I just watched a basic you tube vid on making a DIY tesla type power pack to dump excess PV generated electricity into to use outside generation period.
It uses 2nd hand 18650 3.7V lipo cells found in old laptop batteries and power tools etc.
I've about 60 odd from knacker Makita 18V batteries I've been collecting, most are individual good as the problem is usually only with the first cell in the 10 pack.
Anyone got any experience of building up pack with these cells and or useful links
I think 50 odd would give me 500W for an hours if my very basic calcs are correct

I'll stick it outside
found some good info on ebiker.com re pack building

found an nice solderless pack design that's worth more thought
https://endless-sphere.com/forums/viewtopic.php?f=14&t=57810

Ive never been the faint hearted type :)

If you have a heap of these cells and can separate out the good ones then giving them another life seems like a good scheme. I particularly like the idea of recycling old (crashed or partially dead) electric car battery packs. Seems to be a lot of it about - in Norway.

However, I still don't see the point of using this small form factor (18650) if you're starting with new cells, whether as a one-off DIY pack or if you're Nissan or Tesla. Particularly for stationary applications where weight and volume isn't quite so critical, bigger LiFePOâ‚„ cells seem a lot more convenient (less fiddly connection) and safer (the chemistry is intrinsically more stable). It's difficult to compare like with like but the reduced amount of packaging for bigger cells must help, I think.

With the huge numbers of 18650s made for laptops, etc, it's easy to understand why Tesla started off with them - being able to play manufacturers off against each other must be nice. But it'll be interesting to see if the change when they've got their own production facilities running. E.g., double the length and diameter while sticking with the same chemistry.

I can't see how fewer bigger cells with the same chemistry wouldn't have a better power to weight ratio. But you might well be right - it's plausible they'd be more vulnerable in an accident.

Whatever, the trade-offs on chemistry and form factor for portable/mobile vs stationary applications seem different to me.

James, what are your plans for managing the balance of the pack?

Torrent is right... not for the faint hearted, especially not with used cells, and not without tabs. You can weld your own tabs with a simple-ish homebrew setup but if you are buying new cells anyway it's not worth the trouble.

I've soldered quite a few packs of up to sixteen tabbed cells, including adding the balancing circuitry for 2S, 3S, 4S (available cheaply on ebay) into those packs. Using new cells like this is fine for ebike and laptop batteries but not worth the trouble of using used cells IMHO. A single cell failure (made much more likely with used cells) is pretty inevitable and will likely be loads of trouble to detect and fix even if it does not take the rest of the pack with it. Large packs also suffer from local heating at poor cells that reduces the capacity of the cells around it.

I know that the original Tesla Roadster used 7000 of these things but their pack design balancing and cut-out circuitry was a bit fancier and they still suffered a lot of problems with reduced capacity and short life; hence all the interest in battery upgrades for them.

If you buy new 18650s then I would stay away from Ultrafire - they are already recycled and repackaged laptop cells and their performance is very variable.
But if I wanted 500Wh or more I would be looking at new prismatic cells instead of 18650.

There are some interesting LiFePO4 cells and batteries starting to appear (including motorcycle and marine battery formats - see SHorai LFX) but they are still very pricey from what I have seen and I think they still don't match the capacity per volume/weight of older Lion cells.

It's not a dead idea, but requires a lot of thought. Only worth it if you can get the tool pack/laptop cells for practically free and have a lot of time to devote to it!

Some ideas off the top of my head:

1) You need some mechanism to grade the cells. The guys on endless sphere have numerous threads on cell testing involving artificial loads, magnets etc.

2) You should make it as modular as possible. Rather than one big battery I think you should use an architecture that lets you add extra batteries as time goes on. They need to be connected to each other via diodes (to prevent a charged battery trying to dump 100's of Amps into a flatter one!), Schottky diodes at a minimum, better to use an "ideal diode" circuit. (Tiberius on endless sphere or pedelecs forum has a good design).

3) Use something like that no-solder assembly to allow you to build up and tear down the packs easily.

4) With old and mixed cells like this you really are going to need BMS circuits to actively keep the packs in balance. You probably won't get away with just charging and then assembling and expecting them to stay in balance.

Well if not active balancing, at least some row by row voltage monitoring and low voltage cutoff (as described above) is essential, as well as some means of isolating a row to balance charge it off line. Basically with the kind of cells suggested i.e. lots of different types/ages/capacities etc there's more chance of things getting out of whack, so building it with monitoring and maintenance in mind is a good idea.

James if you haven't found it yet here's the gold mine for info on recycled battery pack building:

https://endless-sphere.com/forums/viewtopic.php?t=26383

Well it'd be fun to watch if you tried (outside with a bucket of sand to hand!)