DIY High Cell Count Balance Charger for Cheap

SaabGuy

100 mW
Joined
May 5, 2018
Messages
36
So i've been looking into how to charge my 20s pack for my ebike...

Heres the predicament:
-Cheap BMSs are scary, quality boards cost $$$$
-I couldn't find a simple monitor/equalizer that would bleed off high cells as well as notify me when any cell reached 3.2v (they all trip at ~2.5v for what ive found). This would be used with a low current 84v pack charger.
-Balance chargers above 10s don't really exist.
-Disassembling the battery pack to parallel charge every time would suck.

Eventually id like to build the monitor and equalizer board based off of the bestechpower equalizer boards + DW01 undervoltage protection chips.

But I've found a creative and cheap way to balance charge a 20s pack at 20A for $30. I just need 20x TP4056 1s 1A lithium charger boards. Normally this wouldn't work when the pack is still in series due to ground loops ( :flame: :flame: :flame: ), but by using 20x 5v1A AC power supplies, each charger is isolated.

Also, charging will be done through the balance plugs which are 24AWG (1.5A max) for me and will limit the charge current. I thought I could go up to 3A with two parallel cells, but most balance wires carry the current of two cells. So 1.5A is the limit for my case.

I've sourced the parts from aliexpress. If I had a thrift store nearby to source 20x usb wall chargers, i would have bought the TP4056 boards from the US and had it done in a week :D
aliexpress.PNG
 
your desire to avoid cheap BMS's is good - but then you go for cheap power supplies that will take hours to charge your pack, so you're less likely to want to sit there and monitor it through the charge? seems rather at odds with your initial conclusion of avoiding cheap stuff...

Not sure what size pack you have, but for the several hundred++ you're already spending on cells, spending an extra hundred or so on a bms over the cost of what you're suggesting a here shouldn't be a big problem, and will save you a lot of trouble. Or bulk charge and monitor your cell voltages manually with simple plug in alarms or similar, good packs rarely need much if any balancing. The other issue you'll likely have is if the charge current varies at all between these units then you'll end up with a poorly balanced pack, likely far worse than just bulk charging and just monitoring cell voltages.

SaabGuy said:
So i've been looking into how to charge my 20s pack for my ebike...

Heres the predicament:
-Cheap BMSs are scary, quality boards cost $$$$
-I couldn't find a simple monitor/equalizer that would bleed off high cells as well as notify me when any cell reached 3.2v (they all trip at ~2.5v for what ive found). This would be used with a low current 84v pack charger.
-Balance chargers above 10s don't really exist.
-Disassembling the battery pack to parallel charge every time would suck.

Eventually id like to build the monitor and equalizer board based off of the bestechpower equalizer boards + DW01 undervoltage protection chips.

But I've found a creative and cheap way to balance charge a 20s pack at 20A for $30. I just need 20x TP4056 1s 1A lithium charger boards. Normally this wouldn't work when the pack is still in series due to ground loops ( :flame: :flame: :flame: ), but by using 20x 5v1A AC power supplies, each charger is isolated.

Also, charging will be done through the balance plugs which are 24AWG (1.5A max) for me and will limit the charge current. I thought I could go up to 3A with two parallel cells, but most balance wires carry the current of two cells. So 1.5A is the limit for my case.

I've sourced the parts from aliexpress. If I had a thrift store nearby to source 20x usb wall chargers, i would have bought the TP4056 boards from the US and had it done in a week :D
aliexpress.PNG
 
Not sure what size pack you have, but for the several hundred++ you're already spending on cells, spending an extra hundred or so on a bms over the cost of what you're suggesting a here shouldn't be a big problem, and will save you a lot of trouble. Or bulk charge and monitor your cell voltages manually with simple plug in alarms or similar, good packs rarely need much if any balancing. The other issue you'll likely have is if the charge current varies at all between these units then you'll end up with a poorly balanced pack, likely far worse than just bulk charging and just monitoring cell voltages.

Its a 20Ah LiPo pack and i got the cells for $400, so your assumptions are correct. For the BMSs, the ones i found mentioned on this forum were still cheap chinese ones. I'm sure there are some out there, but I havent found them yet.

I don't have to worry about balancing when using a TP4056 on each cell; it will charge them all to 4.2V. Yes, it will take 20hrs to fully charge (20 * 20Ah cells/ 20 * 1A charge current), but im not worried about watching it all the time due to the per cell charge that's IC controlled.

I definitely understand where you're coming from, but until i find a bms with the proper over/under protection, this will do. Plus someone else might find this kind of charger useful, so why not share? hah
 
For others in the same predicament, planning ahead coming across the thread in google, or

if OP finds out this is not a good approach:

Create standalone 10S half-packs,

extend per-cell "balancing leads" JST XH type to outside the pack,

ideally each with independent BMS for protection, but not balancing.

Charge on a standard 10S RC-style "hobby" charger of desired ampacity, FMA Cellpro, iCharger, lots of excellent units out there up to 2000W, some "Duo" will do both sides at the same time.

They perfect-balance every charge cycle, much better than any cheap BMS, adjystable setpoints and all.

Connect the two in series for 20S usage.
 
That's a pretty good setup you have outlined there. those dual 10s chargers are only $350, so not bad at all.

Wish i could find a good solution to the low voltage protection during use though.
 
Cheap BMS, maybe x2 for redundancy at the cell level

Adjustable voltage LVC at the bulk / pack-level, set to trigger well above the former.

Usually triggering robust high-amp contactors.

Victron 712-BMV has one built in, get a decent SoC meter to boot, cut off based on say 85% DoD rather that unreliable voltage levels that vary too much with current
 
Are all of the typical RC / hobby based balance chargers capable to be connected in series, is that a common design trait to them? i.e. essentially a floating supply?

If so, that is indeed a neat solution. I have 2 chargers of that type actually, how can I test that they are an isolated output? Just connect them up and cross my fingers one doesn't die?

A better solution to testing that would be good. Do they internally ground their output? I mean, there must be some reference point.

The second charger would be floating at 10S, then charge to 20S. But it might be of no consequence, as that second charger might not have any knowledge of the ground at the 0S point, therefore just work in harmony.
 
jameswalker said:
Are all of the typical RC / hobby based balance chargers capable to be connected in series, is that a common design trait to them? i.e. essentially a floating supply?
No, in fact I know of none.

I clearly specified, standalone sub-packs disconnected when charging, re-serialise for use.

But there **are** plenty of bulk charger / PSUs for daily use, only use the balancing charger when needed.

Also a good solution when huge Ah capacity cells would take to long, use the high-amp bulk say 58V charger to quickly get to 4.14Vpc, 95% SoC

then the piddly 3A per cell lead hobby charger to finish up at 4.18Vpc perfectly balancing within 10-20min.


> how can I test that they are an isolated output? Just connect them up and cross my fingers one doesn't die?

Use a DMM, will let others provide a stepwise cookbook, or post a new thread if not forthcoming.

Yes not sharing any common reference is key, as is not tying to Neutral on the AC inputs?

I've seen same-voltage DCDC converters used to provide a high level of isolation, Mean Well's aren't too pricey.

But for this scenario, IMO building the packs to suit common charger configurations makes more sense.
 
I agree, building the packs for common charge methods is a good idea. Thats why I was thinking of using multiple small balance chargers at once, where you said this:

Connect the two in series for 20S usage.

Thasts why I thought, why not use the incredibly cheap and seemingly decent quality 6s chargers in series :)

Are you suggesting use a DMM to simply resistance check the output terminals with the input terminals? Or is there a better way?
 
I would think you could use two 10s chargers concurrently for one 20s pack as long as the chargers are isolated like how my idea works. Meaning, the chargers couldn't share an AC-DC converter, they'd both have to plug into the wall separately. It almost certainly would not work with the dually 10s chargers.

You'd need two connectors that split the pack into 10s sub packs and use two 10s JST balance leads as well. Same kinda thing for 5x4s packs with 4s chargers.

I also looked into a high current (20x 10A 1s charger) version of my idea and at that point you might as well just buy a 2000W dual charger. total cost was looking like $220 for 10 AC-DC PSUs and 10a 1s chargers
 
From that post
john61ct said:
Create standalone 10S half-packs
..
Charge on a standard 10S RC-style "hobby" charger

Connect the two in series for 20S usage.
By "usage" I mean discharge, as opposed to standalone sub-pack units while charging.

That really is neither costly nor difficult, compared to finding high quality + low cost balance chargers for above 10S.

And even for bulk-only, actual chargers as opposed to human-regulated power supplies, higher the voltage the rarer, more expensive, especially for power above 1000W.
 
SaabGuy said:
I would think you could use two 10s chargers concurrently for one 20s pack as long as the chargers are isolated like how my idea works. Meaning, the chargers couldn't share an AC-DC converter, they'd both have to plug into the wall separately. It almost certainly would not work with the dually 10s chargers.

You'd need two connectors that split the pack into 10s sub packs and use two 10s JST balance leads as well. Same kinda thing for 5x4s packs with 4s chargers.

I also looked into a high current (20x 10A 1s charger) version of my idea and at that point you might as well just buy a 2000W dual charger. total cost was looking like $220 for 10 AC-DC PSUs and 10a 1s chargers

it would work as you describe, however you need to ensure that there's a center tap for each charger to each 10s sub pack... ie charger 1 plugs in to cells 1-10, charger 2 into cells 11-20. NOT putting the outputs of each charger in series, and only connecting to the 20s pack + and -.

if charger 1 + output and charger 2 - output don't connect directly to cell 10+/cell 11- then the potential exists for either charger to be hit with the full 20s pack voltage, which will likely let the magic smoke out, and potentially take out the other charger in the process too. Any connector should ideally connect that center tap point first, before the pack + or -, to prevent either charger ever 'seeing' the other 10s pack. The center tap should be able to handle the full charge current too, even though it'll likely see ~0A 99% of the time.
 
SaabGuy said:
Heres the predicament:
-Cheap BMSs are scary, quality boards cost $$$$
-I couldn't find a simple monitor/equalizer that would bleed off high cells as well as notify me when any cell reached 3.2v (they all trip at ~2.5v for what ive found). This would be used with a low current 84v pack charger.
-Balance chargers above 10s don't really exist.
-Disassembling the battery pack to parallel charge every time would suck.

I don't know what you specifically mean by "cheap BMS" or "quality boards", but you might want to look into the ANT smart BMS since it seems to me that it can do pretty much everything you are asking for.

-It can bleed the cells during and also after charge so you can keep them in balance. You can also set up how close you want the balance to stay between cells (0.001V, 0.002V, 0.003V or whatever).

-It can notify you whenever a cell reaches any voltage that you set up, (it makes a beeping sound), and you can set a separate limit so that it cuts off at a slightly lower voltage than that, for complete safety.

-From what I've seen on my friends builds and mine, which are quite high power builds, it is very reliable. I haven't seen one fail yet even at 600A battery draw. It does however heat up while doing the balancing, so if you plan on doing lots of balancing then you will need it to be correctly ventilated. If cooling it is not an option due to some physical consideration (not enough space, or need it to be water tight), then you can set a lower bleeding current. In which case it takes more time to balance. In any case, if you do lots of balancing it means that something is wrong with your cells in the first place.

-The phone app is quite convenient and it includes a little screen that lets you know what is going on (power draw, pack voltage, state of charge, amp draw, individual cells voltage, temperatures, speed, odometer, etc). Very useful on a bike, you don't need a cycle analyst anymore.

So it seems to me that you might want to look into this, unless I missed some of your requirements it looks like it would be perfect for you.
 
john61ct said:
Ampacity choices are 50A/80A/100A/120A

How can it withstand 600A?

No idea where you got these choices values, but I have the 300A version.
See here: https://www.aliexpress.com/item/32826820690.html
It is marketed at 300A but you can put much higher values in the firmware. We got it working at 600A battery without any problem (slightly more than 600 actually, I forgot the exact number). This value was confirmed by both the BMS readings and the controller readings (Mobipus 72600FOC controller).
Can't tell you why it worked, just know that it did. The BMS still works without any issue to this day.
 
600A :shock: that must be on your 250 clone?

I should buy one of those, they look pretty good!
 
The link above was to model#

BMSANT8-24-ST

So many choices, vendors, hard to keep straight
 
SaabGuy said:
600A :shock: that must be on your 250 clone?

I should buy one of those, they look pretty good!

It's not on my motorbike, mine is limited by the controller to "only" 210 Amps battery.
It's on a friend's bike I've helped building.
But yes, I also have this very same BMS on my motorbike and it works really great. Already more than one year, driving every day and it's doing just fine :)

Yeah, those BMS are pretty nice and they are relatively cheap, unless you want to go for really insane amps or for high voltages they are probably the best option. Just be careful to set it up correctly, there are tons of things you can tweak in the firmware so you need to pay attention to how you set them up. It's not rocket science, but you need to know your battery cells specs well enough.
 
I did end up finally making this. The theory was right as I expected. I did add heatsinks to the TP4056 ICs just to make sure they dont current limit due to heat. Now they sit at 40-50degC

The main problem is the 4.20v +- 1.5% shut off variance (that's 4.26v max). I would have much preferred a 4.2v -3% rating on these chips. I stopped the charge when one cell got to 4.25v but next time I guess ill let it go to 4.26v and see if it trips.
 
How do you reconcile the fact that using 1 cheap 20s BMS board is scary with the fact that 20 cheap 1s charger boards is over 20 times scarier, especially since in the best case they are killing your pack much faster and result in greater pack imbalance after one day than I have had in 11 years running packs with no BMS or balance charger? :?: :?: :?:
 
I find it hard to believe that your cells are within 1.5% after repeated use and charging cycles.

By that logic we shouldn't use any complicated electronics cause they're more likely to fail.... it technically does have 20x the failure points, but it was worth it to me if every cell hit 4.2v on every charge. But yes you are right, that's not the case here and I don't like this so im debating what path I want to take.


I need to fully charge this pack one more time. The chips cut out at 4.26v and 100ma charge current, so the pack voltage will drop a bit when it's done charging. I didn't check this the last time because I was too excited to go romp in the snow. Will report back with cell voltages.
 
I bulk charge them check with cellog thru 4-6s
( you 4-5s) sensee wires if unbalance I use an an,9plus+ thru the sense wire charger for balance. But never needed 1,480 charge cycles. Only use 6 times. No bms
 
999zip999 said:
I bulk charge them check with cellog thru 4-6s
( you 4-5s) sensee wires if unbalance I use an an,9plus+ thru the sense wire charger for balance. But never needed 1,480 charge cycles. Only use 6 times. No bms

^^+1.. The “Human BMS” . :bigthumb:
bulk charge and Cellog check, Manual balance if necessary (rarely !)
This system is not infallable, it can suffer from memory fade and lazyitus !..
BUT, it has self learning abilities , multiple alarm responses, ( including swearing) , repair capeabilities, and “hard copy” data logging (via pen and paper),
It is also versatile and portable such that it can be used on multiple different battery packs. :eek: :lol: :lol:
 
The only time I worry is when I do a deep discharge.
Caught to far form home or a curious test of lvc. This is the only time I'm certain. I recharge and watch. I had two problem with signalab bms:s problems one time killing one cell in 4 parallel repalced all 4 on old ping pack. Then stolen.
Hillhater I only use the highest quality human B's BMS Even with a beer. Quality matters. Not some cheap online shit.
 
SaabGuy said:
I find it hard to believe that your cells are within 1.5% after repeated use and charging cycles.

You must be accustomed to the crappy batteries so commonly used. Get yourself some automotive grade batteries and then you'll understand. BTW my packs are many times closer to perfect balance than 1.5% .

Also, now that we know batteries last over twice as long when charged to slightly lower voltage, it's crazy to charge to 4.2V/cell every cycle...and you're planning at 4.26V, which is quite likely to kill them twice as fast as 4.2V.
 
Back
Top