DIY cheap powerbank (technical suggestions R welcome !)

[ym58]

My bottom-line was to build a cheap but yet fully functional and powerful power bank using :
- a PCB salvaged from a Chinese 2euro-powerbank
- two LiPos (8.2Wh each) salvaged from a defective Sony Vaio battery pack

Kind of suspicious association  ... but  let's try it  :


The final boxed prototype looks a bit like a trendy NYC designer's techno thingy  O0 ...


It works  of course but there are two flaws breaking my initial rules :
#1 it takes >12hrs to recharge its two LiPos (from a regular 2A wall charger)
#2 the output current is limited to 1A or so (making it less ***powerful*** that I had expected)

The reason being (well, I suppose) that the FM9833E the salvaged PCB is based on is limited to 0.6A for the battery charging current and to 1A for the output current.


I tried to lower the R3 resistor (down to 1K) in the hope that this could increase the battery charging current but to no avail.
If anyone has any suggestion how to increase the battery charging current, please come back to me in this thread !

If I can't improve this PCB "as it is", my next option will be to try another PCB like this one that I have just ordered on eBay :



Any suggestion is welcome !

[qnh]

Nice (and cheap) work ... sustainable economy  !
I just can't find info concerning the chip that you mentioned (FM9833E).
Do you have any datasheet on your side. ?

[JacquesBBB]

The problem with these boards is that usually you cannot charge the battery while you use it.
I prefer a different setting with
two very cheap boards

- one  for charging/ discharging the battery with protection
http://www.ebay.com/itm/5pcs-5V-1A-USB-18650-Lithium-Battery-Charging-Board-Charger-Module-Protection-/322045031371?hash=item4afb6133cb:g:RDUAAOSwRgJXigY4

- a second one for providing the proper voltage as requested.
It can be USB
http://www.ebay.fr/itm/401089512266?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

- or a boost converter for higher voltage .

they can be connected together in a very small package.

total cost < 1$

[ym58]

Nice (and cheap) work ... sustainable economy  !
I just can't find info concerning the chip that you mentioned (FM9833E).
Do you have any datasheet on your side. ?

Actually, the link was embedded in my previous post !
Here again  >>> FM9833E <<<
Any technische Klärung  will be much appreciated.

[ym58]

The problem with these boards is that usually you cannot charge the battery while you use it.

Hello Jacques,

But that was none of my intentions !

These so-called powerbanks are made to store energy from a wall charger and then to be able to transfer it back to charge any mobile device that might have run short of battery "on the road" ...
In a powerbank, those two different functions (store then give-back) are mandatoty but yet not concurrent (concomitant) !
---
The PCB I have salvaged does the job but does not really fullfill the charging requirement for two (2) parallel 8.2Wh-LiPos , i.e. a CC (constant current) around 1C, that is about 2A per battery = 4A total !

The IC used on this board (FM9833E) can only deliver a mere 0.6A charging current to the 2 batteries, hence the extended duration of the charging process (>12hrs, instead of 3 to 4hrs at 1C !)

[qnh]

I looked at the .PDF you gave me for the FM9833E chip, but it's all in Chinese !!!
I did not see much info relating to the charging current setting, the only thing that I undersatnd is that R1 - 2K  (connects to pin4=PROG) may have a role in this ... but I don't know how !
However, I might have found something interesting for your research, it can be found there :
http://mysku.ru/blog/aliexpress/38446.html
Sorry, it's in Russian this time  , but if you scroll down a bit, you will see that there is some info concerning the same chip as yhe one you use.
Maybe someone here, speaking Russian, can help you ?

[ym58]

However, I might have found something interesting for your research, it can be found there :
http://mysku.ru/blog/aliexpress/38446.html
Sorry, it's in Russian this time  , but if you scroll down a bit, you will see that there is some info concerning the same chip as yhe one you use.

I have Google-translated the Russian document and learned (well, as far as the Google translation is accurate) that the REcharging current, flowing from the FM9833E circuit to the battery(ies) can be decreased from 0.6A by ways of increasing the R1 pulldown resistor (connected to the PROG pin of the chip), but apparently NOT THE OTHER WAY ROUND  
---
I think that FM9833E-based circuits won't fit my requirements, which are not so high (1A charging-CC per battery and 2A output current).
I am not requesting a QC2.0* or QC3.0* compatible circuitry, just a basic QC1.0* would be satisfactory

Has anyone here got any clue ?

* @Qualcomm Inc.

[Ranger14]

See most Chinese use one chip wonder that not available in the US ... I am not sure if there are chips available in the US that charge and boost??? Power banks are a whole another beast when building one .. Most power banks run 18650 cells in parallel which most people recommend not doing.. Another issue is the charging current for the batteries. Most usb sockets can only output 500ma to 1amp and can't charge a single cell at 1c.The charging IC they use is only meant to charge a single 18650 cell at 1amp but they still use it to charge a battery bank.. Personal I can't a charge IC to charge parallel 18650 cells .. The list just keeps going on but can be done if you don't settle on certain specs..

[ym58]

Most power banks run 18650 cells in parallel which most people recommend not doing..

Yeah, I've even seen cheap powerbanks claiming > 20000mAh capacity with FOUR (4) 16850 in parallel !
The sky is our limit
In my case, I selected TWO (2) 8.4Wh Lipos carefully ascertained to be compatible (same discharge capacity, same full charge voltage) before I can use them in parallel.
Besides, those two batteries were already paired in the Sony Vaio Battery Pack that I salvaged for the need of that pick.

Most usb sockets can only output 500ma to 1amp and can't charge a single cell at 1c.The charging IC they use is only meant to charge a single 18650 cell at 1amp but they still use it to charge a battery bank.

Yes, that's why I intended to REcharge my PowerBank prototype on strong wall sockets only (at least 1.5A/5v), but alas the limitation comes from the chip itself and not from the wall socket.
I even used my lab power supply (5A max) during my tests for me to monitor voltage and current more easily and accurately than with that :

---
However,  a right boost circuitry (with the correct inductor), would normally do the trick even from a 0.5A wall socket source.
What's frustrating here is that this chip (FM9833E) can't output any charging current better than 0.6A ...  even 2A would have made me happy (1A per battery) !

[ym58]

The datasheet states 0.6A maximum charging current and 650mR PMOS charging passive element. Which means, at the best case with even no current limitation, you get (5V-4V)/0.65=1.5A.
Also keep in mind that when the inductor saturates, some really bad thing will happen, so make sure the inductor on board saturates at higher than foreseeable max ripple current.

Thanks for the translation, Blueskull.
---
I think that  you're right, the die is cast !
I won't be able to draw anything better than 0.6A from that chip.
---
Hoping that the next board (that's on its way from China) will be better on that point :

---
Do you have any info on that one ?

[ym58]

Actually, I just received this morning the new PCB from HK (via Istanbul !!!)
---
I am just quite upset coz it worked for 30min in discharge mode (one piece of fully charged Lipo connected to B+/B- and a 4R/100W resistor load connected to the USB-OUT 2.1A output).
The LCD was then just working fine, with a nice '95%, 5V-2.1A' message (with percentage decreasing smoothly) and I effectively measured a 1.2A discharging current with my DC clamp ammeter ...

Then, all of a sudden, it just stopped and now it's dead  and the LCD does not even light any more!

I tried to reverse the process and charge another half-charged Lipo thru the microUSB, the LCD lights up correctly and gives a coherent percentage (57%, i.e. the charge state of the current LiPo) but actually it does not charge at all the battery, the current is near to zero ... no way, the PCB is dead !
---
Since I had ordered another one to another supplier, I must now wait another couple of days/weeks before I can resume my tests.
---
However, I made a pic of the dead PCB and indicated the chips references, just in case someone would got a clue whether this PCB (in working condition ) could possibly charge LiPos with currents > 1A ... ?

[edavid]

Why not charge each cell with a separate TP4056, and combine the outputs with some big Schottky diodes?  You'll lose a little capacity that way, but at least it will work.

[ym58]

Why not charge each cell with a separate TP4056, and combine the outputs with some big Schottky diodes?  You'll lose a little capacity that way, but at least it will work.

A TP4056-based circuit will only handle the CHARGE of the LiPos( or Li-Ions).
What I want is a reversible circuit (exactly as those cheap ones described earlier in the thread) that can not only CHARGE but also manage the DISCHARGING (with correlative protection) of the LiPos into the external load (smartphone, iPad, tablet ... you name it).
Actually, I think they call this a BMS, for Battery Management System.
---
And also, my bottom line was to make a funny prototype with what already exists ans sells a few bucks, instead of designing something that will eventually cost more than a real 20000mAh LiPo-based powerbank (around 25€-35€) !

[Kilrah]

instead of designing something that will eventually cost more than a real 20000mAh LiPo-based powerbank (around 25€-35€) !

Good luck then, since I don't even actually know of any "real" powerbank that does what you want (charge at >1A).

[ym58]

Good luck then, since I don't even actually know of any "real" powerbank that does what you want (charge at >1A).

At least, that one does (amongst dozens) :
https://www.conrad.de/de/logilink-powerbank-12000-mah-1040721.html
---
Let's make a quick calculation if you will :
This 12000maH PB (powerbank) claims it can deliver 12Ah (ouch !)
That means that it either includes ONE 12A LiPo or TWO 6A LiPo (in parallel) ... I dont really think it could be more than TWO (see the thickness).
Then 1C for each battery is either 12A (1 batt) or 6A (2 batt) but not less !
---
Therefore, the charging current (I said CHARGING, not DISCHARGING ) should NORMALLY be either 6A or 12A to stay within the frame of NORMAL charge cycle diagrams (that is a few hours charging, generally less than 6hrs depending on the battery technology and manufacturing quality).

Then, if you are right (see your quote), that would mean that it takes that PB MORE THAN TWO DAYS  to get charged ... no way  !
CONRAD.DE could not possibly be so dumb to include such a poor device in their listed products !
---
Don't forget that even if a wall socket can deliver only 1A to the PB's micro USB charging connector, some internal switch-mode circuitry is intended to increase this current !

[NottheDan]

Looking at the linked pdfs they quote a charging current of 200-1600mA and a charging time of 15 hours at 1.5A.

Though, looking at a 19600mAh powerbank of the same brand they only vie a charging time of 12 hours at 1A. And the 20000mAh one gives 10 hours at 2A.

[ym58]

Looking at the linked pdfs they quote a charging current of 200-1600mA and a charging time of 15 hours at 1.5A.

Thanks Dan, shame on me  I did not even look at the specs in French (I only sticked to the second .PDF, less precise ...!).
---
So, 'Max 15 Heures sous 5V 1.5A' is not THAT bad.
But still, how come one can't find any cheap chip  that could deliver higher constant currents to charge LiPos (or Li-Ions) ... how can the mainstream market possibly accept such poor charging time performances ?
---
Any LiPo (or Li-Ion) can be TECHNICALLY charged within a few hours with a proper CC (constant current) equaling its 1C capacity.
How can we, customers, accept specs stating 15Hrs to charge a 5-ounce battery ?

[Kilrah]

Then 1C is for each battery is either 12A (1 batt) or 6A (2 batt) but not less !

Nothing requires you to charge a Li battery at 1C, it's the maximum you can do on normal cells. The cell count in the powerbank is also irrelevant, it's the energy capacity that matters.

About your Conrad powerbank:

Pack "rating" (not always realistic, let's assume it is) is 12Ah @ 3.7V so 44.4Wh, so 8.8Ah equivalent @5V. They say "up to 15h", 8.8/15 =.... 0.6A.
So this one also charges at the usual 600mA (losses ignored, it will probably actually draw about 7-800mA or so).

The reasons for the low charge speed is a mix of convenience, cost, size and "no need for more".
Having an USB charge port allows not supplying a proprietary charger (expensive, inconvenient to carry around for the user). But given that you have to assume the user might use any USB charger, including one that might not be rated for >500mA, and a charge controller that handles that actively is too much to ask from the cheap manufacturers.
Then a 5-10W charge circuit would start to get pretty big compared to the power bank itself. Lastly, as we can see, that's how things are and nobody complains, most will actually rather want a cheaper powerbank than a more expensive one that charges faster... most people have them in their bag just in case "for emergencies", once back home it's not a problem to charge it all night. There are some that charge faster, but of course they cost more.

Even I who have been using standalone Li-Pos for more than a decade and have universal chargers have never found the need to spend the 5 minutes needed to crack open my powerbanks to wire them to a fast charger, I just plug them to an usb charger laying around and forget them there for a while.

[NottheDan]

Random thought: Having you looked into the balance chargers available to RC model enthusiasts?

[ym58]

https://www.amazon.com/Xiaomi-20000mAh-External-Battery-Portable/dp/B01A89C3K2/ref=sr_1_3?s=wireless&ie=UTF8&qid=1470507890&sr=1-3&keywords=xiaomi+power+bank
Spend some more money for better quality. Xiaomi is one of the best Chinese personal digital gadget manufacturer.
When they say their products passed CCC/CE/FCC, they mean it.
Warning: Chinglish.

Apparently, this one is QC2.0 certified, that's a big step forward compared to my cheesy stuff !!
---
Besides, reviews are not all good, is Xiaomi experiencing QC problems or is it just coincidental ?
---
BTW, your Chinglish's quite good, way better than my froglish !!!

[ym58]

Nothing requires you to charge a Li battery at 1C, it's the maximum you can do on normal cells. The cell count in the powerbank is also irrelevant, it's the energy capacity that matters.

You're bloody damn right, but on the other hand nothing requires us to be happy with long-lasting situations ... otherwise there would be no chance for any progress !
Personnaly, I always charge my batt packs at 1C, be it with my lab P/S or my iMax B6 !

[Kilrah]

You're bloody damn right, but on the other hand nothing requires us to be happy with long-lasting situations ...

Of course not, but given the current situation if you wanted to do better $5 Chinese boards is not really the right place to look. Of course there are solutions to do better, but then you need to forget the condition I quoted in my first post becasue you're going to be making what amounts to a "premium product".

[Siwastaja]

Most power banks run 18650 cells in parallel which most people recommend not doing..

... Which is a prime example why you shouldn't base a design on a group vote by random people who have no freaking clue what they are talking about.

18650 cells are most definitely used in parallel, which is their intended way of use, and has always been.

[Siwastaja]

Personnaly, I always charge my batt packs at 1C, be it with my lab P/S or my iMax B6 !

Charging any random energy-optimized (as opposed to power-optimized) li-ion packs at 1C is very likely to cause severe lifetime issues on some, if not most cells.

Most energy cells are rated at 1C absolute maximum charge rate, compromising the lifetime already, even with non-aged cells which still work optimally.

On random, aged, salvaged cells I strongly suggest limiting yourself to about 0.5C maximum occasional charge rate, preferably less than 0.25C as your everyday charge rate, to maintain good cycle life.

RC toy cells are optimized for high power discharge, and as a consequence, can tolerate higher charge rates too, so 1C is quite a normal everyday charge rate there, occasional peaks being somewhere around 3C. But this is not the case with consumer electronic cells which are optimized for higher energy density and lower price, not high power. There will be serious lithium plating, causing the cycle life to plummet. It can happen suddenly, a certain cell can work for 500 cycles just fine at, say, 0.6C, and be destroyed in less than 50 cycles at 0.8C. You can't know without destructive testing, so keep that charging slow enough.

[ym58]

18650 cells are most definitely used in parallel, which is their intended way of use, and has always been.

I agree, and I suppose this remark should also work for LiPos.
The best example being laptop battery packs that contain Li-Ions or LiPos which are ALWAYS paired in parallel.

On random, aged, salvaged cells I strongly suggest limiting yourself to about 0.5C maximum occasional charge rate, preferably less than 0.25C as your everyday charge rate, to maintain good cycle life.
RC toy cells are optimized for high power discharge, and as a consequence, can tolerate higher charge rates too, so 1C is quite a normal everyday charge rate there, occasional peaks being somewhere around 3C. But this is not the case with consumer electronic cells which are optimized for higher energy density and lower price, not high power. There will be serious lithium plating, causing the cycle life to plummet. It can happen suddenly, a certain cell can work for 500 cycles just fine at, say, 0.6C, and be destroyed in less than 50 cycles at 0.8C. You can't know without destructive testing, so keep that charging slow enough.

Thanks for the advice, that's good to know and take into consideration when it comes to pricey devices.
---
Concerning the prototype described in this thread don't forget that LiPos have been salvaged (from a Sony Vaio laptop battery pack) and that I don't really care about their durabilty ...
My main concern is to reduce significantly the charging time that stretches over 12-14Hrs at 0.6A (that is 0.3A per battery) !
I would prefer to have them charged in 3-4hrs (as most smartphones/laptops do) even if this means I will have to dispose off them after 50 cycles !