Many readers will have at some time or another built their own lithium-ion battery packs, whether they are using tiny cells or the huge ones found in automotive packs. A popular choice it to salvage ubiquitous 18650 cylindrical cells, as [limpkin] has with this 48 volt pack. It’s based around an off-the-shelf kit aimed at the e-bike market, but it’s much more than a simple assembly job.
Faced with a hundred salvaged cells of unknown provenance, the first thing to do was ensure that they were all balanced and showed the same voltage. Some might do this the inefficient way by hooking each one up to a charger and a programmable load, but in this case a much more radical route was taken. A huge PCB was designed with sockets for all hundred cells, connected in parallel through individual series resistors. This allowed them to balance to a common voltage before being discharged to a safe voltage for assembly. Their individual ESRs were the measured, and the best performing examples were then spot-welded into the final 13s-6p final pack.
We all use lithium-ion batteries, but how many of us know how they work?
Um, why not just use 4x 12v lead-gel batteries?
So many cells are a big factor of uncertainty, I think.
48V 570Wh on lead-gel is a lot of dollars. 18650, in comparison, are literally lying around in dead battery packs. Also, “hackaday” not “purchaseaday”.
I’m sorry, I was thinking in terms of reliability. I didn’t have money on the mind.
If you want reliability with lead cells, you can only use 10% of the full discharge capacity each time. A common problem with cheap chinese e-scooters that run on a pair of 12 Volt lead-acid (golf cart) batteries is that they’re usually dead within 1-2 years, or just out of warranty.
Maybe. But I’m still using such a battery that’s from 2007. It’s being recharged daily by a solar panel/solar charger combo. It surely has lost quite some capacity, but it’s not dead yet.
It’s just that in an e-scooter you get the range from going to a very deep depth of discharge, and that kills lead acid batteries very quickly. They say you can go 80 km on a charge, but you’re flogging the battery by going past 8 km.
Weight of lead acid or SLA is about x4 that of lithium ion, which may limit portability
Um, that behemoth here isn’t exactly something I would call portable, either.
I mean, okay, by 1980s definition of “portable” , it might fit the term. 🤷♂️
Best case is lead-acid weighs 5 times as much.
Just use lighter lead. So simple. :-)
The behemoth here isn’t the battery pack, it’s for rapidly testing and cycling used cells that will get built into the battery pack. Look at the last sentence.
Excellent comparison
Agree with you 100%. This battery pack definitely doesn’t have portability in mind, so weight should not be an issue. Cost?…good carrier-grade deep-cycle lead-gel will be “at least” half the price compared to the same capacity in Lithium…generally a 1/4 when looking at really good name-brand Lithium cells (which include LiFePO4).
Lifespan and reliability? Lithium still has a lot to prove, and lead-acid was king for decades for a reason. My first inverter battery set (4 x 200Ah), lasted almost 8 years…being load-shed 3 times a day for most of the year (4 hours every 8 hours). Sucking 30% out of a battery every 8 hours, followed by charging is murder on any battery, no matter the chemistry. My 2nd set has now done just over 2 years, and I am not detecting an obvious drop in capacity (yet, but it will come at some point). As far as reliability is concerned, I think lead-acid is still king, and best bang-for-buck around. They also don’t suffer as much under 0 deg C. Try getting a lot of energy into or out of Lithium in the cold…
It is clear you haven’t tried any lithium based chemistries. LTO is 100% king. It will outlive you, discharge AND charge at stupid high rates, super stable chemistry(won’t catch fire) and way more energy dense than ANY lead acid battery with 100% charge/discharge use. Cost is a tad high, but you can’t beat a cycle life of 15,000+(@100% DOD) for some of them.
What’s the shelf-life?
I wonder if y’all even read the article before formulating an opinion that you decide to share. Your entire comment is based on the idea that the photo you see is the battery pack…..which it is not. It’s the PCB he designed to balance the batteries before packing in their final configuration.
The Hackaday law of comments is that highest ones are commenting strictly on the headline and picture, and will only be full of nitpicks and disapproval.
Why write informed comments when you could be first?
Thats good point about the price and reliability.
If you’re speaking of sealed lead acid batteries like those used in home security systems, to get an equivalent capacity would require either 4 incredibly expensive, large batteries in series or a bunch of smaller ones. It’s rare to find used SLA batteries that still work. Whereas I can always find hundreds, even thousands of good working 18650s at e-waste salvage locations (in my experience, lithium packs are often mistook as being no longer viable when in truth their voltage has simply dropped below the protection circuit’s threshold, requiring a manual charge, bypassing the circuitry using small alligator clips or even soldering directly).
simply because these second hand cells were quite cheap :)
It’s not cheap if they explode like sticks of dynamite. Playing with rechargeable batteries is dangerous, like playing with fireworks. Not all types are high-voltage/high-current capable, for example.
Lead batteries are different here, they can handle a certain amount of abuse (shorts, power surges). They also are sealed, but can “let off steam” if too much gas is being developed inside.
Anyway, I was just wondering.
I didn’t think about economics here, either.
If I have to choose between safety (my life) and saving money, I go for safety.
But each to his/her/they own, of course.
I agree with the reasoning, which is why part of the reason of making that page was to get feedback from the community.
I hope to have used all precaution that could prevent bad surprises:
– capacity and ESR matching
– using the same cell brand
– tested BMS with under/over voltage and over temperature protection
– external fuse
– additional thermal monitoring
At the end, it is a calculated risk… I’d even say that I may have been more cautious than some electric scooter manufacturers.
Without knowing the charging history, voltage/capacity/ESR doesn’t tell you everything about the state of health of the battery.
The reason is that the SEI layer that forms on the electrodes can grow pretty thick before it starts to impact cell performance, so the cell can seem healthy but it’s actually close to keeling over in a matter of a couple years or a couple hundred charge cycles.
oh thanks for that info!
In that case, only Xray would give that info right?
The same goes for the loss of lithium from the electrolyte. You start with an excess and slowly the ions become locked up and unusable into side reactions, but everything works fine as long as you still have extra remaining. Then one day you don’t.
This is the classic picture of how a lithium cell fades: you may be at cycle number 900 (or at calendar year 9 since manufacture date) and think “This battery is still fine. It has plenty of capacity and a low ESR”. Then the capacity crashes from 80% to 60% and nosedives towards zero.
https://www.batteryblog.ca/wp-content/uploads/2013/07/BAK_1-1600-Cycles.jpg
> only Xray would give that info right?
I don’t know what magic tricks they have these days, but impedance measurements at different frequencies should give you more info about the state of electrodes and the electrolyte, or at least show that two cells are behaving the same across a wide span of frequencies, so they are at a similar state of health.
A very good measure of this is the per-cell state after discharge.
In series, you can read the voltage of each cell easily enough. When one cell starts going low faster than the others around it, you have a warning sign.
How do they do it in car batteries?
>When one cell starts going low faster than the others
That’s a large current pulse response that corresponds to the state of the electrolyte. Short cycle (high frequency) pulse response gives info about the SEI layer of the electrodes, because they have a double layer capacitor effect going on. The thicker the SEI the lower the capacitance becomes.
Seeing the voltage sag down and keep falling under load is a very late warning that the cell is just about to croak.
That’s what I meant, thank you for your understanding.
It just came to mind, because I’m used to use lead-gel batteries quite often, also because of my CB radio hobby.
The modern rechargeables are used in walkie talkies (in the form of battery packs) and I’ve noticed that
they do sometimes develop certain illnesses. Like losing one of their cells or developing a bulge.
Anyway, it’s not directly related to
18650 type. It’s more of a general observation. There are special rechargeable battery types like the Emerit one, which can give lead batteries some competition.
This site wouldn’t exist if it wasn’t for weird and wonderful projects like this.
I won’t disagree, it’s just that in my youth, construction kits had certain warning notes written in the manual. They had drawn boxes with notes/reminders in them included. Same goes for books written with the tinkerer in mind. But nowadays, “life hacks” and tutorials in general don’t seem to feel responsible to anything what might happen to the reader. Cartoons nolonger say “don’t do this at home, kids” either.
Warning: this comment might be too long
Warning: this comment will contain grammar errors
Warning: this comment might contain hints of sarcasm
Warning: this comment should not be taken personally
Warning: this comment does not contain much useful information
Warning: this comment contains a personal opinion
Warning: this comment might have too many warnings
A long long time ago I stayed at an hotel which had a nice fireplace in the lounge, all nice and pretty. Heavily shielded with the appropriate glass and wiremesh around it. It also had a pretty large sign next to it, it’s text surprised me. It stated “warning: fireplace can be hot”. Why was there are sign, can’t people think of this themselves?!?
Some things are obvious and do not need to be stated. Why does there needs to be a silly sign to state the obvious, dangerous things are dangerous. If every silly thing is accompanied with a sign, people stop thinking themselves and rely on signs to be present all the time. They will start looking for signs to read instead of seeing the true signs of danger. And with everybody expecting to see signs for every little thing this might eventually lead to too much signs, which is a clear sign that something is wrong with society. And what do we have left then right?
Videos not stating warnings, who cares. If I see a YouTuber jumping of a building with it’s bike, I don’t need a warning. If I see people juggling with batteries I don’t need a warning. If I do want to juggle with batteries, I do my own research and if I don’t then I might screw things up which might cause me to actually learn something the hard way and never make that mistake again. Learning is all about failure, the sooner you’ll hurt yourself the sooner you start to think more thoroughly the next time.
Another fun fact about warning people. The evening news sometimes shows horrible images of horrible events. They justify the showing of it by “the news factor” and by warning us a fraction of a second before showing it. They could have decided not to show it at all, after all they do mention it could be too horrible to watch, so they think it is scary, why do they expect me to watch it then? Fun fact… they never tell you when the horrible images are gone. Which may result in people covering their eyes (and ears, because horrible images come with horrible sounds) for waaaaaaaayyyyyy to long because nobody tells them that it is safe too look again.
“Some things are obvious and do not need to be stated. Why does there needs to be a silly sign to state the obvious, dangerous things are dangerous. If every silly thing is accompanied with a sign, people stop thinking themselves and rely on signs to be present all the time. They will start looking for signs to read instead of seeing the true signs of danger. ”
@Jan I understand what you mean, but that’s not exactly what I meant.
Ok, please let me explain.
a) I’m a 20th century relic
b) What I was concerned about rather is today’s carelessness and “living in the fast line” all the day.
Back in the day™, people spent more time studying things and were willing to learn the basics first.
In this case, it would be basic understanding of chemistry and electricity.
Things like Ohm’s Law, Power, reading electric symbols, knowing basic components, how to solder, what resistance/voltage/capacity are etc.
These days though, it seems people nolonger want to bother with that. They see a cool 45 seconds long TikTok video and decide replicate what they see. No research, no practicing. No experience.
Then, there’s lack of interaction with people in real life.
That’s something was available back then. If you bought parts in a Radio Shack type of electronics shop, you could get advice by the staff after getting into a little chat.
Not always, but there was a chance.
This is an improvement, IMO. Let Darwin do his work.
Sheesh, there’s always someone claiming that even looking at a 18650 will make it explode and take out the entire eastern seaboard. It’s not true. These cells are used *everywhere*, in the millions. If they weren’t essentially safe then there would be millions of fires instead of the occasional one due to severe abuse that gets blown out of proportion by the media.
Can you set one on fire? Sure. Can you treat them with basic respect and never have to worry about them causing a single problem? Also yes.
“if they weren’t essentially safe then there would be millions of fires instead of the occasional one due to severe abuse that gets blown out of proportion by the media.”
True. On the other hand, the original USB connector wasn’t meant to be used for power chords, either.
But that’s exactly what it got popular for. I’m thinking of USB A and Mini USB in particular. The USB-C connector, by contrast, was designed for that purpose.
>the occasional one due to severe abuse
Note that this article is about recycling and re-using cells from other (broken) battery packs, which pretty much constitutes abuse on these cells. You don’t know how bad they had it, until they start smoking on you.
Mmmh. Actually not so. 99% of the ‘used’ battery thing isn’t from broken packs, but ‘old’ packs. Fun thing is, it’s more cycling than age that kills batteries ( some nuance there but overall true).
So take an old modem pack that’s been pegged at full charge it’s whole life, and shuck out the batteries. Generally, they’re still pretty healthy. No spring chickens true, but good. Add a good bms circuit and short of physical abuse (like skewering a tesla with a guardrail abuse) or general shorts, the risk of explosions is negligible.
And while bulky, I could lift 100 18650s. The same capacity of angried up pixies in lead gel would need wheels to have any hope of moving by my hands.
>it’s more cycling than age that kills batteries ( some nuance there but overall true).
I find it the opposite: in most uses you never reach the full cycle potential of the battery and instead the cells and the pack become broken by having always been kept at near full charge, charged or discharged at very high currents, possibly at extreme temperatures, or kept drained of charge for extended periods of time – all until the user notices degraded performance or the BMS calls it quits and shuts down the battery.
Things like electric scooters. Imagine you have a 50 km range and a pack that nominally gets 2,000 charge cycles. Are you going to ride 100,000 km on that scooter? Probably not. Same thing with laptops: do you cycle the battery every day? Probably not – it’s sitting tethered to the wall cooking the cells to death instead. A power bank? A power tool? A flashlight? Again, rarely cycled, most of the time spent forgotten at the bottom of a drawer.
So you got a battery pack that was neglected and abused, showing symptoms of wear, charge imbalance, or cutting off when loaded. Now some of the cells are in near self-destruct mode while others are fairly OK, and yet others are so-so that will become dangerous later on if paired with good cells – the difficulty is in telling which ones are the bad eggs.
>pegged at full charge it’s whole life
That’s actually one of the ways to kill a lithium ion battery sooner than later. High cell voltage and/or high temperature causes leakage through side reactions which degrade the electrolyte.
Describing 18650 cells like “little sticks of dynamite” is the problem here. Describe what actually happens: a release of gas/liquid and if ignited, a chemical fire. Which the original poster mentions several times throughout his blogpost.
Don’t be a safety nerd. Don’t discourage learning/experimenting in the name of safety. *Especially* in an environment like Hackaday that encourages projects like this.
DO encourage safe practices like people in the poster and people in the comments mention including cell balancing, BMS, fuses, etc.
What causes the release of gas/liquid is usually a thermal runaway event, which also causes the battery to self-ignite. It’s not like a stick of dynamite, but more like a small cartridge of wet gunpowder.
It’s perfectly safe, as long as you don’t put many of them in a sealed container.
Hmm, well done, a nice post & it makes sense.
Although & I understand the setup primarily a static aspect, I’d be tempted to add a
dynamic aspect, need to think on it as a tad concerned paralleling blocks & then serial them,
I’d go the other way, serial as many with IC BMS on each then parallel a 42 v block (matches
a typical electric bike oldie lead acid batt) – then less chance of singular catastrophic
failure pulling massive current in one shorted cell – ie part engineering and part
connection efficiency with risk assesment at the expense of actual $ bear in mind $ cheap
in relation to leverage…
But, thats just me. Have a layout in design but, too many abstract things demanding
attention ;-) Will get there though & yes its true these cells more common and most seem
to be rather good condition – offers opportunity for even more products/exploration, tah
Good to see posts like these, thanks
I’m actually working on a project that would allow parallelizing and monitoring battery pack like these!
I made sth similar for 6 6s drone lipos.. Absolutely get the soldering part, this took 2 hours with testing
https://postimg.cc/V521Ttbm
way more portable than mine :D
They’re lipos? They just look like chunky resistors.
WoW, never seen these, can you advise a source and cost – bulk discount ?
Thanks
the battery project I am working on is useing
2v 208ahr cells,rated at 8 hrs,and I have 48 for
2 strings at 48 volts plus a dozen spares
as noted,heavy
I get them from a steel supply/metal recycle
business that is WAY ahead of the curve and
it is possible to buy anything that they comes in
and as I buy lots of new steel,they give me a guy
with a forklift to dig a pile of a batteries out
or stailess tube of used I beam for cheap
lithium batteries
,which I know will be very difficult to run through any current(ha) recycling system,which
is realy quite violent,and the chance of fire is 100%
so they will be land filled unless people can take on the danger of many tons on variously damaged
and used batteries
point bieng:they will remain a cheap source of good storage for anyone willing to do the work
plastic cases with dumb funny screws,maybe glued,and or more funny click locks,then its all
welded together,and maybe leaking the flamable
electrolyte,hey HEY!
I built 2 48v volt battery packs for my wife and my bikes. Based on new 18650. The BMS did not go down to the individual cell level. Years later they still serve us well.
Love it. I’ve been building lithium-ion batteries for years and this is kind of intriguing.
For all the ‘it will explode’ people, seriously how else are you to keep life interesting.
I keep telling the neighbours that a kid or two might get vaporized, they keep sending them to help.
If you’re going to make a PCB to balance 78 batteries in parallel, and worried about having to order 5 PCBs, then why not make a PCB to charge 16 batteries and be able to string 5 of them together? Overall, it would be much cheaper.
It indeed crossed my mind to make stackable PCBs. In the end I didn’t go for it due to the extra wiring “complexity”
The cost savings on the PCB order (as well as being able to get them reasonably from virtually any fab) and ability to scale the size of the conditioning bank, as well as be able to employ any excess of the smaller boards for other projects seems like it would have still been a good tradeoff. Note that the boards needn’t have been “stacked” – mounting holes could have been on each board in/out parallel and serial configurations could have been run to edges, and they could be screwed down to a backer board and in the end looked pretty similar to your large PCB, sans the large band at the bottom. The extra wiring for interconnecting would have been minimal compared to the rest of the work. I somehow doubt the traces on the PCB would last long if you did try to pull 100A through it though.
Another benefit to an array of PCBs: if something went wrong (say some cell starts weeping, or you burned out a trace or two while testing output), you could extract the damaged boards and individually replace them: with one big board you have more work in front of you to effect a repair.
I’m a fan of LFP (LiFePO4, Lithium Iron Phosphate), because while they may not be as energy dense (Wh/Kg) as NMC, they are MUCH less prone to catastrophic failure if subjected to some abuse. While not my intention to abuse them, in high power stationary projects, you’re more likely to burn your home down if something goes wrong. So, the safer chemistry seems wise.
Also, I have the benefit of having access to some nice batteries: 50Ah LFP cells with threaded studs, which are easy to configure with beefy interconnects (stout wire with crimped ring terminals OR metal plates with two holes drilled in them and thick shrink tubing over the midriff), and studs means no spot welding needed and cells can be swapped in and out easily if a problem does develop.
At 50Ah, each cell is equivalent to approximately 20 2500mAh cells in parallel (though being LFP, they’re at a lower voltage than their NMC cousins).
So, what are you going to use the battery for? What sort of PSU are you using to charge the pack (since 13S NMC tops out at 54.6V) ?
A more appropriatelabel for the board would be “non-extinguishable fire and deadly fumes when shorted!”
Great write up on making a battery!
You said that EV battery packs used huge cells. That is not true. All the cells are smaller than a “C” alkalline battery.
Some cars used LiFePo cells which can be quite large.
For instance, https://www.youtube.com/watch?v=VHKYNrVR9Dk
Er no, there are a few sizes even as far back as 2020 – yes 3 years ago – the 4680 by Tesla
https://www.aboutenergy.io/post/the-tesla-4680-does-size-matter
I want a few of those – all sorts of reasons *grin* anyone know a retail supply ?
“RELEASE THE ANGRIED UP PIXIES”
good one Will
And while bulky, I could lift 100 18650s. The same capacity of angried up pixies in lead gel would need wheels to have any hope of moving by my hands.