Hackaday Prize Entry: Laptop Batteries For A Power Bank

USB power banks – huge batteries that will recharge your phone or tablet – are ubiquitous these days. You can buy them at a gas station or from your favorite online retailer in any capacity you would ever want. Most of these power banks have a tremendous shortcoming; they need to charge over USB. With a 10,000 mAh battery, that’s going to take a while.

We already have batteries with huge capacities, are able to charge quickly, and judging from a few eBay auctions, can be picked up for a song. [Kumar] is working on a device that leverages these batteries – and the electronics inside of them – to build a smarter power bank.

Right now, [Kumar] is working with Dell Latitude D5xx/D6xx replacement batteries that he can pick up easily. These batteries have an SMBus interface, and with a low power ARM microcontroller and a TI BQ24725a, he has everything he needs to efficiently and safely charge these batteries.

[Kumar] says he’s looking for some community suggestions and feature requests for his project. If you have any, be sure to drop them over on his project page.

The 2015 Hackaday Prize is sponsored by:

48 thoughts on “Hackaday Prize Entry: Laptop Batteries For A Power Bank

  1. I might be wrong, or just plain stupid, but to me this project doesn’t make sense. Let me clarify, this project makes sense but using laptop batteries doesn’t make sense.

    Why bother reverse engineering a laptop battery? First, you have already paid for R&D+Profit+whatever markup to design and manufacturer, said battery. Now you are spending further time reverse engineering said design and adapting/compromising your needs to what battery design can/can’t do.

    Why not go with, RC batteries, variety of sizes/capacities/cell combinations/voltage/etc. Plus the fact being bare cells, much much cheaper. Another route is prismatic cells.

    Now for controller, sure one can spend some time designing and making a controller, but Minibox OpenUPS (I have no connection to Minibox, never even bought anything manufactured by them) is off the shelf solution available.
    – 1s/2s/3s/4s/5s/6s battery
    – Li-ion, Li-po, LIFEPO4, Lead Acid, etc
    – Charging voltage is 6-30V/3A
    – Input is 6-30V/6A (10A peak)
    – Output is 6-24V/6A (10A peak)
    – USB interface

    So what was the point of this project again? So once again, this project doesn’t make sense to me, other than excerising reverse engineering, or if you have a pile of laptop batteries (for same model).

    I apologise in advance, please do correct me/show me the benefits.

    1. The only part that might make some sense is that the laptop manufacturers have much higher volume and they can negotiate a much better pricing than the regular Joe or even smaller OEM manufacturers. So in the end if you pay yourself nothing, you might ended up being able to get the cells cheaper assuming you can buy them in bulk. Now the problem is that you are actually paying the retail price to get those batteries. So you are not much worse off buying the raw cell.

      The problem of that is now your product (if that’s what the intent is) will be rely on someone else’s product cycle. You waste a big chunk of that when you have to reverse engineer their product and use their part. If they decided to change the form factor while your product is selling like hot cakes, you are SOL.

    2. We’re talking about availability and price,especially in the secondhand market.
      They can often be found for free if you’re paying attention when someone throws their old devices out.
      Additionally, there’s usually some very sophisticated battery health monitoring and diagnostic capability built into these smart batteries.
      Dealing with bare cells yourself is its own special kind of hell.

    3. If it is a one shot project, sure whatever. As for dumpster diving, you have no idea what you are getting.
      If the intention is to make a product or someone to duplicate the results, then you want to have some quality control of the batteries you are buying. 2nd hand or after market sources can differ greatly from batches even if you were to buy from the same ebay guy/shop. OEM batteries at least have some level of assurances.

    4. I have quite a few issues as well.
      1. From the starting description of the project the problem is long charge time. This is limited by usb because the idea is to charge it with the same thing you charge your phone with. So you could just use a more powerful charger to fill it up faster.
      2. I really have a hard time seeing somebody throw away a laptop that is not old enough for the cells to be pretty dead.
      3. I can buy bare 18650 cells from good distributors for less money than them in a laptop battery.
      4. I have hard time justifying investigating some obscure protocol the battery uses.

      1. [#1] Understood, though I also make the assumption that there might also be a laptop charger (60-90W or something) lurking around, so why not use it to charge faster? I’m not trying to make someone buy yet another power supply. If you think you are OK with leaving your existing power bank charge overnight, use it. But if you need it urgently and you realize the pack is flat, good luck charging it.

        [#4] Also these batteries typically share the Smart Battery specification, which is well available if you search for it. So it’s not some obscure protocol. As an example I hooked up my battery’s SMBus lines to my BeagleBone Black, ran “smartbat” [http://www.linurs.org/battery.html] and I had all battery parameters on the screen right away. So as long as the battery you use is “Smart”, this should work.

        [3074ap] What I view this pack as it should be modular, in the sense that it offers multiple outputs and each of them could be individually controllable. You can see that in the block diagram, also mentioning that it will have a small display to report core parameters. Also the design of OpenUPS is closed [I see no schematics, no firmware sources, no board layout], if you needed to use something like that and don’t care about how it’s done, use it by all means. If you want, you can also think of it as an open-source alternative.

        Start building something directly from the cells would typically end up having to add gas gauge, protection circuits and allied functionality which is built-in to these battery packs and available at your disposal with two wires. But we have to start somewhere, so let’s just start with a laptop pack and see where it can be taken.

        1. [Abhishek] But that is my point, RC battery packs are modular, Power connector and balance connector come standard on pretty much every pack. Of course prismatic/cylindrical cells and diy packs are different story.

          With OpenUPS, its all configurable and if its being used with a computer, it behaves like a battery with capacity gauge.

          And if we want to go open source, why not go with RC battery packs which are much simpler/cheaper/readily available/longer life cycles. Plus the fact that we will not need to reverse engineer and mold our design to what OEM battery pack was designed to do.

    1. What you want for portable welding is a little oxyacetylene set. Significantly more heat and for far longer than you could accomplish with similar size / weight batteries, and as a bonus you’ll be able to cut, solder, braze, heat things, and also weld many metals that MIG is not typically capable of.

    2. Being that any decent welding needs a couple of kW, you’d either have very limited welding time (a few minutes at most) or a very heavy and expensive device…
      Battery powered spot welder might actually be practical though.

  2. Just grab few 18650 Li-ion batteries, connect them in parallel, add charge controller based on MCP73841 to get higher charge currents and a boost converter with USB connector. Connecting batteries in parallel removes necessity for balancing circuitry. Adding undervoltage and overcurrent protection, or better yet, designing complete charging/boost controller with build-in battery protection would be even better. I would certainly buy that, because I have over 150 LI-ion batteries from old laptops.

    1. The thing about parallel cells is that they each get a different current through them due to differences in intermal resistance, so while you’re discharging them one cell empties before the others, and vice versa.

      They do automatically balance, but only at rest. That’s why you still need to pick cells with matching properties, or a balancing circuit that makes sure you don’t destroy your cells.

      1. Also: the internal resistance of a cell decreases with temperature, so the cell being drained the most will heat up and get more current. Hence why parallel packs can be problematic.

        1. I don’t think it will matter that much in real life. Most ready to use boost converters for power banks have current limit of 1A. Divide that by 5-10 cells in example bank and current per cell will be limited to about 100-200mA. Not enough to cause much heating with internal resistance of cell. Let’s assume worst case scenario, old cell with internal resistance of 500miliOhm in power bank of 5 cells. With current of 200mA gives 20mW to dissipate. At the same time there will be voltage drop of 100mV, which will be enough to increase current flow from parallel cells to balance voltages. Best case scenario: internal resistance 75miliOhm, with the same current will have to dissipate whooping 3mW. Voltage drop across cell will be about 15mV.

          1. You are not considering the uneven current draw of unbalanced batteries. Even if you have 10 batteries, if one of them has a significantly lower internal resistance, it will end up delivering most of the current. So instead of only getting 1A / 10 = 100 mA, that battery may end up delivering much more of the current. If the difference in batteries is enough, that one battery may end up delivering almost the full 1A. 1A * 500 mOhm = 500 mW. That is certainly enough power to generate heat in the battery.

            The other thing you are failing to consider, is that the mismatch in battery may result in A) more current being drawn from the battery than it is rated for or B) drawing current from the circuit even once the battery has fallen below its minimum voltage. Either of these could lead to a battery failure. Even if the failure of that battery is not catastrophic, it could lead to subsequent failures of the other batteries due to increased loading.

            There is a lot more to consider with batteries than you might think.

          2. @phreaknik, you are almost right. The word “almost” is the key. Li-ion cells increase their internal resistance with use and abuse, not decrease. New ones have internal resistance of about 75-150 miloOhms. Dead and dying ones have above 500 miliOhms. Even if one of cells has significantly lower internal resistance, then after some time it will increase to match with other cells. And when it matches, the rate at which all cells will degrade with increasing internal resistance will be the same for all of them.
            Also you need to consider this: each 18650 cell can deliver 1-3C of current constantly, and a lot more when shorted, 200mA per cell is 0,1C or less. Any imbalance between cells caused by uneven internal resistance can’t even happen. Even with electrolytic capacitor banks it can’t happen. Go on, test it. Make five banks of electrolytic capacitors, ten or more in each, values don’t really matter as long as ESR will be lower than 0,1Ohm. To one of them connect 0,47Ohm resistor, to two 0,1Ohm resistors, and to the rest of them 0,22Ohm resistors. Then connect them in parallel to simulate typical power bank with different internal resistances. And then charge and discharge it with constant current of 1A and see, what happens. Look at currents that flow trough each resistor.

          3. “Even if one of cells has significantly lower internal resistance, then after some time it will increase to match with other cells.”

            Yes. By wearing out. Then you have one middle-aged cell among many barely used cells and the pack fails because the one cell reaches the end of its cycles faster than the others.

            That’s why you have to match them.

          4. When you buy new cells, they will be probably close-matched. Especially because internal resistance in case of Li-ion batteries depends on shape and size of electrodes. I have over 150 slightly used ones, all of them holding charge pretty well, and I don’t really care, if one of them dies faster than others. I can just replace them. And because power banks don’t demand high currents from cells, it won’t overheat them. If one is really picky and worries about unbalancing cells during operation of such bank, there is very simple solution. Just add a 4,7-10Ohm resistor in series with every cell and Bob’s your uncle. With 5 new cells and 4,7Ohm resistors the total equivalent resistance will be 0,96Ohm. If one of them is close to death, it will have resistance of 0,5Ohm, and total resistance will increase to 0,97846Ohms. And because of its bad condition it won’t be able to deliver enough current across that resistor to be discharged faster than other cells. This will also protect cells during charging.

          5. “When you buy new cells, they will be probably close-matched.”

            That depends on where you buy them, and what you buy. You can easily get cells from different batches.

            Things like differences in separator or electrode thickness causes variation in the cells because they’re basically rolled from sheet. If the sheet thickness changes, the cell properties will differ.

            If you’re buying prismatic cells for cellphones or similiar devices that operate on single cells, there may not be any attempt to make them uniform in quality since they’re not meant to be used in packs.

      2. Parallel cells are so much better and easier than serial cells. They effectivelly behave as a larger cell.
        It is true, that when connecting them you should make sure to start with identically charged batteries, preferably as discharged as possible.
        But over time, the load balances by itself because the internal resistance increases as batteries discharge. It’s a self regulating process.

      1. There are none. You must make your own circuit. Check out Microchips’ site for datasheet. It uses external P-MOSFET and current sense resistor to control charging process. You can make with it even 20A charger, but going higher than 5A for even big power bank won’t be necessary. Li-ion batteries shouldn’t be charged with currents higher than 1C, and lower are better to keep cells from overheating. For example I have a block of 8 cells with capacitance of almost 19Ah, With standard charger it takes 19 hours to charge. With charger made for RC batteries with current set at 5A it takes only 4 hours.

        You can get on eBay cheap Li-ion battery charger based on TP4056, but it’s limited to 1A. It costs about two bucks per board. For charging 2-4 cells it’s enough.

        1. There are quite a few 18650 enclosures available on the market. A guy did detailed examination of few too.

          Basically his conclusion was, using them is highly risky, discharging 18650 cells was no problem but charging was shonky at best with ripples.

          I am personally using XiaoMi batteries and they have been great. Decent price but proper charging with temp management. Here’s review of one http://goughlui.com/2014/10/15/review-teardown-xiaomi-mi-ndy-02-ad-gold-10400mah-power-bank/ Unfortunately there are more fakes than real.

          If you go through http://goughlui.com/, he explains it all in detail.

  3. THis post makes me think of all the breakthrough batteries that will allow my phone to charge in seconds.
    Yeah, right now it needs 90 minutes with 10W to charge, charging in 30 seconds would require 1,8KW of power, good luck making me carry that charger around.

    I have a couple of 10.000mAh power banks that I only use in long trips, so I don’t see the problem of them requiring an overnight charge.

    1. Back in the day, before batteries were not removable, or hidden behind the case, I had a Motorola MR20.
      I had spare batteries for it and a desktop charger which did the phone and the battery.
      When I had my 6130, again, removable battery and desk charger with spares on standby.

      I think we all went wrong somewhere with the need (??) for slimmer phones.
      We’re still carrying around spare batteries.
      But now we cannot easily replace the phone battery like it’s a AA cell, we we’ve taken an odd step of having a portable power pack.
      Which we charge and discharge so effectively takes longer than it would to charge and insert a battery.

      This process is illogical.

      1. Aw man, you deserve internet points, this whole scenario of non removable batteries is really illogical, but when consumers demands super thing electronics, every single 0.1mm saved makes a difference i guess.

        I used to get Samsung Galaxy phones just because of removable batteries and the fact that Zerolemon batteries were available (Zerolemon has genuine/decent batteries and not fake ebay crap). Since S6 that plan went down the drain.

        I mean if I am going to carry an external powerbank, then why not have a bit thicker phone? In the end it will still be lighter than both combined.

        1. The point is to force you, and everyone else to replace their phones rather than replacing batteries. With sandwich-sized phone (to accommodate bigger screen) it doesn’t really matter, how thin it is. But if you can’t replace battery when original one dies, you will buy new phone. It’s a win for corporation.

        2. Yeah, same reason for me. After about 2-3 years i give the phone to somebody with a new battery, which makes is operational for about 2-3 years again.
          Problem is most people are stupid and don’t do that, so it makes sense for the manufacturer not to care about the few who want removable battery.

  4. I suggest NOT going LiION and use LiFePo4 instead. a lot easier to charge, higher capacities, and significantly more charge/discharge cycles along with a far far lower self discharge rate.

    1. LiFePO4 is also much safer — it doesn’t have the “catch fire” failure mode of Li-ion.
      Same charging strategy though: CC/CV — no difference other than terminal voltage.
      Also, 4 cells in series is just about exactly the same as a 12V car supply: perfect for using 12V appliances.

      I use a Bioenno 12V 15Ah battery: http://www.bioennopower.com/collections/12v-series-lifepo4-batteries/products/12v-15ah-lfp-battery-blue-w-pvc-pack It can put out >300 watts for a half hour, can drive inverters, can power my (50 Watt) ham radio directly, charges in <4 hours, easily fits in a backpack. Lighter/smaller/cheaper sizes are available too, for those who don't need the grunt of this one.

      USB power out is trivial to get via 12V converters. Lighter plug is a lowest-common-denominator output, but a much, much better 12V power distribution solution is Anderson Powerpole connectors. The battery comes with them.

        1. PCM board doesn’t seem to care. It works fine in float service: I often use it while charging. Charge input is (internally) the same terminals as power out, but uses a separate dedicated barrel connector. This (unfortunately) means that charger input works fine as a power output, and it’s not fused properly for that (the charger leads are much smaller than the discharge leads). I solved that safety risk by simply cutting the charge connector off, and charging through the (output) Powerpole connector.

          The charger is OK. It has the usual safety certifications, has a red/green charge status LED, doesn’t get very warm even at full charge rate. Build quality is not top-notch but it’s fine.

  5. I don’t know what happened to my earlier comment, didn’t got moderated or something. But I needed a battery bank for a project (Security system for a yacht with solar cells) and after researching I came up with OpenUPS. In the end, project got scrapped and i didn’t go ahead.

    What I found good was, OpenUPS gives choice of 1s/2s/3s/4s/5s/6s battery, Li-ion, Li-po, LIFEPO4, Lead Acid, etc, Charging voltage 6-30V/3A, Input 6-30V/6A (10A peak), Output 6-24V/6A (10A peak), USB interface and pretty much every single factor used configurable. Biggest thins was, I had the choice to use cheap RC battery packs, or make my own pack using prismatic cells/18650’s/whatever tickled my fancy and it still balance charges up to 6s packs. But choice was still there, to use pre-made battery packs or making my own.

    PS:- I am not affiliated with minibox in any way and so far haven’t bought anything from them.

  6. Charging over USB could be a LOT quicker. According to USB.org’s Power Delivery page, there are existing standards that allow power delivery up to 100 W.

    At 5V output, that would mean that the connectors and wiring will have to carry twenty amperes. I don’t know that I would want to run that much current through a USB connector, much less the wire in any of the USB cables I have ever seen.

    My phone has a 1785 mAh (6.8 Wh) battery. If I am figuring correctly, it should take just over 4 minutes to completely charge the battery (I know you couldn’t charge a real battery that fast, but this is theoretical). I would imagine any gas station cable would burn through completely in that time, potentially igniting anything around it.

    The third paragraph is here so several people after me do not have to type it.

    1. It will take longer than the 4 minutes to full charge the battery as you are charging at full current (assume that your battery can take it) during the CC stage, but not in the CV stage where the current is reduce to limit the battery voltage.

  7. Am I the only person who got points docked from his lab writeup for not simplifying my units????

    It’s not 10,000 milli-Amp-Hours; it’s 10 Amp-Hours!

    Marketing seems to be slowly eroding everyone’s ability to think intelligently. It’s rather irritating (to me anyway) and not simplifying the units appears redundant. If we’re to keep going down this road, why not call it 10 million micro-Amp-hours? The number’s bigger, so it must be better!

    1. Good question.
      It’s rated in thousands of milliamp hours rather than amp hours for the same reason 100nF capacitors are specified as 0.1uF, 1nF capacitors as 1000pF, and 1mF capacitors are labeled as 1000uF.

      It’s because there is more value in being consistent with your notation than elegant. (Also an interesting juxtaposition of engineering vs scientific ways of thinking.)
      A good student follows their instructor’s guidelines, of course. :-p

    2. It’s also a really stupid way to specify the capacity of the pack. It’s useful only if you also know the voltage.

      Example: I have a car “starter” pack like this one: http://www.amazon.com/PowerAll-PBJS12000R-Rosso-Portable-Starter/dp/B00D42AFS8 It outputs nominally 12V and is rated at 12000 mAh, implying a total energy storage of 144 watt hours… But No, it has three LiPo cells in series, producing only 11 Volts (on a good day), and the pack capacity is actually the SUM of the mAh capacity of those three cells, despite them being in series, not parallel. The total pack capacity is actually just 44WH (nominal). The actual usable capacity is actually much less than even this, as the voltage drops below 11V quickly.

      Your typical “power” (i.e., *energy*) bank with cells in parallel can in fairness sum the capacities, but they use a boost converter to make 5V out of the 3.6V cells, and not at 100% efficiency, so a “10000mAh” pack you might hope can charge your 4000 mAh tablet cell at least twice. In practice, you’ll get just 30 Wh out of that pack, or 6000 mAh at 5V. You might get a single charge in your tablet from that.

      Lies, damned lies, and marketing.

      1. You know Paul, I’ve had the exact same thought many times.

        What’s dumb – and you’re totally right about this IMHO – makers of battery packs just love telling you how much juice is packed in their battery *cells* and neglect to mention the efficiency of their regulators or the percent of that battery capacity that goes unused.
        They can’t predict how much energy will lost in the charging cable and in the charging circuit of your battery-powered device. Trying to calculate effective power delivered without being able to analyze the entire system is kind of pointless.

        But it does make a certain kind of sense when you think about it. As long they’re (modern) Li-Ion cells, we know the voltage is nominally 3.7V per cell (a 1000mAh Li-Ion is not equivalent to a 1000mAh Alkaline, but it is equivalent to two 500mAh Li-Ion cells).
        Ballpark your expectations with a few approximations. Multiply by 85% for regulator losses, times 95% for cable resistance, etc. Your guess is as good as mine how efficient a phone’s charging circuit is.

      2. You’re absolutely right on the watt-hours thing. We should go on a promotional campaign to ask each and every vendor to rate their battery packs in a way that more accurately reflects the potential capacity.

        Your estimate is actually pretty close. I do have one of these battery packs and it only charges my tablet about once before it’s dead. Its better for my cell phone than anything else.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.