Series connecting dissimilar battery packs.

[BalorNG]

I want to build a velomobile for long trips.

Currently, I have 7s 80Ah 'main' battery and after a bit of a fiasco with a 200kv motor I'm contemplating a 130kv motor - but to reach desired RPM more voltage will be required.
I contemplate getting an extra 3s ~90Ah auxiliary battery.
Each will have it's own BMS.

First battery will be used as a motor battery.
Second will be wired in series with the first AND will have a separate 12v output for things like lights, cooling fan, etc.
Batteries will be charged separately.

Otherwise using a voltage booster is possible, but at a hit to efficiency which I don't like.

I'm new to this, but it sounds doable to me... or is it?

 

[john61ct]

Not a good idea at all.

Get / build a quality single new pack at the xPyS layout you need.

Lower-volt converter running off the pack as a whole.

10S charging is easy, but if you want to split 2*5S, OK, but same exact cells, mfg date and all.

 

[amberwolf]

I want to build a velomobile for long trips.

Currently, I have 7s 80Ah 'main' battery and after a bit of a fiasco with a 200kv motor I'm contemplating a 130kv motor - but to reach desired RPM more voltage will be required.
I contemplate getting an extra 3s ~90Ah auxiliary battery.
Each will have it's own BMS.

Theoreticaly it will work, however:

--If the BMS's FETs aren't both capable of the full voltage of the whole series set, then if one of them shuts it's FETs now have the full voltage of both packs across them, and can fail if not capable of that. FETs tend to fail shorted, so the pack would then be stuck "on", and unable to stop discharging. Since it was already trying to shut off to protect the pack for whatever reason, this is a problem.

If the FETs are all rated higher than the packs will ever charge to, then that won't be a problem.


The booster idea on the single existing pack will probably work, but it doesnt' just have an efficiency hit, it's also another potential point of failure. :/

 

[BalorNG]

I want to build a velomobile for long trips.

Currently, I have 7s 80Ah 'main' battery and after a bit of a fiasco with a 200kv motor I'm contemplating a 130kv motor - but to reach desired RPM more voltage will be required.
I contemplate getting an extra 3s ~90Ah auxiliary battery.
Each will have it's own BMS.

Theoreticaly it will work, however:

--If the BMS's FETs aren't both capable of the full voltage of the whole series set, then if one of them shuts it's FETs now have the full voltage of both packs across them, and can fail if not capable of that. FETs tend to fail shorted, so the pack would then be stuck "on", and unable to stop discharging. Since it was already trying to shut off to protect the pack for whatever reason, this is a problem.

If the FETs are all rated higher than the packs will ever charge to, then that won't be a problem.


The booster idea on the single existing pack will probably work, but it doesnt' just have an efficiency hit, it's also another potential point of failure. :/

Oh, I see.
Will some sort of shunt help?

 

[amberwolf]

Will some sort of shunt help?

What specifically do you mean?

 

[BalorNG]

Will some sort of shunt help?

What specifically do you mean?

Actually, I have no idea . Maybe, some sort of relay arduino-controlled relay that automatically bypasses the battery once it detects shutdown? Talking of extra points of failure though...

 

[john61ct]

extra points of failure though...

Yes the more extra gadgets attached while in use, the less reliable the system.

Put the investment in simplicity, known-good Grade A brand new cells

retired before they've reached the dangerous EoL stage.

And checking testing gear, monitoring against when-new benchmarks while not in use, so no surprises.

 

[amberwolf]

Actually, I have no idea . Maybe, some sort of relay arduino-controlled relay that automatically bypasses the battery once it detects shutdown? Talking of extra points of failure though...

This could not be fast enough to prevent FET failure if the FETs aren't capable of simply handling it on their own.

Additionally, you'd require a relay that could handle the full current of the system (at the lower voltage of whatever one pack is), as well as the full voltage of at least the pack it would be bypassing, and could be GUARANTEED to never bounce the contacts on rough roads, or else there is a risk of it shorting across the pack while the pack is still active, which could be very bad depending on what the BMS of the pack does under those conditions (and assuming the FETs don't fail).

The same problem could occur if the BMS of the pack detects the pack recovered enough voltage and turns back on: it'll be connecting itself to a direct short (the relay's closed contacts).


It would be safer and simpler to just make the plugs for the battery-to-controller in such a way as to allow you to unplug the battery taht has shutdown, and connect the still-usable battery to the system on it's own. As long as it's keyed or laid out in a way that makes it impossible to connect either one backwards or across the other, etc., then it is much safer than using a switch or relay to bypass a pack by shorting across it.

 

[john61ct]

But wouldn't too great a voltage drop mean a super slow limp-home mode?

Whereas paralleling the "extra unmatched" modules, it's only range you lose when they collapse.

 

[amberwolf]

But wouldn't too great a voltage drop mean a super slow limp-home mode?

Or even a nonfunctional system because the pack could be below controller LVC.

 

[BalorNG]

Actually, I have no idea . Maybe, some sort of relay arduino-controlled relay that automatically bypasses the battery once it detects shutdown? Talking of extra points of failure though...

This could not be fast enough to prevent FET failure if the FETs aren't capable of simply handling it on their own.

Additionally, you'd require a relay that could handle the full current of the system (at the lower voltage of whatever one pack is), as well as the full voltage of at least the pack it would be bypassing, and could be GUARANTEED to never bounce the contacts on rough roads, or else there is a risk of it shorting across the pack while the pack is still active, which could be very bad depending on what the BMS of the pack does under those conditions (and assuming the FETs don't fail).

The same problem could occur if the BMS of the pack detects the pack recovered enough voltage and turns back on: it'll be connecting itself to a direct short (the relay's closed contacts).


It would be safer and simpler to just make the plugs for the battery-to-controller in such a way as to allow you to unplug the battery taht has shutdown, and connect the still-usable battery to the system on it's own. As long as it's keyed or laid out in a way that makes it impossible to connect either one backwards or across the other, etc., then it is much safer than using a switch or relay to bypass a pack by shorting across it.

Hmm, good point, though in my case the amps will be rather small - it is not intended as high-powered drive, but about 150 (300 at most) watts of assist via a 'RC crank drive' (a chain puller, actually), but intended to be used for hours and hundreds of miles.
As for mechanical 'contact bouncing' - is that a frequent problem?

Controller LVC is not a problem - I'll be using a VESC with programable voltage cutoffs.
As for 'limping home' - well, that IS a problem, because voltage is tied to cadence, and too low a voltage will basically stop assist, but if I'll make sure that 'secondary' battery is at least as large or more Ah wise then the main one, it should not happen.

 

[john61ct]

As for 'limping home' - well, that IS a problem, because voltage is tied to cadence, and too low a voltage will basically stop assist, but if I'll make sure that 'secondary' battery is at least as large or more Ah wise then the main one, it should not happen.

Size is not the issue.

For this purpose, you want to **parallel** connect modules that are at the **same** full-pack-level voltage.

They will then stay "in sync" with each other - both voltage and SoC, as if it were a single pack at a higher Ah capacity.

If one module is faulty or just too old, the other is still at the correct voltage, just lower range.

Prevent a catastrophic failure while riding by periodically testing capacity, and proactively remove / recycle / discard and replace it when SoH has dropped too low.

I use 70-75% as EoL myself, others take higher chances.

Do not connect lower-voltage modules together in series, unless done intentionally from the beginning using all matching brand new quality cells with the same mfg datestamp.