Modular 48v 100Wh battery build?

[scm007]

I am wanting to build a modular ebike battery with packs in 100Wh increments so that I can fly with my bike (when COVID is over). Has anybody done this before? What would be especially awesome is if someone has made the modules in a way that they are waterproof so that I could travel with my iSUP, SCUBA scooter, and eBike and be able to use the same battery packs with each. One battery pack to rule them all...

 

[john61ct]

Grin sells 100Wh packs for exactly that, but dunno their voltage.

My idea is to bring out the 6 pairs with a modular connector, allow for wiring any S/P layout.

 

[scm007]

Do you have any idea about this modular connector? I'd also like to not have a BMS per pack and instead have one BMS that charges once the modules are connected?

 

[goatman]

why not use power tool batteries like milwaukee or dewalt

 

[john61ct]

Do you have any idea about this modular connector? I'd also like to not have a BMS per pack and instead have one BMS that charges once the modules are connected?

There are many dozens that would be suitable.

If water resistance is important I like TE Deutsch.

And sure so long as the serial count is consistent, the same pack-level BMS could handle varying counts of paralleled cells down at the 1S group level.

 

[scm007]

Ideally I'm thinking I would set up something like the following:

Starting with 3000mAh cells @ 4V = 12Wh per cell, so 8 of these cells gives us the max allowed by the TSA.

Link these 3p3s = 9Ah @ 12V. This is one subpack.

Each subpack has a connector that allows it to be connected to another pack in parallel or in series.

Using these connectors we connect 4 of these in series to create a 9Ah @ 48V = 432 Wh. This becomes our ebike battery. If we want to create a monster battery we can do this twice and connect in parallel to give us 864 Wh.

Basically I'm hoping for a BMS that can take these 3p3s subpacks and manage them as independent units? Am I doing this right? Would anybody be willing to serve as a consultant on this, I can pay you?

Alternatively a subpack could be 6 cells 1p6s = 3Ah @ 24V. Might be easier to work with high voltage subpacks as all of my electronics will run off 24 or 48V.

 

[john61ct]

IMHO

In order to give good flexibility, the 8-cell (or 6-cell, or 7-cell) sub-packs should not be pre-wired for any particular voltage layout.

48V is a low bar these days, 54V much better, many want 60V or 72V or even higher.

And that's just for ebike use cases, for others you will want 19V or 24V. . .

So the connector just brings 6 pairs out, and the xPyS wiring buss is implemented in a separate

"wiring module" which may also carry the fuses, BMS*, maybe a switch and a display or two.

Inexpensive hobby chargers could be used for charging the sub-packs at lower voltages, quickly with fuse-protected parallel boards

or balancing precisely one at a time when that is needed, each cell then being monitored independently, unlike the usual stupidity when they are welded in parallel.

Any bad cells easily revealed over time and just as easily replaced without disturbing the pack overall.

Since charging safety (HVC functionality) and testing / health monitoring, and balancing functionality are all taken care of by the chargers

the only things left for the BMS is LVC based on 1S parallel-group voltage, if only healthy cells are used could even be based off full-pack voltage

maybe SoC% estimation and display,

maybe an overtemp warning?

If the rider is watching the voltage readout, and can just use familiarity / intuition / common sense to avoid drawing DoD% too low

then maybe the in-use-discharging BMS is not even required. . .

 

[john61ct]

Forgot to mention

3S is way lower than what standard "12V nominal" needs to be, for any LI chemistry.

4S LFP at 3.2Vpc nominal happens to be pretty spot on, but much lower energy density,

a 100kWh pack in LFP would be bulkier and heavier than one built from the usual 3.6-3.7 Vnominal "li-ion" flavours more commonly serving propulsion use cases.

Standard 48V packs using those chemistries are usually 13S, which even then is actually a bit too low, 46.8-48.1Vnom, 52-54.6V charging

which is why 14S packs are very popular, em3ev calls it "50V" more commonly called 52V, actually 50.4-51.8Vnom, 56-58.8V charging.

If you did want a "hard-coded" sub-pack voltage then 7S would make a good increment, two in series gets you that 52Vnom,

still lets you take advantage of cheap hobby chargers, sub-packs on their own give a standard 24Vnom,

just use DC-DC buck converters if you really needed 12V circuits.

 

[john61ct]

Also next multiple for 7S is 21S or 76-78Vnom 85-88V charging

Actually still well below most controller component maximums, e.g. makes best use of Grin's Phaserunner

Lowering charge voltage / capacity utilization just a bit - which will extend cell longevity - can substitute for 20S as a "high 72V"