Even Newer 4 to 24-cell Battery Management System (BMS)

[GGoodrum]

The CellLogs don't do anything for balancing. They are just monitoring devices. They do have LVC and HVC functions that will trip an alarm output, but these need to be isolated with an optocoupler. An optocoupled HVC signal could be used with the MW Charge Controller to keep a cell from being overcharged, but it would simply limit the current to all the cells. It does nothing for balancing.

Getting back to the full BMS, I just got some test samples of the following:

View attachment 1

These are the 3 x 8 version of the 24-Channel full BMS. There is also a 4 x 6 configuration, but I only did a test run of the 3 x 8 version. Here's what the two look like together:



They both share identical charge controller and "smart" LVC active cutoff sections. The latter function can differentiate between hard, under load, LVC "strikes", which will pull down the throttle signal, and slow drains like leaving a controller on. In the latter case, the main power will be cut. The charge controller does HVC throttling/protection and it has the low current/end-of-charge auto-shutoff function.

This is about as close to a "finished" version as we can get. The 1A shunt cell circuits have been well tested. The only difference between the two versions is basically how the connections to the cells are made. The charge controller and active cutoff sections have also come from previous, well-tested variants. The only thing we haven't tried is everything on one board.

I've gone to great lengths to make this as "one-size-fits-all" as possible. Actually, it is more like two-sizes-fits-all. The 3x8 version can be cut down to a 16-channel variant, and the 4x6 version can be made to work with 12-channel and 18-channel setups. If active cutoff is not used, this section can be trimmed off, or left unpopulated, as can the charge controller section, if a MeanWell supply, with a companion MW Charge Controller, is used. Both boards are designed to fit in a pair of Hammond extruded aluminum boxes, stuck in a side-by-side configuration, but both of these can be used like the previous v2.x versions, in a standalone fashion. In this case, however, a larger heatsink bar should be used across the shunt resistors. In the "boxed" version, there is a 3/16"x3/4" bar that goes between the shunt resistors, which are mounted from the bottom of the board, and the bottom of the case(s). This allows the whole case to be used as a heat sink.

One of these test boards is on its way to Andy, and one is going to Richard. I will build up one as well, but probably only 8 channels. I just need to verify the proper operation of the charge controller and "smart" active cutoff sections. Anyway, my hope is to put in "production" orders for both variants by the end of this week, or the first part of next week. Not sure of the cost yet, as it is totally dependent on quantities ordered. Just to be clear, I will only be offering the PCBs-only. no kits. There will, however, still be BOM Import files for Mouser and/or Digikey (they now also have a BOM import function...), to automate the parts ordering, and an illustrated set of assembly/test instructions. This is still basically a DIY application, and this shouldn't be the first printed circuit board you've put together/soldered. For those wanting an assembled version, you need to contact Andy (AndyH here...), but not until he's had a chance to do a build/test unit.

I know this is about two years late, and not something a noob might want to tackle, but the result will be a fairly capable system that can be adapted to a wide variety of configurations.

-- Gary

 

[AndyH]

Nice work Gary. Like Trappermike and others, I also intend to install this on the bike and desire a one-plug 'charge and forget' solution. I'm looking forward to tortur..I mean testing this new version!

 

[slowlane]

Resistor type overcharge protection question : Can the same value resistor be used in parallel per cell for a failsafe circuit ? (2 or more resistors per cell) Is resistor failure a concern ?

 

[trappermike]

Ah yes,further back when I was referring to the CellLogs,I was thinking of the Battery Medics which balance cells too.I still like that idea.

Are you Gary,or Andy going to publish a kind of data sheet here on the new 24s BMS,explaining it's technical data and features,and the price when you're done testing ?
I myself am looking for a unit that's ready to go,without assembly,I just kinda' wondering about the weight and size dimensions also,so I can plan for the room in the battery box on my bike,but I'll be patient while you guys get this worked out.

Once I've got my new battery pack and BMS mounted in my bike ready to go,I will need to find out how to "Break in" a new LiFePO4 pack,I think it may be somewhere here but I can't find it,so maybe someone can enlighten me on that...

 

[NickW]

One of my 2 series packs just fried its 12s BMS last night. I left it until morning to diagnose. It had melted a hole in the heatshrink. I went to pull it apart, and it started smoking and caught fire a couple times. It's a bit nerve-wracking trying to take something apart that keeps catching fire, especially when there's no apparent cause. There was no exposed metal for the board to contact, and the wires were all intact and bolted down where they belong. the board just kept smoking and lighting up.

Long story short... I'm now looking to get a 24s BMS very soon.

Any guess what the board will cost? approximate BOM cost?

The soldering isn't a problem.

-Nick

EDIT: By no apparent cause I mean that I couldn't see anything shorts or anything directly causing the fires. I think the original failure was probably caused by water getting into the pack.

 

[trappermike]

I'm no expert Nick,but in my shopping around I was told it was best to use 1 24s BMS for 2 12s packs in series. So in effect I'll have 3 separate components,2 12s packs(1 on each side of the bike)and the BMS. I'll have the BMS in 1 saddlebag with a pack and my motor controller with a pack in the other bag,and a 72v-12vdc-10a converter in one of the bags also. Then I will utilize a high quality 72v-6a fan-cooled charger.
And probably a set of 4 battery Medics to monitor charging and help final balancing.

 

[NickW]

Trappermike: Yes, that's what I'm planning to do now. The two 12s ones will work, but you need to add some large diodes to protect the BMSs. Otherwise, when one of them shuts down at lvc, it ends up with both pack's voltage across it, and will likely fry.

In this instance, I don't think either the configuration, or the BMS design itself was at fault. I've been running them that way for close to a year with no problems. They came with the packs and I couldn't justify replacing them, since everything was working well.

In the end, they just weren't sealed well enough against the almost constant rain we get here in Vancouver.

 

[NickW]

Resistor type overcharge protection question : Can the same value resistor be used in parallel per cell for a failsafe circuit ? (2 or more resistors per cell) Is resistor failure a concern ?

Paralleled resistors will act as 1/2 the resistance, so no, you can't really use them for redundancy without other supporting electronics.

Hypothetically, several higher value lower wattage resistors could be used, then if 1 failed the value would change, but less drastically.

In practice, resistors are pretty reliable compared to a lot of other parts. (notably electrolytic capacitors)
Assuming they're not being run excessively hot, I wouldn't worry about it.

 

[GGoodrum]

The boards are designed to use two 5.1 ohm/3W power resistors, per channel, in order to get a max shunt current of 1A. Should one of these fail, something I've never seen, the worst that would happen is the shunt current for that channel will go down.

-- Gary

 

[NickW]

So, in case it wasn't already obvious, I'm very interested in this, especially since my bike is down until I choose a replacement BMS.

I was wondering about the charge controller. Is this set up to just work with any appropriate voltage power supply, or does it need a feedback pin (meanwell?), or is it meant to be used with a standard CC/CV type charger?

 

[jsoar]

I built the TTpacks charge controller kit but ran into a question as I was getting ready to attach it to one of my pair of Meanwell S-350-36 PSUs. The S-350-36 pc board looks a little different from the photo in the instructions, in the area where the “2.48V Sense” wire connects to a resistor there. It even looks suspiciously like the layout of the resistor in the S-350-36 it might be reversed from the one in the photo.

So I'm wondering which end of the resistor to make the connection to. Is there any easy way to know, or will I need to do some voltage checking with a multi-meter, or need to take the board out of the Meanwell (not easy to do) and see where the pcb traces go.
Thanks for any help,
jsoar

 

[eRev]

jsoar - I'm right behind you on the S-350-36 x2 implementation - TP Packs controller pcb is on order and so too are the Meanwell units, so my ears are tweaked by your modification hoops - keep us posted if / when you get things working and I will be happy to share notes too.

BTW - the Meanwell S series is now obsolete (still on-hand in China per one ebay source found) being replaced w/ the SE series in 350-watt format - currently the SE-450 series is readily available - my gut says they "should" be similarly designed by Meanwell to where this charger concept can continue to be a viable one (old news to some I'm sure).

 

[NickW]

Back on the subject of the full BMS PCBs...

Any guess when they'll be ready so I can buy one?

I want to get my bike back on the road soonish, but I've been holding out, hoping that I can put one of these together and do it up right all at once, rather than opening up my controller and switching back to 12s in the mean time.

I don't mean to be a pest. I realize that these things always take longer than planned.

 

[Indubitably]

I too have begun to chomp at the bit. Any scraps of info that we can chew on until testing for the board is finished?

Also, is there any info on what flavors of assembled unit Andy might offer once he is ready to start taking orders? I vaguely remember hearing someone mention potting the entire thing in order to mount it on the bike without fear of damage, which sounds appealing since in my case the whole point of this venture is to provide myself with a plug-in based primary form transportation.

 

[GGoodrum]

The testing is going well. Andy should have his end wrapped up this coming week. right after that, I will do a run of the final boards.

As to versions, there will initially be two basic versions of the PCB that will be done. Both are 24-channels, but one is in a 3x8 configuration and the other will be a 4x6 setup. these will cover pack configurations of 8, 12, 16, 18 and 24-channels.

-- Gary

 

[neobeam]

after a fall off the (front x5303) bike ,at 28 mph,i built a x5304 with the dual headway 48volt 10 amp/hr.
i had a 50 mile range,when it was stolen,if anyone sees it ...its a trek navigator 100 ,thought
to be in nyc...(reward for recovery!)

got enough parts in the group buy to build a new system,and though i have a 48v bms for my cells,
i own a kelly KBL-12151. the controller will take 100 volt operation....SO,
is there a way to use this BMS for 32 of the lifepo from headway?
(i am so tired of duplicate ,high current components...)

thank you !

 

[nitters]

First post... but I've been watching this forum/thread for quite some time.

A question before you do the final run of boards; I'm hoping to build a 12s BMS based on these boards. In the available space, I'd like to stack 2 of the 6s sections on top of the charge controller section. It seems there's only 3 lines connecting each section, so it won't be too difficult to do so. However, it would be a lot easier if there were some solder circles for a 3 pin connector to each level... (a bit like the "evassemble" capacitor bms it will be replacing - I assume you've seen those)

Thanks,
Mark

 

[neobeam]

answer found,ty.
awaiting the "good news" of availability....

 

[AndyH]

A.l.m.o.s.t...

Mouser shipped my last 'test board' parts this evening - I'll have them tomorrow afternoon. The channels work well and appear as rock-solid as the 4.1.5 board I've been torturing for some time now. The control section works well on the bench as well.

Tomorrow's parts will let me finish another 4S segment that will be cycled on the bench like a 'real' BMS.

This is looking very very good!

Andy

 

[geoff57]

Hi
Gary you are making these full BMS for 6 and 8 cell packs, do all of the channels need populating or can say 5 cell packs be used by on a board designed for 6 cell packs only populate the first 5 channels of each section.
I know on the forum most favour 6 cell packs but there quite a few of us out there that use 5 cell will these boards work with 5 cell packs or do we have to wait for a 5 cell version to come out.

One other thing I have both meanwells and a different PSU not a meanwell copy / clone this does not have the same setup for adding a wire into the PSU so which is it with the BMS, is the extra wire to the PSU need like with the meanwells or have you now worked out a way to bypass the need for that wire?

Geoff

 

[fechter]

Yes, you can use any number of cells on the BMS as long as the total cell count is over 4. Just don't connect the unused channels (like the LVC boards).

The relay-based charge control circuit will work with any CC-CV charger. No interface is required.

 

[NickW]

The relay-based charge control circuit will work with any CC-CV charger. No interface is required.

I think the question is whether a CV supply is sufficient.

Does the charge controller do pwm current limiting?

 

[fechter]

I think the question is whether a CV supply is sufficient.

Does the charge controller do pwm current limiting?

No. You'll need a CC-CV supply.
PWM current limiting does not work well at higher power levels.
Virtually all switching mode power supplies can be made CC with the same little Meanwell limiter circuit, it's just a matter of locating the voltage reference input. http://endless-sphere.com/forums/viewtopic.php?f=31&t=21768

 

[neobeam]

NickW wrote:
I think the question is whether a CV supply is sufficient.

doesnt "any series resistance" that i would add,consitute a CC (or even Maximum Available current)?
and as the load was rising in voltage,the charge current would therefore reduce....
for instance,i was using a bridge rectifier,and some lamp loads in series when i was charging my
SLA pack...right off the "line".

i am aware that if the source voltage is high enough the max current thru the lamp,will be harder to calculate
for varied d.o.d.....
.but over a limited range , i have used this with good results.
once a set of times,and currents are noted,the lamps selected would dim when approaching the end of cycle.
though i havent ever done this with my lifepo,i would try it if i went to a 32s1p,
as i have no 100v+ regulated power supply

(adding and removing series limiter lamps for the "end characteristic" with a float current )


(and here we are using a 10 -20 volt max difference so ,constant current of No More Than
5 Amperes ,would require approximately , a 4 ohm series load...)


i was testing this ,using a variety of large stage lamps,and found i was even able to keep a set of
sla operating (for 3 years!) since the pulse of the line frequency(withoout ANY "smoothing" capacitance),
was probably keeping that set from getting sulfonated.
i was prepared to use a toaster ,or "iron" ,if the lamps were "too small".....lol
most of the suggested theory for this,is from old texts ,like "audels",and "elements of radio".
sections like "electricity on the farm",battery eliminators" etc...
not the best way,by far...but it will do the job!
books as old as 85 years!

 

[fechter]

NickW wrote: I think the question is whether a CV supply is sufficient.

doesnt "any series resistance" that i would add,consitute a CC (or even Maximum Available current)?
and as the load was rising in voltage,the charge current would therefore reduce....
for instance,i was using a bridge rectifier,and some lamp loads in series when i was charging my
SLA pack...right off the "line".

As long as the current at the beginning of charge (when the batteries are fully depleted) is not higher than the rating of the charger circuit, then things will be OK. The charge control circuit typically stays on until the batteries are nearly full, at which point the current should be tapered down to a fairly low level.
A big resistor in series with a CV source will do that, but might get pretty hot during the first part of the charge cycle.