Are high #'s of BMS Channels (200) possible? affordable?

[robots4joey]

Hello Everyone!

I've been reading alot here, and I am excited with what I have learned so far.

Here's my situation:

I have received about 3,000: 18650 lithium-ion rechargeable cells for free! <2200mAH, 3.7V> Most of them are perfectly fine and hardly any cycles on them. So, I was thinking about building a EV, but I'm going to start semi small...

I have a golf cart with dead batteries, so I thought I'd start there: a 37Volt, 100 AH battery pack made from these li-ion cells.

From my calculations, that would take about 500 cells; 10 series sets of 50 in parallel... I was thinking that I would try to make a BMS system to balance 10 parallel cells at a time. (I picked 10 because of the cost issues with each BMS channel- more later on that)

Using these figures I'll need 50 BMS channels to make this pack work. AAHH! If I use the current BMS DIY system on this site (By GGoodrum) That would cost around $300! I don't have that kind of money for a golf cart

So, Are there other designs out there for large numbers of BMS channels? I would like to make a full EV with these cells eventually- that wold take 200 channels of BMS... so I'm looking for a super cost effective solution... (maybe $1 a channel??? too much dreaming?)

unfortunately I am not super electrically inclined... I have done circuits before and I understand the principles, but to design it myself would not be real likely...

Maybe I could charge more than 10 in parallel? I was thinking that # was high already though.

Thanks for the advice!

-Joey

 

[Doctorbass]

Welcome here robots4joey!

Prepare yo to enter the E-S team of used cells second life !!

for 37Volt, 100 AH this is more like 10s 50p and not 50s 10p!.. this last one would give 185V at 22Ah

Ok.. so you said you have 10s 50p right?

With that i can see that you will only need a bms of 10 chanel no more!

You need one ch of BMS for every parallel groups you have anyway your paralle group have one or 100 cells.

All cells are connected in parallel in a paralle group so the voltage to monitor is the same for every cells in it.



Where come from all your cells? laptop? powertools? sputnick?

Doc

 

[fechter]

Doc is right, you really only need 10 channels of BMS.

It would be good to try and test the cells before assembling them into a pack since it will be nearly impossible to locate a bad cell in a parallel group. Also keep in mind that these kind of cells have the potential to catch fire rather catastrophically. Don't try to use them without a BMS of some sort.

For only 10 series, it might be easier to use 10 separate chargers or dc-dc converters like Doc uses.

 

[robots4joey]

Ok... This is kinda funny.

s is for series and p is for parallel right? just had to make sure I got that straight.

I was under the impression that it was advisable to only charge like 3 18650 cells in parallel (so 3.7 Volts and 6600 MAH 'groups')

I was going to jump the shark and charge 10 at a time, all in parallel (3.7 Volts and 22AH 'groups') So I had thought that I had to break up my 50 cells in parallel in groups of 10 and charge them all separately.

Isn't that kind of the point of cell balancing, to charge as few as possible in a group so that if a bad cell exists it doesn't bring the rest of the battery pack down with it?

Or would my method only be for narrowing down where the bad cell is? I do have the problem of finding the bad cells... the batteries are internet server battery backup units. so they pretty much sat there being recharged when they lost any charge. They may have had 2 or at most 10 discharge cycles- because they are only the backup system. However, there is something wrong with every pack... mostly the charging circuit went out for one of their BMS channels. so I'll just recycle those cells and keep the rest.

So if I charged all 50 cells in parallel I'd need to charge with a lot more than an amp or two... like 20A? just to get a descent charge rate...

Thanks for the help!

 

[fechter]

Yes, s is for series, p is for parallel. Our shorthand lingo here.

You could break up the pack into smaller groups for safety. It makes things a bit more complicated to wire up all the BMS circuits. This is something like what the Tesla roadster uses.

Putting all the parallel cells into a single group is just the simplest way to charge. Each individual cell in the parallel group should have a PTC or fuse to protect against shorts. Many laptop cells have a PTC built into the end. There are a few examples of people here combining many 18650 cells into a single parallel group (see Xyster's battery).

Yes, you'd need a big charger to do a 100Ah battery regardless of how the pack is configured. You can use multiple smaller chargers either way.

 

[GGoodrum]

The BMS has two distinct functions, low voltage protection for each cell during discharge, and it has individual circuits for each channel during charging that will bypass current when that block of parallel cells become full, so that the next block of cells can finish charging, if it is a bit behind. Without this bypass function the first block that is full would cutoff the current and you can end up with blocks that don't get a full charge. SLAs work different in that the cells have a handy built-in feature that lets them absorb some extra current, even after they get full, so all the cells can "catch up". The important concept to understand is that all the current has to go through all the cells, all the time, so if one cell limits the current, that's all the current that any of the rest of the cells can see. With this BMS bypass circuit, we are basically just simulating what an SLA cell does for Lithium-based cells, which don't have this current absorbtion capability when full.

These bypass, or "shunt" circuits don't kick in until the cells are almost full, so unless you have blocks of cells that are way out-of-balance, with respect to the rest, it doesn't take all that long for the low cells to catch up, for a typical ebike pack, as the basic BMS version allows up to 1/2A of bypass current. With a 100Ah pack size, however, even small imbalances could take a couple of hours. What you could do, however is use lower shunt resistor values (3.9 ohm instead of the standard 6.8 ohm versions...), along with the larger BD136 shunt transistors, which will double the available shunt current to 1A, but you will probably need to either add heatsinks, and/or add a fan, as this will generate a fair amount of heat.

Another option you might consider is using 10 individual DC-DC supplies like these: http://endless-sphere.com/forums/viewtopic.php?f=9&t=7392 These can be used with a standard 48V power supply, which are pretty cheap on ebay. Each one can supply 3.7V and up to about 20A.

-- Gary