Hacker News Reader

Comment by mhb | Hacker News Reader

I'd like to do this for my old Thinkpads since I don't trust the random brand replacements to either last a decent amount of time or not blow up. Is there a source of loose matched cells and the protection circuit? Ideally with a guide like the one in this post.

View on HN

Those batteries are generally 18650 cells welded together, but you do need a spot welder for that, and to figure out their voltage so you know how many to put in series. You may be able to just take the protection circuit out of the pack you have now.

Sony and Panasonic are good 18650 vendors.

I love to (or used to before we switched to WFH) check in on the battery recycling bins in the office and look for discarded laptop battery packs. Usually they are full of good 18650 cells (except for 1-2 bad ones that bricked the pack).

Oh yeah, that's usually like $40-$50 worth of cells just tossed.

All the laptop salvage cells I have test with high internal resistance and low capacity, though the packs I salvaged them from were not dead, just low in capacity.

They may be worth $5/cell new, but not so much in their current state.

It's not recommended, but I have soldered in replacement cells with a regular 30W iron. It can be as easy as open the pack, identify and replace bad cells, manually balance charge the cells (one at a time), tape the pack back together. It's a bit janky, because the plastic cases are sonic welded and you have to butcher them a bit to get them open and then glue or tape them back together, but it can be difficult to locate batteries for old laptops and I don't want to spend money.

The problem with soldering them with an iron is that you get the cells way too hot by the time solder has managed to melt, even if you're quick about it.

Not if you work quickly and use 63/37 0.6mm rosin core tin-lead solder with a 30W iron. I never need to apply the iron for more than half a second. I've done this more than a few times, and I realize that it goes against standard safety recommendations, but I'm going to keep doing it.

And if you're recycling cells from other packs, they are likely to have remnants of old tabs still connected, so you can solder to the tabs instead of directly on to the cells, which gives a little more thermal wiggle room.

You can also stick the cells in the freezer before soldering. It adds a fraction of a second to coming up to temperature, and gives the cell proper a lot more thermal wiggle-room. Nothing bad will happen if you keep them above about -20° C.

Definitely file this under "bad advice you probably shouldn't follow, and don't tell them I sent you", but 'allegedly' this works...

That's extremely, extremely risky to do if you don't know the cell's composition, especially the liquid electrolyte, and its performance in cold temperatures. Last thing you want to do is manage to get it frozen and physically damage the layers.

Yes, you should look it up, and pretty much any 18650 will have -20° C or lower as the lower end of its range.

It would be extremely, extremely risky to sell a product which could be destroyed by leaving it in the car in the winter.

I think this advice only earns one "extremely" or maybe a mere "very"...

What are the consequences of getting cells too hot by soldering? Immidiate danger or danger while discharging/charging?

immediate danger

I didn't think that the absolute series voltage would matter much - the right number of cells will be close enough for the voltage regulation in the laptop. My big concern with that approach is matching the cells. It's not practical for me to do that myself since I would need to test a lot of cells. Or maybe I am overestimating the negative consequences of using unmatched cells. But I don't think so.

How do you mean matching them? Just get 8 NCRs or whatever and that's that.

The absolute series voltage matters in that you don't want to over- or undervolt it by 3.7V usually.

Matched in terms of voltage and discharge characteristics. Many of the battery guides mention the possibility of the weakest cell in an unmatched pack being improperly charged. And that that would result in shorter life for the pack or worse consequences.

Eh, sure it matters somewhat. But if you think you're getting matched cells when you buy anything with a battery pack from China you're kidding yourself. So a home made pack isn't going to be any worse off that the cheapest kit you already have. If you go to the trouble to assemble your own pack with matched cells you'll be ahead of the game, if that floats your boat.

Hmm, I was assuming the charger would have a balancer, if it doesn't you can get in trouble, yes.

Maybe you are right that it is a balancing charger. That would simplify things a lot. Although some of the packs have some of the cells in parallel so would have to be carefully matched.

FWIW, old thinkpad batteries make for a great source of 18650 cells to use in flashlights.

Usually only one of them in a dead pack is bad and the rest are great.

Using them in flashlights that take only one 18650 - fine. Using multiple ones - I am not taking the chances (mismatched li-ion is asking for trouble).

Assuming you charge them separately, I don't think there is an issue.

(though usually I use flashlights that take one 18650).

It can be. The really bad outcome is reverse-charging, where one cell is completely drained, then charged to a negative voltage. The probability of an explosion is significant.

If the flashlight has a low-voltage shutoff for the series at 2.5V per cell, it would take a series of four before that's possible, and even then it's pretty improbable. You'd have to mix a discharged cell with three full cells; a moderate difference in capacity or internal resistance wouldn't be enough by itself.

It is easily possible to over-discharge a cell under those conditions, after which charging and using it again is risky, but if you're salvaging laptop cells you probably already know that.

Thanks! I hadn't even considered the possibility of reverse charging really imbalanced cells. Not an issue for anything I use those recovered cells for, but certainly something I'll watch out for.

[this also reaffirms my decision to store lion batteries and their charger in a metal cabinet in a detached garage... because clearly I didn't know all the ways I could turn them into bombs.]

It's my opinion, having been using and reviewing Li-ion flashlights for several years now that single Li-ion cells are very easy to handle safely provided they and the devices you use them in are of reasonable quality. It's mostly a matter of avoiding short circuits and mechanical damage.

Multi-cell devices require a little more knowledge and care, but the main thing is just to make sure the cells match. Being the same model and charged to full before installation will do for new cells, but salvaged ones need to be tested and binned.

Not that easy. They have a battery management controller which contains cell charge data and a fuse which bricks the pack if it’s not handled properly.

Have a good look around on the internet.

I've looked plenty and I think I can handle the issues with the controller though it would be better if I could just replace it. I guess I could also get a quality battery pack (Turnigy?) and take apart the cells, but I'd prefer to buy matched loose cells from a known-good brand (Sony, Panasonic). I haven't seen anyone selling matched loose cells though.

For a while I was scrapping dead laptop batteries to build an 18650 pack for my esk8 build. I ended up building a battery validation rig with a Turnigy battery charger that charges the battery to 4.2V, drains it to 3.8V, and charges it back up to 4.2V to figure out its true storage rating. Then I sorted them by manufacturer, model, and storage capacity, with a recycling bin for anything below 80% of their rated capacity, open cells, and shorted cells.

It's a lot of work, but the validation was all automated via the charger. The pack I ended up with had healthy cells all within 10% of each other in terms of capacity.

By "matched", do you mean binned by the vendor for capacity and internal resistance? I don't know of anybody doing that, but dealers like Illumn might be willing to sell you multiples with the same lot number. I'd be surprised if most OEMs do more than that when building packs.

You can get a cheap analyzing slot charger ($30-40) and do it yourself. That won't be quite as accurate as a big expensive industrial battery tester, but it will be accurate enough to make safe battery packs for consumer electronics.

I usually buy dead Lenovo batteries and take the 18650’s out not the other way round :)

Depends on the model, for T420/T430/W420/W430/X220/X230 there is this:

https://hackaday.com/2020/09/02/building-an-open-source-thin...

https://beta.aceparent.me#/battery

Thank you for the links! By the way there is a follow-up: https://beta.aceparent.me/#/battery2

And the related HN discussion: https://news.ycombinator.com/item?id=24319839

I have the same problem except it is the battery pack for my original Sony Aibo (1st gen robot dog.) I had the cells replaced once using an eBay service but they didn't last very long.

Reply on news.ycombinator.com