DC to DC booster with multiple 18650 cells

[Crimson13]

Question for all the brains out there!

I have a couple simple, single cell, cheap, usb powerbanks.  Each has a single 18650, would there be anything I should look out for, or would prevent me from taking the other 18650 and putting it in parallel with the other on one dc to dc booster? Would there be a limit on how many cells I could add?

I'm pretty sure that since I'm doing a parallel setup, I don't have to worry about balancing the cells.

For anyone wondering, both 18650's are teal ChangJiang cells.

Thanks!

[janekm]

It's not quite clear from your message whether you're planning to parallel them after the 5V boost converter in the power bank, or before?

[Crimson13]

After, looking to expand the capacity of the usb powerbank.

Edit: Basically this: (very poor sketch done on my phone)

[eKretz]

That would be before, not after. And it shouldn't be a problem other than your charging will be slower unless you are charging them outside the booster's internal charging system (if it has one).

You should charge the batteries individually so they are the same voltage before joining them also. Without a method of monitoring individual cell health you could have a catastrophic failure that destroys both cells if one of them fails before the other - but if the cells are the "built-in" protection variety you should be fine IMO.

[Crimson13]

It does have internal charging. It's one of those small power banks with usb out and micro usb in for charging.

Yeah I'm charging the cells now and making sure both are still viable cells, I've had them for a bit.

They are unprotected cells, so I'll just try this for now (if both cells check out) and I'll look into getting some new cells down the road.

[BravoV]

Suggesting to look for power bank branded Tomo V8-4 or Soshine E3, they're relatively cheap imo, as they came "without" the battery, you will have to fill it yourself.

Its capable of accepting four 18650 cells, and each 18650 cell is charged and monitored independently.

[Crimson13]

I've seen those, they look like they work really nice.  But this specific project is constrained by the shape of the container it's going into, and one of those just wouldn't work.  Trying to get this into a ~4inch sphere.  I could fit all of the Tomo in the sphere, but there'd be no place for the lcd on the side of the sphere.

[Zero999]

After, looking to expand the capacity of the usb powerbank.

Edit: Basically this: (very poor sketch done on my phone)

There's still not very clear.

Is the battery on the input or output of the DC-DC converter? Generally inputs are shown on the left of a schematic and outputs on the right, so that would imply you've got them on the output. If this is the case, you can't just use an ordinary DC-DC converter. It needs to be a battery charger.

If the battery is on the input to the DC-DC converter, then there should be an undervoltage cut-off to prevent the cells from being discharged too much and a fuse, so the battery doesn't overheat if there's a short circuit.

[Crimson13]

Technically this would be both.  The DC to DC booster both charges the cells (via mini USB) and provides the five volts out to the standard USB port. It's what would be inside something like this:

(Sorry if the image is large, I'm on my phone at the moment, and can't really judge it's size well.)

[eKretz]

Pretty sure he just drew it backwards to conventional.

[vealmike]

If you are putting unprotected Li-ion cells in parallel, you need to balance them very carefully first.
It's not uncommon for these cells to have a very low internal impedance. Think 30mOhms or so. Unless the cell voltages are almost identical when you connect them up, there will be a very large current and the cells will die...

Inside these unprotected cells, there should be a pressure relief / current fuse. If the cell pressure increases too much or too much current is shipped, the fuse is permanently blown and the cell is dead.
The function of this one time fuse is to disconnect the load / charger before the cell catches fire. They are pretty reliable (unless of course you were daft enough to buy  cheap Chinese cells), but they should be regarded as the last line of defence. You should always use an electronic protection circuit with Li-ion cells.

If you insist on doing this, please, for your own safety, this is how you should balance your cells.

1/ Charge both cells individually.
2/ Connect both cells together as follows. -ve cell 1 to -ve cell 2. +ve cell 1 to terminal 1 of 1 Ohm 10W resistor. +ve cell 2 to terminal 2 of 1 Ohm 10W resistor.
3/ Put voltmeter across resistor, set meter to mV range.
4/ Wait for 0mV on meter.
5/ Wait an additional 1/2 hour.
6/ Remove resistor from circuit, connect +ve terminals together.

If you're really unsure, work outdoors. Paralleling up cheap Chinese Li-ions isn't complicated, but if you get it wrong you could have a lithium fire on your hands.

Oh, if you ever fly with this home made pack, please let us know your flight details! 

[Crimson13]

Thanks for the info Vealmike, I'll be sure to make sure I balance them if I go ahead with these cells. (I'll balance any ones I buy too)

[rs20]

I always imagined that if I were assembling a parallel array like this, I'd connect all the negatives of some charged cells together, and then connect all the positives to a common point using, say, 10 ohm* resistors. Then just wait 24 hours*, and the cells would surely end up balanced, but with the 10 ohm resistor limiting the current to make the operation safe? And then you can remove the 10 ohm resistor and replace with wire links, being confident that no enormous currents would flow at that point, since the voltages of the cells are basically enforced to have been equalised. Is there something wrong with this approach that I'm missing? Seems a lot more reliable that just "fully charging" individual cells, whatever that means.

* I'm pulling this value out of thin air, you'd want to think carefully about worst case voltage differences and currents, both in terms of thermal safety of the resistor and charging current of the lower-voltage cells.
** Also pulling out of thin air; calculations would also reveal the optimal waiting time.

[ebclr]

I would love to see you balancing cells in parallell

[rs20]

I would love to see you balancing cells in parallell

Are you making some smartass comment that I'm "misusing" the word balance to mean "equalise voltages prior to arranging in a parallel array" as opposed to the more conventional "equalise voltages in a series array", or do you actually have something useful to contribute here?

[vealmike]

I would love to see you balancing cells in parallell

I'm missing something. I can't see anyone stating that the cells are "balanced in parallel". Li-ion cells must be balanced before placing them in parallel, because if you don't the current that can flow in the series loop around the two cells can be (very) destructive.

Lithium ion cells pack a very large amount of chemical energy into a small volume and have the ability to release this energy in an extremely short period of time. That's the definition of an explosive.

Unprotected Li-ion cells should not be toyed with lightly. Especially cheap, over the internet, cells of iffy origin. Like smoking, we're never going to stop people using these things. The best you can do is provide a bit of education and hope that you're never on the same aircraft as one of these home made packs.


OP,
You really *should* add protection to these cells. Relying on the internal cell trips would be iffy on a cell from a well respected manufacturer like Samsung or ABC, relying on it when using cells of unknown origin (ordered over the internet they could be anything under that label) is lunacy.

These are really cheap and easy to use parts.
http://www.sii-ic.com/en/semicon/products/power-management-ic/lithium-ion-battery-protection-ic/
I'll help if you need it, post a link to your cell datasheet.

[vealmike]

Is there something wrong with this approach that I'm missing?

No, that's exactly what I'm advocating.

Seems a lot more reliable that just "fully charging" individual cells, whatever that means.

Fully charging here means that in the constant voltage phase of charging, the current drawn by the cell has dropped to 0A. In practise, it never does, so the charger typically cuts off when the current drawn by the cell is <10% of the current used in the constant current phase of the charge. Google Li-ion charging profile for detail.

* I'm pulling this value out of thin air, you'd want to think carefully about worst case voltage differences and currents, both in terms of thermal safety of the resistor and charging current of the lower-voltage cells.

Max cell voltage for a fully charged Li-ion is ~4.2 volts, dependent on chemistry. If one cell is fully charged and the other is deep discharged @~2.5V, then there is a difference of 1.5V. 1.5V/10 Ohms = 150mA. 150mA*1.5V=225mW. So even if someone mucks up and tries to connect a fully charged cell to a deeply discharged cell, it's safe. I'd only balance two cells at a time though.

** Also pulling out of thin air; calculations would also reveal the optimal waiting time.

Harder to do than you'd think. You need to know cell capacity in mAh. Even then cell voltage doesn't change linearly with capacity. And even then the cell will always relax (voltage creeps back) after a charge or discharge as the chemical changes in the electrolyte spread slowly from near the plates into the rest of the electrolyte. Meh, why bother?
Bung a voltmeter across the resistor. 0V means 0A, which means balancing is complete. Then wait a bit longer for luck (and because your meter may not be reading the odd mVs still there) and Robert is your parent's sibling.

[rs20]

I'd only balance two cells at a time though.

Why? If you have many cells and give each cell its own 10 ohm resistor, and connect to a star point, the experience of any given cell will be no more stressful than if you have two cells with just a single 10ohm resistor between them. And although the process may take twice as long (since any pair of cells have 10+10 ohms between them, not the single 10 ohms you propose), you end up with your entire parallel array ready in a single operation, rather than being only one step into some complicated tree arrangement of progressively doubling up cells.

[vealmike]

I'd only balance two cells at a time though.

Why? If you have many cells and give each cell its own 10 ohm resistor, and connect to a star point, the experience of any given cell will be no more stressful than if you have two cells with just a single 10ohm resistor between them. And although the process may take twice as long (since any pair of cells have 10+10 ohms between them, not the single 10 ohms you propose), you end up with your entire parallel array ready in a single operation, rather than being only one step into some complicated tree arrangement of progressively doubling up cells.

Of course you're correct. Brainfart on my part.

[ebclr]

Based and previus posts this guy is very luck, no explosion at all