Author Topic: DIY cheap powerbank (technical suggestions R welcome !) (Read 38529 times)

qnh · « **Reply #25 on:** August 08, 2016, 05:16:56 pm »

Quote from: ym58 on August 08, 2016, 01:53:58 am

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 !

I agree with you because Lithium batteries are getting cheaper and cheaper, and even free (as those you salvaged !).
For me, it makes no sense to 'maintain cycle life' (quote from @Siwastaja) at the expense of charging time ... it simply does not worth it : 12-14hr versus 3-4hr ... this range is just too wide !
In Germany, we have recycling bins in nearly all shops, with treasures inside : Li-Ion, Li-Po etc ... recycling and using those batteries after testing and selecting them is definitely a sustainable attitude ...

Siwastaja · « **Reply #26 on:** August 08, 2016, 07:53:19 pm »

When charging at 0.25C, the charging time to almost full is four hours. At 0.5C, it's two hours. Talking about 12 hours is nonsense. We are talking about topping the final <1% in the CV phase, which is irrelevant. CV phase is never much more than one hour max in any practical case, and you don't need it at all if you don't want to do it.

At low C rates, the cells are already very close to full when the CV phase begins. For example, at 0.5C, a typical laptop cell charged starting from 0% will reach approximately 90% charge after 0.9*2hours = 108 minutes, and topping off the last 10% would take maybe another hour. However, there is absolutely no reason to go to 100%, if you are happy with, say, 95%, which is quickly reached.

So, when you seldom start from real 0% (but something like 10% instead), and don't need to get to 100% (95-97% being enough), C/2 charge is really very close to 2 hours.

Note that the CV phase gets shorter when the charging current is lower, so it compensates for the longer CC phase!

About maintaining cycle life: what I mean is that your salvaged cells abused at 1C charge may be dead after just 5 such cycles! It's not just worth doing, building the pack takes your time and physical resources. You are exceeding the design margins. Just go a bit lower. Even a tiny bit helps, there may be a huge difference between 1C and 0.9C. I recommend below 0.5C to have some margin. Only charge at the rate you really need. If it must be fast, then sure. But note that you can't fully fast charge anyway, the CV phase will take longer accordingly when you increase the CC current.

Note that "Battery University" is quite bullshittish. This graph, for example, shows li-ion cell being charged starting from 1V, which is utter bullcrap. As usual, they have no clue what they are teaching about.

Many sources, including cell datasheets, also list CV phases that are way longer than actually needed in real life, giving the idea of very slow charging. The time required blows up exponentially when you are trying to squeeze the last 0.5%, but I never understood why you would want to do it. It's logical for the manufacturers to specify the capacity this way, however.

Kilrah · « **Reply #27 on:** August 08, 2016, 11:04:50 pm »

Quote from: Siwastaja on August 08, 2016, 07:53:19 pm

When charging at 0.25C, the charging time to almost full is four hours. At 0.5C, it's two hours. Talking about 12 hours is nonsense. We are talking about topping the final <1% in the CV phase, which is irrelevant.

Not at all... we're talking about commercial powerbanks that house 4 or more 18650s i.e. 12Ah@3.3V or so, yet still charge them with only 600mA@5V out of a USB port... which is means charging with less than 1/10C and is what results in the whole thing taking those 12h or so.

ym58 · « **Reply #28 on:** August 09, 2016, 02:43:48 am »

Quote from: Siwastaja on August 08, 2016, 07:53:19 pm

When charging at 0.25C, the charging time to almost full is four hours. At 0.5C, it's two hours. Talking about 12 hours is nonsense. We are talking about topping the final <1% in the CV phase, which is irrelevant. CV phase is never much more than one hour max in any practical case, and you don't need it at all if you don't want to do it.

No one will argue on this ... you're right.
But as Kilrah just explained in the previous post, we are not talking about an "ideal" single 16850 charge process but about a "strong" pack of two (or more, but in my case it's two) 8.2Wh 3.7V LiPos wired in parallel !
1C, in such a configuration, would theoretically be 4.4A (for the two batt)... that means that charging them at 0.6A equals to charging them at 0.15C or so, hence the unacceptable charging time ... and hence my quest on this thread to get a better circuit than FM9833E with its low 0.6A max charging current !

Quote from: Siwastaja on August 08, 2016, 07:53:19 pm

Note that "Battery University" is quite bullshittish. This graph, for example, shows li-ion cell being charged starting from 1V, which is utter bullcrap. As usual, they have no clue what they are teaching about.

"Battery University" is not that crappy in the sense that it sums up in a readable and structured form all pertaining knowledge of the battery world including the latest technologies.
Some credit should be given to them, at least on these grounds.
That being said, you surely have noticed that there are numerous reader comments attached to nearly all the articles published in "Battery University" ... which means that some constructive criticism is posted there, allowing readers (like me) to put the published articles into perspective !
---
Do you know of any other knowledgeable site about the battery world that you could recommend us to read ?
---
BTW, I have corrected the improper 1-volt extrapolation that you came across in their graph, that will make it more acceptable I guess ...

Siwastaja · « **Reply #29 on:** August 09, 2016, 07:14:22 am »

OK, I understood you guys wrong, because of how the OP stressed the need to charge at 1C, at the same time talking about 12 hours. Indeed, charging at C/12 takes 12 hours by definition, and it's impractical because it's considerably longer than a night. To fix that, I recommend charging at C/2 if absolutely necessary, or preferably a bit less (C/3 or C/4).

The problem with Battery University is the amount of misinformation and urban myths originated from there. The way the information is presented clearly does not help; quite the opposite, it leads the reader to enjoy reading and to believe it's a credible source. However, it appears to me that the quality has gone up considerably. Just some 5-10 years ago, it was total bullcrap to such a high level that it was a huge cringe to try to read through if you knew how the things are in reality. Still, most readers took their word because of the nice style of presentation, and most internet forum urban myths regarding li-ion originate from there, even though most of that stuff has already been removed or rewritten.

Showing faked graphs with severe misinformation "just as examples to point something else out" is a huge problem, you should never ever do it, especially when trying to teach someone else! People will always refer to such graphs and interpret every piece of information they can get from those, because graphs are powerful format of conveying facts using data. This graph doesn't even state that the voltage information is nonsense, which would be the minimum requirement. Although, it would be easiest to simply use a graph with correct information - but that, of course, requires that you have the correct information, which is not true with "Battery University". I guess the author just simply thought that li-ion cells are discharged down to 1V. From the charging algorithm point of view, the starting voltage is an extremely important parameter - a cell starting at 1.0V doesn't go though that kind of charging graph at all.

ym58 · « **Reply #30 on:** August 09, 2016, 07:39:50 am »

Quote from: Siwastaja on August 09, 2016, 07:14:22 am

Showing faked graphs with severe misinformation "just as examples to point else out" is a huge problem, you should never ever do it, especially when trying to teach someone else! People will always refer to such graphs and interpret every piece of information they can get from those, because graphs are powerful format of conveying facts using data. This graph doesn't even state that the voltage information is nonsense, which would be the minimum requirement. Although, it would be easiest to simply use a graph with correct information - but that, of course, requires that you have the correct information, which is not true with "Battery University". I guess the author just simply thought that li-ion cells are discharged down to 1V. From the charging algorithm point of view, the starting voltage is an extremely important parameter - a cell starting at 1.0V doesn't go though that kind of charging graph at all.

Actually, there is a second graph just below that one on the same page, did you see it ?

Siwastaja · « **Reply #31 on:** August 09, 2016, 08:06:05 am »

That graph looks less wrong and more realistic, although many [new] cells today perform better, i.e., reach much more than 60% at the end of CC even at 1C charge.

That being said, even this graph shows the starting voltage clearly at or below 2.5-2.8V, which would most likely cause the real charging algorithm to go to the "conditioning" mode (or simply report "broken cell" and refuse charging), so even this graph is wrongish, although not catastrophically so.

qnh · « **Reply #32 on:** August 09, 2016, 04:15:35 pm »

I think we are getting a bit off topic ... being obsessive on why we should NOT let any Lithium battery discharged under 3 Volts is sort of pointless as it makes sense to everyone here.
I am also interested in finding a chip (preferably a "cheap chip" as ym58 quoted previously) capable of getting one or two Li-Ions/LiPos single cells charged in a decent time (few hours) and also able to provide a min of 2A output current.
What would then be the best alternative to FM9833E ?

crazyguy · « **Reply #33 on:** August 09, 2016, 04:55:00 pm »

very cheap chip

Power Bank SoC
http://www.injoinic.com/english.html

IP5108
Charger Current 2.5A, Boost Converter Output Current 2.1A
http://www.injoinic.com/doc/IP5108_brief.pdf

http://www.gearbest.com/chargers/pp_421906.html?wid=21
http://www.chongdiantou.com/thread-2357-1-1.html

ym58 · « **Reply #34 on:** August 09, 2016, 06:28:26 pm »

Quote from: crazyguy on August 09, 2016, 04:55:00 pm

very cheap chip
Power Bank SoC
http://www.injoinic.com/english.html
IP5108
Charger Current 2.5A, Boost Converter Output Current 2.1A
http://www.injoinic.com/doc/IP5108_brief.pdf
http://www.gearbest.com/chargers/pp_421906.html?wid=21
http://www.chongdiantou.com/thread-2357-1-1.html

That's just unexpected ... there has been so much blah-blah on this thread that I was close to give up finding credible alternatives to the FM9833E chip !
Let me review those specs accurately and I'll surely have further questions to ask you then ...
Thanks to you HK crazyguy

for being so concrete after all this claptrap

Be back soon.

qnh · « **Reply #35 on:** August 09, 2016, 06:45:23 pm »

Quote from: crazyguy on August 09, 2016, 04:55:00 pm

very cheap chip

Power Bank SoC
http://www.injoinic.com/english.html

IP5108
Charger Current 2.5A, Boost Converter Output Current 2.1A
http://www.injoinic.com/doc/IP5108_brief.pdf

http://www.gearbest.com/chargers/pp_421906.html?wid=21
http://www.chongdiantou.com/thread-2357-1-1.html

Siwastaja · « **Reply #36 on:** August 10, 2016, 06:06:57 am »

Quote from: ym58 on August 09, 2016, 06:28:26 pm

That's just unexpected ... there has been so much blah-blah on this thread

I may have saved hundreds if not thousands of your hours by informing you of the problems related to your idea of always charging li-ion cells at 1C (partly based on my own, yet-to-be-published research), and by informing you that Battery University is not a good source for a beginner like you, which you, ignoring my expertise, started to blah blah against without better knowledge. I clearly should have seen at that point that you are not worth replying to.

Designing something is more than just taking a random solution and applying it; understanding the relevant principles is paramount. I always try to stay relevant when talking about said principles, to save everyones time and avoid diverging to unnecessary nitpicking and blah blah.

Everything in this thread has been concrete and important, and if you deny to see it, it's your problem and you'll just face it later.

I didn't feel I need to diverge to charging voltage, but wanted to reply to the claimed premise that Battery University is "just fine" by showing all the severe errors that come up with every graph posted originating from them. If they can't even get the basic voltage profile right, how can you trust them in other aspects, either?

It would have also helped if you could write more clearly. When you keep blah blahing about charging at 1C, and reason it so that you don't have to wait for 12 hours, I logically conclude that you have have been misinformed (by sites like Battery University, for example) that charging at less than 1C takes 12 hours. This kind of misconception is not uncommon; not all people understand C rates at all, and that's all OK, as long as you are willing to learn.

But now I know I don't need to invest any more of my time in you. Thank you and good bye.

Skimask · « **Reply #37 on:** August 10, 2016, 06:49:48 am »

^^^^
This
(or is it "That"?)

With that being said, I've been using those cheap ebay PCBs that come with the TP4056 & DW01 along with a bunch (bunch = 12) of relatively equally used 18650's (same brand, same 'color', look to be roughly the same age, etc) for a few months now, with the pack feeding a TPS63001 prototype PCB upconverting to 5V @ ~1A, connected to a 4pk of USB-A plugs. Takes forever to charge the whole thing if it's dead-dead, but lasts forever (maybe slightly less) on a charge. 12x ~2200mah 18650's = a bit less than 100WH...for the cost of some junk

Pictures to follow...(Don't expect too much. It's thrown together with duct-tape, bailing wire, and clothespins. Someday I'll find that fire-proof for the thing...just in case)

ym58 · « **Reply #38 on:** August 10, 2016, 07:51:30 am »

Quote from: crazyguy on August 09, 2016, 04:55:00 pm

very cheap chip
Power Bank SoC
http://www.injoinic.com/english.html
IP5108
Charger Current 2.5A, Boost Converter Output Current 2.1A
http://www.injoinic.com/doc/IP5108_brief.pdf
http://www.gearbest.com/chargers/pp_421906.html?wid=21
http://www.chongdiantou.com/thread-2357-1-1.html

Hello Crazyguy,
A few questions :

Thanks !

ym58 · « **Reply #39 on:** August 10, 2016, 07:59:27 am »

Quote from: Skimask on August 10, 2016, 06:49:48 am

^^^^
This
(or is it "That"?)

Quote from: Skimask on August 10, 2016, 06:49:48 am

With that being said, I've been using those cheap ebay PCBs that come with the TP4056 & DW01 along with a bunch (bunch = 12) of relatively equally used 18650's (same brand, same 'color', look to be roughly the same age, etc) for a few months now, with the pack feeding a TPS63001 prototype PCB upconverting to 5V @ ~1A, connected to a 4pk of USB-A plugs. Takes forever to charge the whole thing if it's dead-dead, but lasts forever (maybe slightly less) on a charge. 12x ~2200mah 18650's = a bit less than 100WH...for the cost of some junk
Pictures to follow...(Don't expect too much. It's thrown together with duct-tape, bailing wire, and clothespins. Someday I'll find that fire-proof for the thing...just in case)

We're in the same boat together, buddy ... trying to make practical things from scrap/junk without spending a dime

Awaiting your pics.

crazyguy · « **Reply #40 on:** August 10, 2016, 08:11:28 am »

Quote from: ym58 on August 10, 2016, 07:51:30 am

Quote from: crazyguy on August 09, 2016, 04:55:00 pm
very cheap chip
Power Bank SoC
http://www.injoinic.com/english.html
IP5108
Charger Current 2.5A, Boost Converter Output Current 2.1A
http://www.injoinic.com/doc/IP5108_brief.pdf
http://www.gearbest.com/chargers/pp_421906.html?wid=21
http://www.chongdiantou.com/thread-2357-1-1.html

Hello Crazyguy,
A few questions :

Thanks !

(1) The IP5108 brief datasheet said VREG (pin 12) is 3.1V LDO Regulator Output, 50mA (max)
I think it means the L4 (pin 13) is connected to the VREG (pin 12)

IP5108 datasheet
http://www.injoinic.com/doc/IP5108_brief.pdf

(2) the table
green color -> IC Part Number, IP5101 ....
yellow color -> discharge current (A)
pink color -> charge current (A)

(3) according to the package, I would guess it is a IP5108 chip
https://item.taobao.com/item.htm?id=534722661076&ns=1&abbucket=5#detail

ym58 · « **Reply #41 on:** August 10, 2016, 09:43:10 am »

Quote from: crazyguy on August 10, 2016, 08:11:28 am

(1) The IP5108 brief datasheet said VREG (pin 12) is 3.1V LDO Regulator Output, 50mA (max)
I think it means the L4 (pin 13) is connected to the VREG (pin 12)

I am not quite sure I understand how VREG works.
What that 3.1V LDO Output is supposed to control ?
And why should TWO outputs (VREG and L4) be connected together ?

Quote from: crazyguy on August 10, 2016, 08:11:28 am

yellow color -> discharge current (A)
pink color -> charge current (A)

Just to make sure I got it all clear :
Charge = constant current (CC) delivered by the chip to the batteries during the charge cycle
Discharge = current that the chip is able to output to the load (smartphone, iPhone ...) via the two 5V/VOUT pins (5-6)
Is that correct ?

ym58 · « **Reply #42 on:** August 10, 2016, 09:56:46 am »

Quote from: blueskull on August 10, 2016, 09:50:28 am

Form the datasheet it says VREG is regulator output. The reason they are connected to L4 is because (I guess) L1 and L4 are also logic inputs sampled at power up, to determine whether the IC operates in 3 LED, 4 LED or 5 LED mode.

crazyguy · « **Reply #43 on:** August 10, 2016, 10:00:54 am »

Quote from: ym58 on August 10, 2016, 09:56:46 am

Quote from: blueskull on August 10, 2016, 09:50:28 am
Form the datasheet it says VREG is regulator output. The reason they are connected to L4 is because (I guess) L1 and L4 are also logic inputs sampled at power up, to determine whether the IC operates in 3 LED, 4 LED or 5 LED mode.

yes, you are right, 3, 4 and 5 LED configuration used for battery level indication.

crazyguy · « **Reply #44 on:** August 10, 2016, 10:14:39 am »

Quote from: ym58 on August 10, 2016, 09:43:10 am

Quote from: crazyguy on August 10, 2016, 08:11:28 am
(1) The IP5108 brief datasheet said VREG (pin 12) is 3.1V LDO Regulator Output, 50mA (max)
I think it means the L4 (pin 13) is connected to the VREG (pin 12)
I am not quite sure I understand how VREG works.
What that 3.1V LDO Output is supposed to control ?
And why should TWO outputs (VREG and L4) be connected together ?

Quote from: crazyguy on August 10, 2016, 08:11:28 am
yellow color -> discharge current (A)
pink color -> charge current (A)
Just to make sure I got it all clear :
Charge = constant current (CC) delivered by the chip to the batteries during the charge cycle
Discharge = current that the chip is able to output to the load (smartphone, iPhone ...) via the two 5V/VOUT pins (5-6)
Is that correct ?

IP5108 is a 2 in 1 power bank solution (charger + booster)

correct, Charge means (battery charger current), Discharge means (boost converter output current)

ym58 · « **Reply #45 on:** August 10, 2016, 10:56:24 am »

Quote from: crazyguy on August 10, 2016, 08:11:28 am

(3) according to the package, I would guess it is a IP5108 chip
https://item.taobao.com/item.htm?id=534722661076&ns=1&abbucket=5#detail

I am very interested in the IP5109, but don't feel like making my own PCB around this chip

Do you know of any powerbank that would use this chip (as the Miller ML-106 with the IP5108) ?

crazyguy · « **Reply #46 on:** August 10, 2016, 02:17:53 pm »

Quote from: ym58 on August 10, 2016, 10:56:24 am

Quote from: crazyguy on August 10, 2016, 08:11:28 am
(3) according to the package, I would guess it is a IP5108 chip
https://item.taobao.com/item.htm?id=534722661076&ns=1&abbucket=5#detail
I am very interested in the IP5109, but don't feel like making my own PCB around this chip
Do you know of any powerbank that would use this chip (as the Miller ML-106 with the IP5108) ?

IP5109 chip in operation is very hot, > 115C

IP5109 demo board review
http://www.chongdiantou.com/thread-1731-1-1.html

no idea which powerbank use this chip

Miller taobao shop
https://miller162.taobao.com/
ML-106 V1.0 (IP5206)
https://item.taobao.com/item.htm?id=534400802077

Charger ML-102 V9.0 2015
use IP5206
http://lygte-info.dk/review/Review%20Charger%20ML-102%20V9.0%202015%20UK.html

ML-102 V9.0 Power Bank (Russian)
http://mysku.ru/blog/china-stores/35821.html

IP5108 and IP5206 are pin-to-pin compatible

ym58 · « **Reply #47 on:** August 10, 2016, 05:09:02 pm »

Quote from: crazyguy on August 10, 2016, 02:17:53 pm

IP5109 chip in operation is very hot, > 115C
IP5109 demo board review
http://www.chongdiantou.com/thread-1731-1-1.html
no idea which powerbank use this chip

---
I'd like to test this IP5109 board and use chunks of this :

here :

and there :

What do you think ?
Where can I purchase one ?

crazyguy · « **Reply #48 on:** August 11, 2016, 03:50:57 am »

use a large heat sink will increase the size of the powerbank

I don't think you can purchase this demo board anywhere, send a email to the IC company maybe they would be kind enough to give you a copy for testing

ym58 · « **Reply #49 on:** August 11, 2016, 04:51:52 am »

Quote from: crazyguy on August 11, 2016, 03:50:57 am

use a large heat sink will increase the size of the powerbank

Well it's not so large, the width is only 14mm (0.55 inch) and the height 6mm (0.23 inch) !
And I was thinking I could trim its length according to the hot surfaces it has to be glued on on each side of the PCB ...

Quote from: crazyguy on August 11, 2016, 03:50:57 am

I don't think you can purchase this demo board anywhere, send a email to the IC company maybe they would be kind enough to give you a copy for testing

That's what I'll do.
Is Changdiantou a distributor only or the original IC company ?
Apparently, their website is about video rentals ... ?!?


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: DIY cheap powerbank (technical suggestions R welcome !) (Read 38529 times)

Share me