Anyone else using old laptop batteries wired together instead of buying proper cells?

Done this exact thing as a proof-of-concept for my shepherd's hut build before committing to proper cells.

Pulled apart about 15 old Dell and Lenovo laptop packs, tested every 18650 individually with a cheap capacity tester off Amazon (the blue OPUS one). Ended up with roughly 60% of cells being usable — anything below 1500mAh or showing high internal resistance got binned immediately.

Built a small 4S4P pack, balanced it manually, and ran it through a Victron BMV-712 to monitor properly. Honestly? It worked. Powered LED lighting and a small 12V compressor fridge for weekend testing. But here's where I'd pump the brakes for anyone considering it seriously:

• Cell matching is absolutely critical — mismatched internal resistance causes imbalance under load, which gets worse over time
• No BMS = genuinely dangerous. I used a cheap Daly unit, which I wouldn't fully trust either
• Capacity is unpredictable — the cells degrade differently and your usable capacity drifts

For anything permanent I wouldn't bother. I eventually went with Fogstar Drift 280Ah LiFePO4 cells which cost considerably less than I expected and the peace of mind is worth every penny. The 18650 experiment probably cost me 20 hours of faffing for a pack I trusted about 40%.

That said — for learning the fundamentals of cell chemistry, BMS configuration, and capacity testing before spending real money? Genuinely valuable exercise.

Anyone else gone down this rabbit hole? Curious whether anyone's managed to source higher-quality pulls from EV battery packs rather than laptops — those 21700 cells from older Teslas seem interesting on paper

@SprinterProject curious about your testing process — did you do capacity testing or just voltage/internal resistance checks on each cell?

I've been tempted to try something similar for my narrowboat as a low-stakes experiment before committing to a proper Fogstar or EVE cell build. The main thing holding me back is the sheer time investment sorting cells, and whether mismatched capacities cause headaches even with a decent BMS.

A few questions if you don't mind:

• What BMS are you running with them?
• How are you grouping cells — matched capacity or just matched IR?
• Any thermal concerns given the mixed age/chemistry?

The shepherd's hut use case seems quite forgiving compared to marine, which is probably why I'd want to test the concept somewhere less critical first.

@SprinterProject what did you use to test capacity on the individual cells? I've got a pile of old ThinkPad and HP packs sitting in a box on the boat that I keep meaning to pull apart, but I'm nervous about committing time to it without a proper cell tester.

Is something like a SkyRC MC3000 worth the outlay for a one-off job, or is there a cheaper route that's still reliable enough to trust the results? Also wondering how you're handling the BMS situation — are you running each salvaged pack's original BMS, rolling your own, or something like a Daly or Overkill Solar unit across the whole bank?

The capacity variation between cells must be a headache for balancing.

Been down this exact rabbit hole with my tiny house build. Pulled apart a mountain of old Fujitsu and HP packs before I'd committed to proper Fogstar cells.

The capacity tester that changed everything for me was the OPUS BT-C3100 — not glamorous, but it'll charge/discharge each cell and give you actual mAh figures rather than just a snapshot voltage reading. Internal resistance alone will lie to you; a cell can look healthy until you actually pull current from it.

@MuddyTinker ThinkPad cells are surprisingly decent survivors in my experience — IBM/Lenovo historically used better quality cells than some budget laptop makers.

One thing nobody mentions enough: sort your tested cells into tight capacity-matched groups before wiring anything in parallel. Mismatched cells drag each other down constantly and it gets ugly fast over a winter.

@SprinterProject this is almost exactly where I started with my shepherd's hut before I eventually went Fogstar cells. Tested probably 200-odd 18650s from old Dell packs and ended up binning about 60% — anything under 1500mAh or with internal resistance above 150mΩ got chucked.

The ones that passed I ran in a small test lash-up over a winter. Main lesson: cell drift over time is the killer. Even matched cells start diverging after a few charge cycles without a proper BMS keeping them honest.

What BMS did you end up using? I cobbled together a cheap Daly unit which worked but felt nervous about it. Would genuinely like to know if there's a better budget option before I consider doing this again for a van conversion project I've got in mind.

Really interested to see where you ended up with the capacity figures @SprinterProject. The sorting and grading process is genuinely the most time-consuming part - you'll end up with maybe three tiers of cells and the temptation is to use the middling ones "somewhere" which usually means they just sit in a box forever.

One thing worth flagging for anyone else reading this: do check the internal resistance alongside capacity. I've had cells that tested reasonable on mAh but had terrible IR, and they'll drag down whatever parallel group they're in. A decent charger like the Opus or a dedicated tester handles both measurements.

@MuddyTinker the Liitokala Lii-500 is a solid budget option for testing individual cells if you don't already have something.

Curious what voltage you're targeting for the overall pack @SprinterProject - 12v nominal or going higher?