2 Raspberry Pi units lasting maybe a day depending on what they are doing – without no mains.  The smaller unit is going to sit in my car for taking to hotels etc., – I’m always running out of phone or tablet power. Just bear in mind the charging power input limit however.

Here’s the UK Amazon link. If anyone can do better  - do let us know in here.

https://www.amazon.co.uk/RAVPower-26800mAh-Portable-Charger-External/dp/B012V9H3WA/ref=sr_1_2?ie=UTF8&qid=1466008221&sr=8-2&keywords=ravpower

The Chinese alternative: After I wrote the first part  I got into a conversation in Google+ on the subject and a couple of people mentioned solar powered boards which look like the picture on the left below. I’ve covered these before and the general consensus was that they are a con to the very last one. Here’s a comment I made on Google+

solar charging - notThis is typical of the solar panel adverts - I don't know if Amazon simply don't check advert specs, don't understand them or are deliberately trying to deceive customers..but they are all at it - the specs are absolutely meant to trap people who can't do maths and it is not the worst by any means:

http://www.amazon.com/ZeroLemon-SolarJuice-20000mAh-Portable-Technology/dp/B00NIOGKL8

So look at the spec – this has a 20,000MAH rating (the unit IS big enough to handle 7 batteries – there are many ads claiming this and the units are just NOT big enough). That is 20 amps for one hour. That's at 5v so given that the 3v6 -4.2v cells need switching up at maybe 90% efficiency that requires 27AH capacity. If a Lithium battery does 3.5amps - that's 7 of them being generous. My new and large unit has 5 of them in there - I doubt very much if the others you see on the market have 7 batteries of that capacity.
So then there is efficiency in charging. Lets be generous and assume 100% efficiency. So assuming those units that claim 20,000MAH are real (they are not, they are a con generally but lets give them the benefit as I've not tested every one) - you need 27 amps for one hour, or 1 amp for 27 hours or any combination, to fully charge them.

Let's take a look at the CLAIMED power of that Amazon solar unit - 1.2w they claim. If we're needing 27AH at 5v that's 135WH, so you would need 112 hours of good, direct sunlight - even here in Spain who's sunlight DWARFS the UK, that is 11 DAYS to charge the battery. In the UK it would be more like several weeks.

You could argue my math either way but not by THAT much - it's a con.

So - take the unit I first blogged about above...16750mah - I'm going to call that 15AH being conservative... so VERY roughly we need to charge at 1.5amps at 5v for 10 hours. I CANNOT find powerful 6v solar panels - so assume an efficient switcher and a panel for 12v.... what.... 700ma at 12+v ??
http://www.ebay.co.uk/itm/10-WATT-MONOCRYSTALLINE-SOLAR-PANEL-10-WATTS-12V-PV-WITH-DIODE-/191641832274?hash=item2c9ebe4b52:g:dAkAAOSwu1VW4tgg
That's 37cm by 25cm of solar panel.

And as for the phrase “Please do not place SolarJuice on or over-heated place for continuous charging.” – HELLO!!! – don’t put a solar charger in a hot place for continuous charging. Perhaps the designers don’t get out much but usually when you have decent sunshine – it’s hot! Here in Spain I can’t touch my solar panels, they are that hot.

Facebooktwittergoogle_pluspinterestlinkedin

72 thoughts on “Ravpower the ultimate Pi UPS?

  1. I've been DOING the exactly the SAME thing... since my FLIGHTAWARE USB RX didn't need an EXTERNAL
    +12V inline pre-amp/power inserter...
    And I LOVE anything with a LED 'emergency light'...
    THANKS!!!

    1. I see that Amazon ALSO sells a 13,000mA batpak for $25...1/2 the cost of the
      27,000mA. But that that price, I think I could use TWO, cascaded with a short
      cable.

      1. I've not tested the Amazon unit mind you - you need to know if it will work without an on-button, how it will handle a flat battery and whether it will handle simultaneous charging and use. If you know better - do let us know. I'm actually quite excited about my testing tonight with the little cheap PCBs but the RavPower units are in a different league - with nice sturdy looking construction and lots of power.

      1. The one above is 15w - so my maths says that 15w/5v is 3A? The specs say it needs 1.5A minimum to charge. Obviously they could be over stating it - and even if not, it is the 'optimum' but it would be an interesting experiment. Particularly in Spain where you get lots of sun.

  2. It works with one of those phone charging sticks, powered by an old Blackberry phone charger.

    Thanks for the idea, and now a UPS for the router.

  3. It came to my attention looking at an old blog on the subject - that of course you can also detect power loss. If you are using a wired Internet connection - look for loss of internet connection (as your router would not presumably also be on battery backup). What you do then of course depends on your requirements. I've not tested this but worth a read maybe.

    http://weworkweplay.com/play/rebooting-the-raspberry-pi-when-it-loses-wireless-connection-wifi/

  4. Pete,

    I want to emphasize (or confirm) your doubt. Most of the 20.000 mAh solar battery packs are useless.

    I needed a solar panel with charger (I had a LiPo from a notebook already), so I ordered one from Ailbaba:

    New high quality 20000 mAh Waterproof powerbank,Solar Charger Power Bank portable charger External Battery for iphone/samsung
    ( I can't copy the image here).

    When I opened it it contained a 1.600 mAh battery (8% of the defined one).
    So be careful.

    George Fenyvesi

  5. Well..... I have a few of these https://www.amazon.com/Charger-Portable-Waterproof-Shockproof-Dustproof/dp/B00YOQBIKM/ref=sr_1_1?ie=UTF8&qid=1465559731&sr=8-1&keywords=dizauL+5000mAh

    I took them apart, to separate the battery and circuit board (it also says not to leave it out in the hot sun) from the solar cell. The circuit does not allow you to charge and use at the same time so I connected a voltage regulator directly to the battery. I need 3.3V so why in the world would I wan to take 3.7V up to 5V only to bring it back down to 3.3? Connected to a ESP8266 Olimex Wifi Mod, with DHT22 and DS18b20. Set to deep sleep and wake up every 10 minutes, take temp, humidity and battery voltage via ADC, send to a server and go back to sleep. It has been running since February with the solar cell in a window facing west. It does not get a lot of direct sun, but I can see a difference in the voltage. We have had a lot of cloudy/rainy weather this spring in Tuscany, so I have seen it drop some when we have had a week or more of no sun. When I put it out where it can get sun most of the day, I expect it to be fine all year. While it may not be perfect, it has worked just as well as using 4 NiMH AA batteries with a larger solar cell.

    1. Hi Freddy - well for starters if they are Lithium batteries you need to be able to supply up to 4.2v - and Lithium charging is nothing like Nicad or similar charging - overdo it and you could end up with a fire. Hence a 5v starting point is probably as good as any for the charger. Many of the units out there will not allow simultaneous charging and use - which makes them useless as UPS - and some need to be turned on with a button - which makes them nearly as useless. That's why the Ravpower is a good discovery. Right now it is sitting plugged into a 5v usb charger - and it's 2 outputs are each powering a Raspberry Pi 2 (my main one and the identical backup). All is well.

    2. Hi Freddy,

      If you want to make the ESP solution draw even less power than you're currently using then consider using the inbuilt RTC memory to store readings on wakeup and then only connect to Wifi in order to upload when the RTC memory is full.

      Andreas Speiss has a video on this subject.

      https://www.youtube.com/watch?v=r-hEOL007nw

      Andy.

  6. Yes, but the neat trick is the Dizaul solar powerbank contains all the circuitry for properly charging the Lipo fron the solar cell, I am just not using the 5v output and connecting a LM3671 3.3V regulator to the battery. I am using the included circuitry to charge the battery, with the included solar cell.
    This was just the cheapest way I could find to get a 5000mAh Lipo and charging circuit, I think it was 13 euro with free shipping from Amazon.it.
    I expect the reason most will not allow charging and use is because it takes more circuitry, which I am adding with a LM3671 3.3V. This Solar battery bank also requires you to push a button to use the USB output, again totally bypassing the 5V output by going directly to the battery.

  7. Interesting post, I recently purchased the next model up from this which is 26800mAh / 5.5A (compared to 16750mAh / 4.5A). There are 3 outputs instead of 2 and 2A input charge instead of 1.5A. No emergency light I'm afraid, but you do get a nice pouch. I've not tried it with the Pi yet. I was lucky enough to get it at £21.40, it's a bit higher than that at the moment at £34.99, but for comparison the one referenced here is £22.99 now - prices tend to fluctuate on Amazon - put it in your wishlist and watch for a price drop or daily deal 🙂

    Here's a link to the Amazon page: http://amzn.to/2318tlL

  8. Hi all.

    Thinking on the charge/discharge of this pack, are the batteries being charged individually or in parallel.

    If in parallel thus will not give balanced charging and may cause problems if you get a defective cell.

    The only problem that I see with

    1. That of course is a very good question - one would IMAGINE they have thought of this but if someone wants to rip theirs apart to find out and let us know ... that would be good 🙂

  9. Hi AndyW,
    That is very interesting. I need to see what I can do with that after I get another problem resolved. Since you can save configuration settings and boot image, I would expect persistent memory which could be used, and I think it would work really well for some applications.

  10. I'm trying to get a Pi 3 based Plex Server up and running for long car trips and I was wondering if you would recommend this as a solution, especially with a WD MyPassport Ultra plugged into the PI 3 and acting as the media library.

    I basically want something to plug the Pi3 into while driving that would give me time to initiate a shutdown (using the RemotePi Board from MSL Digital Solutions) once the car is shut off.

    Also, would you suggest using an inverter and plugging the powerbank in with a wall wart or would using a 12v charger and connecting the powerbank with a USB cable be better?

  11. Hi Peter,

    a little bit OT:

    i want to power a esp8285 with some environmental stuff on it and i'll want to use the sleep mode. so i bought the ravpower RP-PB17 with 6700mAh. unfortunatly the current consumption is so less then the powerbank goes in to sleep mode. the workarounds with a basic consumption like LED are not good for reaching my goal (keeping the ESP up for a loooooooong time). Do you have any suggestions for solving this prob.?

    sorry my english is a really mess 😉

    thx a lot - i like your site!!

    1. I would imagine these units are not designed for very small amounts of current - mind you - I've never noticed mine turn off... Contact RavPower - they're helpful... but if that is designed in - then a small load (as small as you can get away with) might be necessary.

  12. can you comment on uninterruptability of output when input disappears/reappears on the 26k8 device? I've read the 22k stops output when plugged in 🙁 seems those ravpower devices behave differently

    1. I can only say that I have two Ravpower deices - and that in both cases they continue to operate regardless. I have one I've left in Spain in charge of two Raspberry Pi 2 devices - and the power over there is awful. The other powers a single device here. Again no problems. Contact Ravpower...

  13. I just tested the 22k device (model RP-PB052) and confirm the output drops as soon as input power is restored 🙁 I guess the new generation of devices will not serve as a UPS

    1. OH DEAR ME - why do people do that... oh well, time to start looking again...

      Anyone any success out there? The supply needs to power up without a button press, stay up after power loss and while on charge and if the power runs out, when it gets sufficient charge needs to come back up again unmanned.

  14. update: I just got the 22k8 device you tested (model RP-PB41) and indeed it works as you described \o/ gonna send back the PB052

  15. I have tried the RavPower ES RP-PB19 and the RPPB41N and neither one works. My raspberry restarts whenever I unplug the charger, even when the battery is full.
    Any suggestions?

    Thanks in advance.

    1. I can only speak for the models I have tested. Contact RavPower - they'll most likely be able to offer help - but as the units are not intended for this purpose you should only try models specifically suggested.

  16. Wish I'd looked-for and found this page before. I'd bought 3 of the Ravpowers by July '16 plus the rugged ... all 4 with the great UPS friendly characteristics. Plus not noticed any diminished capacity after basic manual testing after months of UPS duty.

    Bought 3 from Amazon UK while cheap (£17) in December and another model with 3 outputs ... came to use them in earnest this week but only today realised issues I've been having are down to this new batch. I was blind to their new characteristics when troubleshooting - thinking myself clumsy or other issues, because I'd so thoroughly tested the older designs. Now I have 4 power banks I don't really need! No packaging - thrown out because of erstwhile earned confidence. Can't really send them back - it was never promised that they behave as UPS's etc. Damn.

    Great page ... will have to explore alternatives mentioned. I thought I'd found the end of the rainbow when I bought my first batch!

  17. Addendum to my previous hasty comment. On closer (still hasty) inspection, all my 16750 Ravpower models, new and old - that I can access to see, say Model: RP-PB19. So these are the ones that behave as I'd like (UPS), and ones that have given me no-end of trouble while I remained ignorant of their short-comings.

    Initial further-investigation with a little more patience and an ancient Velleman oscilloscope (DC / 2V/div & 2sec/div) is providing new information ... on my new 16750 Ravpower units, the 2.1A port is still showing UPS-like behaviour. I'll have to test for longer periods and with a more structured approach to be sure, but that seems to be the case. The 2.4A port drops upon power-insert to the powerbank; the 2.1A port doesn't.

    I had to use a hub for this quick test, so I could plug in a USB male end with trailing stripped wires for the oscilloscope probes, along with different loading on the other ports ... including using a charger-doctor for approx. current on that port ... so using loads of a 3G dongle (sms mode only) ... 0 to 30mA non-constant approx. ... Ravpower port stayed powered and no blips ... partial charge to Android phone ... approx. 400 to 500mA ... no blips, and a 5 Ohm resistor (can pull 1A through this particular hub) ... no blips. (No blips while pulling or inserting the power cable to the Ravpower, or adjacent loads).

    I tested 2 of my 3 new ones that I could get at, and 2 of my 3 older ones. The new 3-port RP-PB052 I have drops on every port.

    Anyway ... I might not yet be able to confidently run my Bay Trail compute sticks on the 2.1A port ... will have to test higher loading and longer durations and more power-delta scenarios ... but I'm hopeful that Pi 2's and maybe Pi 3's might be ok on the 2.1A port acting as a UPS. I could always throw in a few dirt-cheap 1F super-capacitors to be on the safe side. 🙂

    Oh! Nearly forgot! Yeah ... with these tests i.e. 1A load very approx. plus something for the hub, I was seeing around 1.15 to 1.25+ A going into the power bank (portabow DC power monitor). Don't know if it will scale yet with higher load ... didn't notice a difference with the phone plugged in but couldn't put that through the charger doctor too ... so these results all a bit fuzzy.

    Will investigate more when I get a chance. But phew ... not quite all wasted money on batteries.

  18. One more thing (Columbo - style). Not done much more testing yet - had to take a break from it. But ... doh! Yeah - push button activation. 🙁

    Hadn't noticed because inserting a load directly/immediately into the 2.1A port activates that port. (Not just a load-detection ... maybe it's mechanical or capacitive detection or something). You can also activate the port with a DEPRESS of the unit's button.

    With more accurate measurement, my 3G dongle draws fairly continuous 50mA (no data connection) (I said 0 to 30mA before but I guess it was too low for my charger doctor to get a handle on). By chance, the charger doctor draws around 30mA. I found that 30mA wasn't enough to keep the 2.1A port alive ... goes down after about 1-minute. 50mA is sufficient on the unit I tested (but testing only for fairly limited amounts of time).

    I also found that after the unit is activated with direct port insert or button DEPRESS, that a further press and continual hold on the button, keeps the 2.1A port alive (at least for the 10-min or so I had patience for) rather than the 1-min timeout, with no load at all.

    But ... even though I could keep the port alive so long as the unit's alive even after the load switches off, it's looking like the unit might not behave in the way I want if there's a black-out and the batteries go low-enough for the unit to switch off ... the question is will it bring-up the 2.1A port again upon power-resumption. I don't think it will. ;(

    So ... if I want to use it as a UPS I'll have to hack the way the port activates somehow if it's imperative that port comes alive again after extended black-out scenarios such that the battery gets drained ... otherwise it looks like it might be ok for Pi's ... but ... and there's another but ... I'm still not sure how the power-input to the unit now matches the power output. Certainly using both ports on the unit affects things in ways I'm not yet clear on ... and I'm not sure that pass-through is behaving in a way that will play nice with the battery charging management so as to keep the cells in relatively good condition for extended periods. 🙁

    So I'm thinking at the moment, the "UPS" like behaviour of these later-units albeit sporting the same model-number-badge might be strictly limited to non-critical (as in ok if it doesn't recover from extended black-out unattended) and fairly low-average-draw situations (but ok upto 2A ish peak) where I can use the single 2.1A port solo. That's my gut/perception at the moment.

    1. Hi David,

      If I'm following you correctly here, the 2.1A port on the new revision of this RP-PB19 powerbank will work as a passthrough UPS (by which I mean the powerbank is plugged into AC power, the RPi3 plugged into the powerbank, if the power goes out it the RPI3 would switch over to battery power from the powerbank without shutting off until the power comes back on or the battery is exhausted) for an RPi3 in a situation where the power goes out and the powerbank battery has enough juice to outlast the length of the blackout, but if it's a long enough blackout that the powerbank is exhausted, when the power comes back on it would need my intervention to get the 2.1A port working again?

      I could deal with that limitation, but was just looking for confirmation before I ordered one of these. Thanks for testing this out!

      1. Ha - I got so confused I can only say "I think so". 🙂

        I have started a sort of blog / website (wordpress) but it's in a right mess still at the moment with lots of half-started stuff and scaffolding - nothing to be proud of yet. Bit embarrassing. And I did start augmenting the newer battery pack externally with a solenoid to push the button but not documented it yet or added the crucial arduino. I'm doing/learning so many things at once I'm really suffering from context thrashing. But screw-it ... I'll link to a picture anyway:
        https://blog.xarta.co.uk/2017/02/raspberry-pi-2-vpn-and-battery-ups/

        Err - augmented battery picture a bit lower down.

  19. Ok - one^2 more thing ... with initial testing (holding with fingers) a small 12V 300mA pull-type solenoid thingy seems to have just enough punch (impulse with given shaft mass and full extension) off a 9V battery to depress the button on the unit - and bearing in mind that when power is restored to the Ravpower 16750 RP-PB19 - either old or new ("old" as in June/July-maybe 2016 vs "new" as in November/December 2016+) the 2.4A port always comes up (if off before) regardless of load, until time-out period passes (unless approx. 50mA+ load present) ... then I'm thinking - have that output charge a 12V capacitor or 16.5V-ish with limited charge (fancy-step-up ready-built high quality things only cost a few £ now rather than building something yourself - I use them for 9V network switches from 5V supplies) (and 5.5V 1F supercapacitors only cost £1-odd as well!) (but ok: I haven't thought about the maths - just throwing possibilities there) ... and without logic necessary per se you could have some threshold switch the capacitor to the solenoid - various methods. So in effect you could daisy chain the ports going-up with no complicated mucking about or hacking open the Power bank or anything - to arrive at similar behaviour to earlier builds in that respect. (Desirable to get the 2.1A port up again as it doesn't drop with Powerbank input-power inserts/withdrawals so long as load is maintained and juice left in the tank).

    When I get something reliable working I'll a have to do a video & link to it - sorry about getting carried away with a deluge of experience-related info. here. Just thought it might be interesting/relevant ... but I know I can be a bit overwhelming sometimes! I want to get back to Docker on a Bay Trail compute stick (MQTT image) and eventually back to blasted programming so I can maybe finally get coding job ... but things keep getting in the way! Nothing works!

    1. Hi Phil

      So two things...

      1. Please DO let us know how you get on - at 1 amp these are not sufficient for the likes of PIs but certainly worth further investigation for use with ESPs etc.. Especially ESP32 as this is powerful enough to consider using as a central controller.

      2. Please DO test as this looks to have similar components to the ones I have though a different layout.. Test it with a FLAT battery, applying the LOAD first - which should do nothing - THEN apply power... if it comes up after a short while - that is GREAT - but the ones I have do not. If you disconnect the load just for an instant - all is well, but of course that is not practical.

      I'm also looking at an idea I discussed with Antonio, to take a £2.50 Arduino NANO and a 50p relay... the NANO uses very little current and so SHOULD not present an issue for the power supply. The idea would be simple - monitor the voltage - if it FALLS below a certain level, warn the Pi (digital signal on an input) then 30 seconds later turn off the power. Then wait until the battery voltage is much higher before turning the Pi ON. Voltages to be determined. Assuming it works I'll put the (simple) code here. Clearly a flashing light would be used to indicate the state of affairs. the NANO would be running on the 5v output to avoid any issues with the relay and would measure the battery voltage directly. Hopefully by stopping the Pi before the battery voltage gets too low, then powering the NANO for a day or two on what's left of the battery should not be an issue. All needs testing.

      1. I will do some testing when they arrive, did have one but the plug I used for the battery had the polarity the wrong way round and I blew up the chip...... This should be fine running the pi zero W as when I ran it off my supply it used less than 200mA.

        1. The ONLY thing about the MOSFET is voltage drop... you know the issue with the slightest impedance in the power supply that we've discussed elsewhere - and in the case of this MOSFET you would be raising the ground line slightly. Good idea but I'd want to see that tested in practice.

          Meanwhile I'm doing a little work on the NANO idea as against the ATTINY because firstly the cost difference is negligible (and I happen to have a NANO handy) and secondly it has a proper UART. The plan would be to offer two methods of detecting low battery for the PI - firstly a GPIO bit going high - handy for the simplest setup - so when you see that bit high, you instigate a shutdown. In addition however I will output a JSON package including the state of the battery and any other important info like pending shutdown. I woke up in the early hours thinking about this and after the market today I should have working code.

          1. or, when you have all ready, switch to an arduino mini which is available in 5v and 3v3 versions, so no need for level shifting the serial with the PI, which is 3v3...
            https://it.aliexpress.com/item/New-Pro-Mini-atmega328-3-3V-8M-Replace-ATmega128-For-Arduino-Compatible-Nano/32292926009.html

            oh, look at that Pimoroni board i published here yesterday... look at how it uses 2 pins, 1 for detecting and 1 to complete power down, here an excerpt:

            "We've put together a one-line-installer to install the clean shutdown daemon. It watches the state of BCM pin 17 and, when pulled low (pressed), it initiates a clean shutdown. Last thing, just before your Pi shuts down, BCM pin 4 is pulled low to completely cut power to your Pi."

            https://shop.pimoroni.com/collections/new-products/products/onoff-shim

      1. I thought maybe it's a fairly cheap implementation where the focus is on the 12V inverter, and the 5V is just an off the shelf component add on or even some kind of tap or something. The main purpose of this device I think is for network gear and they often state a 12V 2A supply requirement.

        I assume, that if reasonably efficient, this device might be about 25Wh ... about half the RavPower. Looks interesting though. Oh ... for the Pi ... I've run a Pi B for months non-stop as a regular-use OpenVpn server for streaming for friends etc. and I've never had an issue with that on the 1A supply I feed it. Pi 3's might need a tad more spiky oomph occasionally.

        Looks like this battery is a proper "UPS" ... i.e. doesn't have the cells on a float charge and switches the output between battery power/AC etc. ... so as long as the quality is reasonable then shouldn't be a risk I think.

        I'm interested in them for my pfSense box, modem, and a smart switch - all running off a normal UPS at the moment which has uses 'cause of the UPS packages in pfSense ... but eventually I'll have a Pi zero managing all my power so these might be handy. I've seen the plug-in types before but I prefer this form factor and the 1A 5V supply would have uses. Plenty of cheap/efficient 12V to 5V USB converters too for the higher power 5V -
        devices - just at the expense of over-all efficiency.

  20. QUESTION: I love USB for power supply / distribution but it would be really handy to have a USB A male to USB A female slim inline form factor adaptor to insert into power-only connections or maybe with the option for data too, with either a reed switch or relay or opto-isolated relay or solid state relay or power mosfet ... something ... just for simple power switching for USB devices. Cheap and simple / no-fuss / no extra cables / no boxes etc. I suppose a voltage sensor and even current sensor possibilities would be nice ... so I suppose an I2C version would be handy ... but at its simplest ... a simple low power normally-on relay would do for me - I could just throw in an extra wire for voltage feedback to the controller.

    I've got an inline device for 12V switching using an ESP thingy ... but I don't want wireless ... I just want "simple" so I can hardwire it to ... a Pi zero say.

    I've searched and searched and can't find anything. I can make my own of course with male A plugs/sockets but it's all fiddly soldering and life's too short.

  21. Ahhh - they're cool ... but you misunderstand me. I mean switches to switch a number of USB devices and off, all controlled from one Pi zero.

    Without having to "hack" stuff ... i.e. so I can plug any old USB cable into a power source on one end, and any old USB cable into a device on the other end, and just switch the power on it on/off. At the most basic level.

    You can get boards that are a USB hub basically where you can programmatically turn each socket on and off but they cost a bit of money and then you're stuck with making cables reach that board/module etc. ... it's all constraints.

    I just want simple. USB male A to USB female A ... no more than a couple of centimetres in between. So I can plug any lead I like in there - any length to suit the situation. And then have a control lead of some kind depending on the actuator and feedback channel. Electrically isolated somewhat at least from the power switching.

    That way I can scale it however I like. If it's simple sense on/off and I need more & more, I could stick them all into some kind of I2C or addressable thingy first etc. ... just flexibility.

    So all I want is something like that wee gadget you linked to, but soldered on one end the USB A Male socket, the other end a USB A Female socket, and maybe a gap for each power trace with a via, so I can link them or stick a switch in. Something as simple and basic as that. I can solder it myself but doing 10 or so of them, especially on vero board etc. is just a pain ... I'd much rather just buy something. I did more than enough laborious soldering as a kid ... and I have too many things to do.

    🙂

    But ta for the link.

    (I could just carefully strip a cable and snip one of the power cables and do that way, but then I'm stuck with a cable that length doing that job ... but I'm thinking now that might be my easiest option. I could get a tiny box to enclose where I interrupt the power-path and whatever actuator I use. If I do it carefully enough shouldn't be too ugly).

  22. Actually - slight modification (been battling a simple email notification issue that's wearing my patience lol ... didn't give my descriptions here enough thought) ... yeah ... I think I wanted slim form factor so I could actually plug the USB Male A straight into a power source.

    But thinking about it a bit more, perhaps I'd quite like a USB micro female to USB A female. That'd probably be more useful ... more of a common-use need. I actually have plastic moulded adaptors in this configuration ... but without any easy way to modify them for my switching needs.

  23. Actually - although I wasn't thinking along those lines because I wanted flexibility for re-purposing and quick redeployment for multiple situations ... and of course I'd have to be careful about possible ground loops if using a common ground for a lot of devices also interconnected with network cables etc. ... but yeah - actually ... that's a great cost-effective idea! I could use one of the ports for some kind of arduino or equivalent with built in USB driver - quick google - something like this maybe: http://www.dx.com/p/tiny-usb-atmega32u4-development-board-compatible-arduino-black-379264#.WQo9jfkrJEY ... and might be able to squeeze extra stuff into the voids left from the vacated rockers. Might already be python examples or something for talking to the Arduino. Sounds very doable ... quite convenient.

    Ta for the tip!

  24. Oh - but I forgot ... I wanted to switch things on and off after "UPS" 5V batteries. Been really busy / tired ... not thinking straight lol. I have 3 Ravpowers and 1 other battery pack all actively playing UPS duty at the moment, and the one I'm "modifying" (when I get back to it after 1000,000's of other jobs including my big network refit ... big to me anyway - the amount of configuration takes enough time).

    So anyway ... with typically two USB ports per battery pack, that's a lot of outputs up for potential switching (bom bom). Mostly (TODO) for given-up-on-mains-returning states, or for last resort off/on power-cycling ... using (probably) a Pi zero for watchdog duty and general management.

    Still might look at using the cheap hub but it still wouldn't easily meet all my needs.

    Some of these plans I formulated months ago ... queued up with zillions of others on the back-burner ... takes a while for it all to trickle back into my working-memory. 🙂 Almost like rediscovery. Can't even jot things down on my website easily right now ... finishing a plug-in I was working on (WordPress plug-in ... need to change over to OOP and make sure the way I was manipulating the loop thingy in it isn't what's having side-effects on my admin side) ... and then clear all that up too. And then I need to get back to .Net stuff.

    Too many Things ToDo. Oh ... and my credit cards have run-out and I really need to get a job again. Haven't done everything I meant to yet ... probably won't be a crappy office admin job again that I get stuck-with ... hopefully I can get some kind of tech job but I don't really fit expected criteria (i.e. I like technology and IT etc. etc. ... that doesn't seem to sit well with people ... recruiters etc. look for people who act for money & status & for their own interests rather than work satisfaction or in the interests of their employers ... someone like me is too alien to them conceptually - especially as I like so many things but also give all my time & loyalty to my employers when in employment ... I don't compute in their small minds unfortunately). But we'll see.

    I'm waffling. I always waffle when I'm exhausted at nearly 2am in the morning with tonnes of stuff still to complete. Using PRTG on my reverse proxy / MX2 Compute Stick ... takes a bit of configuration but really cool. And free for 100 sensors or less ... and can do SMS / Push notifications too ... to the companion app. Just wanna finish getting email round trip monitoring / alarm set-up finished and then I must sleep! Finally fixed my 3 byte vs 4 byte utf8 issue though ... been meaning to do that for years.

    1. As far as I can see Antonio (I'm back from the market) that is simply a power supply with power off sensing - nothing to do with battery backup as such. I may be missing something.

      1. you're right, but the shutdown part and how to really do it is interesting... my rpi3, when i halt it, continues to blink with some leds, so maybe doing that way it really is powered off... can't test now, my pi is far away from me...

Leave a Reply

Your email address will not be published. Required fields are marked *