lithium ion pack partial discharge

[joshcarr]

I have a 36v li-ion battery pack made from 18650s with a BMS. It's main purpose is to power an e-bike but I have a headlight that runs off of a separate 2 18650 pack. Can I just wire two of the 18650s from the main battery pack to the headlight? Will this discharge them disproportionately or would the BMS manage that? Thanks.

[ovnr]

You shouldn't. The BMS only balances while charging, not discharging. And you will find that the BMS will cut off the entire pack once the two cells you connect to run out - which they will do much sooner than the rest of the pack. Recharging it will likely also be really slow and error-prone, due to the large imbalance.

Best solution is to just get a suitable buck converter and hook that up to the full 36V pack voltage.

[joshcarr]

Okay that's about what I figured. I'm having a hard time finding a buck converter that will go from 36v to 7.2v which is why I was wondering about this method.

[ovnr]

Well, there are these 60V units. I've got a few of them, and they work fine up to 40V at least. Not the most terrifically effective solution, but it should work - how much power does your headlight use?

Alternatively, find a new driver for the headlight, something like this one.

Lastly, there are of course "proper" buck converters like these ones, but they're $35 a pop and completely overpowered (250W max output).

[james_s]

I've used those LM2596HV and related parts before, both the Chinese modules (which are often fake but serviceable) and the real deal. I've built a few regulators deadbug style on copper clad and was pleased with the performance, I see no reason why that won't work provided the lamp doesn't draw too much. It's so easy to use these things now, you can pretty much just use the circuit right off the datasheet.

[Rick Law]

Okay that's about what I figured. I'm having a hard time finding a buck converter that will go from 36v to 7.2v which is why I was wondering about this method.

I suspect your problem being 36V in?  Most others I've seen supports only up to 30V in.

How many amps do you need?  If it is 2A or less, try the digital B3603 Buck Converter.  It support input as high as 40V.  I wrote a mini-review for that buck converter some time ago.  I learned a lot more since, but I still love that B3603.  Read that if you are going to use 2Amps, I wrote about the heat issue if output amperage is high.

^{https://www.eevblog.com/forum/reviews/b3603-dcdc-buck-converter-mini-review-and-how-the-set-key-could-be-fatal/}

[ovnr]

I suspect your problem being 36V in?  Most others I've seen supports only up to 30V in.

I'm taking the safe option of assuming that OP has a 10-cell pack with a nominal voltage of 36V.

Which means 42-43V fully charged.

[Rick Law]

I suspect your problem being 36V in?  Most others I've seen supports only up to 30V in.

I'm taking the safe option of assuming that OP has a 10-cell pack with a nominal voltage of 36V.

Which means 42-43V fully charged.

I think your assumption is better than mine!

I did not read the OP properly.

[james_s]

The LM2576HVT is good for up to 60V input and can deliver 3 Amps.

[joshcarr]

So after a month of fiddling and waiting on packages from china the adjustable regulators all only go up to around 6 amps that I can find. My light pulls 15 amps so that's not gonna work. I melted the solder right off of the first one I tried. I looked into some 20 amp regulators like the on suggested above, https://www.digikey.com/product-detail/en/tdk-lambda-americas-inc/I6A4W020A033V-001-R/285-2503-ND/5878834, but I'd rather not spend $35 on a regulator to run a $20 light. Any other solutions or is the $35 one my best bet? Also yes it is 36v nominal, fully charged is 42v

[soubitos]

Quote from: joshcarr on Today at 11:27:42 PM

I'd rather not spend $35 on a regulator to run a $20 light. Any other solutions or is the $35 one my best bet?

Could this

200W 15A DC DC do your thing?

[ogden]

Other solution: separate battery pack for light.

7.2V*15A = 108 watts of LED light!!! Most likely light output is more than most car headlights Anyway - then 250W 35$ converter does not seem overpowered. Maybe you can find similar one around 150W or so.

[joshcarr]

Soubitos that looks pretty good nice find. Ogden it is very bright, I've often compared it to a car headlight myself. It came with it's own pack and I could use that but that's an extra thing to charge and worry about. Getting rid of the separate battery was kinda the point of this whole fiasco lol

[paulca]

For ideas

[james_s]

Other solution: separate battery pack for light.

7.2V*15A = 108 watts of LED light!!! Most likely light output is more than most car headlights Anyway - then 250W 35$ converter does not seem overpowered. Maybe you can find similar one around 150W or so.

Unless it's grossly inefficient then it is definitely more than car headlights. Standard halogen headlights are typically 55W low(dip) beam 60W high(main) beam so with a pair figure 110W/120W. Any reasonable LED light should be *at least* double the efficiency of a halogen incandescent source so that's got to be one heck of a light.

Maybe it's easier to re-engineer the LED driver? Usually high power LEDs are made of several smaller chips in series or series-parallel arrangements using more than 12V across them.

[GeorgeOfTheJungle]

Are you sure? A hundred watts for the headlamp of an ebike is crazy! If you want/need that much light I would put 21 * 2v white leds in series instead, or 23 * 1,8v, or... etc, without any regulator. Piece of cake.

Number of leds needed = (1 or 2) + int(40v/LEDForwardVoltageDrop)

[paulca]

Putting my road safety hat on.  Exceptionally bright LED headlights on bikes work against you, not for you.  There seems to be the belief that brighter is better to be seen.

While it is true that a car will see the brighter light easier, they won't see YOU.  They won't have the faintest idea where you are as all they can see is a scalding star of white light.  You could be 2 meters away or 20 or 200.  A lot of drivers will take that chance you are 200m away.

I find them irritating as I cannot judge the cyclist's position.

Use a sensibly powered headlight and DIP IT towards the road.  This does not dazzle the driver, especially in mirrors and gives me a slight chance of being able to work out were YOU are.

Safety hat off, I'm away to do some donuts.

[metrologist]

the super bright bike lights are even better when they flash - 10 Mega candelas to the eyeball.

[james_s]

I find those flashing ones to be distracting and disorienting. It makes it even harder to see what's going on.

[viperidae]

I've had a few cheap AliExpress lm2596hv's fail while running off 48v input. Ended up sending 48v to a device rated at 9v. Exploded a capacitor with a lot of smoke.

[ogden]

I've had a few cheap AliExpress lm2596hv's fail while running off 48v input. Ended up sending 48v to a device rated at 9v. Exploded a capacitor with a lot of smoke.

No wonder it failed because specified for max 45VDC (attach). Usually you shall have safety margin, especially for clone chips
[edit] Missed "HV" part which means max 60V, but anyway don't assume they use genuine chips

[james_s]

Almost all of those LM2596 boards from China use fake counterfeit parts that are something that functions similarly but is a completely different part. They seem to work fine at modest voltages but if you try to load them up to the full rated current or approach the max rated input voltage they will often pop. Use genuine parts and they will work.

[soubitos]

Quote from: james_s on Today at 10:11:13 PM

Almost all of those LM2596 boards from China use fake counterfeit parts that are something that functions similarly but is a completely different part. They seem to work fine at modest voltages but if you try to load them up to the full rated current or approach the max rated input voltage they will often pop. Use genuine parts and they will work.

It would be pretty cool to know what part it really is they are using under the LM2596 label!

[james_s]

I'm amazed they can sell anything at all for the prices they do, I've bought many complete models of various types for 99c or less including shipping. For that price I couldn't ship the empty packet to the next street, let alone the other side of the world.

[paulca]

Now to address your question, do not discharge from a partial string of series connected Li-ion batteries, unless the power is very low, and you have an active bidirectional BMS.

There is nothing wrong with discharging one or two cells in a 3 cell pack via the balance lead.  (Or 1, 2, 3 cells in a 4 series pack).

As long as you don't put the pack into service again until the cells have been re-balanced with a proper balancing charger.  I wouldn't trust a BMS to do this, you need at least a hobby grade balancer charger.

They may be frowned on for various reasons, but a Turnigy Accucell 6 set to "balance charge" will charge individual cells at about 0.1A until they are within about 100mV before it ramps up the charge current.  Thus if you use 1 cell from the pack to run a 3.7V device and bring that cell down to 3V, giving a pack voltage of 3V, 4.2V, 4.2V the charger will effectively charge the 3V cell up to 4.1V before kicking off the charge properly.

Unless you can confirm this behaviour of the charger in advance I would not attempt this however.

So in the OP's use case, the pack is in service, it isn't at all advised.