New 16-cell Battery Management System (BMS)

[GGoodrum]

Can someone post a link to the latest circut schematics, parts list and board layout? I'm itchin to build one of these.

Thanks

These aren't quite ready, just yet. Richard has a few changes we're going to test, and assuming everything works, I'll then do an updated layout, and get some new boards made. At that point, I'll be in a position to start offering kits, which will include the PCB, all the parts, including connectors and harnesses, and an illustrated set of assembly/test instructions. No idea what that will cost yet, as I haven't added everything up just yet, but it will be close to whatever my true costs are. I'm not trying to get rich on these. Just don't want to lose money again. I wsa $3-4k into the last effort, when I pulled the plug.

-- Gary

 

[mcstar]

Hmm, maybe we can arrange something in the form of a public offering. I'm considering building 36V and 48V battery packs using some solid LIFEPO4's (they are similar to the lifebatts but by another company I think). With a good reasonable cost BMS we could have a nice pack available for all e-bikers. Still the cost will be more than the china duct tape packs, but these batteries have been tested to over 7000 cycles under 30A constant loads for > 10AH per cell by the DOE. Are you using Lifebatts on your bike?

 

[TylerDurden]

I'm not trying to get rich on these.

Screw that...

... get rich.

 

[fechter]

I'm not trying to get rich on these.

Screw that...

... get rich.

Hey, I get a cut...

I still need to nail down a few minor details, but it's starting to look very good.
Bulk charging at 20 amps or more should be no problem. The design is easily scaleable to larger packs/higher currents.

It should be possible to use dumb switching power supplies like the kind you can get on eBay for cheap. SLA chargers may also work if they have enough voltage. Something like a Soniel charger may need to be tweaked, but they have an adjustment pot inside. :wink:

 

[shredator]

will the bms work with a 'bad boy' charger? how much ripple can it tolerate?

 

[fechter]

will the bms work with a 'bad boy' charger? how much ripple can it tolerate?

Probably not unless the "bad boy" has some form of current limiting.

 

[Deepkimchi]

Gary told me about the adjustment pot last night on the Soneil. Mine was cranking out 5.38 A. Thought I'd back it down to 4A.

I fried mine with crazy Korean voltage (my fault). But looking at the innards, they are very well made and easily repaired.

Got mine up and running again..

 

[maxwell652000]

hi gary, im definitely interested to buy your 16 cell bms charger,i have a lifepo4 pack at 48v/20ah and can get them at wholesale prices,i have a supplier of lifepo4 batt packs ,the problem is the charger bms they sent me stopped working after only 10 uses, the mov fried along with a capacitor and temp sensor, im pretty intruiged by the solution you offer and would like to purchase a charger and bms from you if possible,,,,also just to let you know i can get crystalyte motors at cost and of course lifepo4's,,,in the meantime im back to sla's and shedding a few tears until i get to use my my lifpos again ,so i hope you can help.

 

[Patrick]

Say, regarding the "bad boy" charger - a question for the pros. I've been shopping on Ebay for a decent PS - they ain't cheap. Sure, sometimes you can get a real steal, but that's sometimes, or more like rarely. Maybe i got a 60V, 10A supply today for $35, but I bet it doesn't work (DOA warranty, or i wouldn't have). So . .

How about a Variac, with full bridge and some LARGE caps? Current limited based on the Variac (let's say 500W or 1000W), any voltage you want (within reason), all the intelligence is in the BMS. Definitely no logic to confuse. Easy to build. Easy to repair. Easy to blow stuff up if you have a kid, or are one.

What sayest thou?

 

[fechter]

How about a Variac, with full bridge and some LARGE caps? Current limited based on the Variac (let's say 500W or 1000W), any voltage you want (within reason), all the intelligence is in the BMS. Definitely no logic to confuse. Easy to build. Easy to repair. Easy to blow stuff up if you have a kid, or are one.

What sayest thou?

In theory it would work OK. You might not even need the caps. You would want a GFCI, a fuse and an ammeter. Those aren't isolated usually, so you could get zapped easy.

Switching power supplies off eBay are getting pretty cheap these days.

 

[GGoodrum]

Progress...

We've made some significant progress, in the last couple of days, and I just submitted what I think is the "final" version of the layout, to have boards made. The changes we have from this point forward will most likely just be some resistor value tweaking, and/or maybe a change to a slightly higher current rated NPN transistor.

Here's what the boards look like:

What is different from the first layout is that Richard added a gate driver chip to square up the pulses going to FET gate, and we got rid of 2nd TC54 chip that we were using to to detect when the shunt was fully on. Richard figured out a clever way to do this without the TC54 that actually works better.

Basically, the way the circuit works is that once a cell's voltage reaches about 3.65-3.70V, the LM431 programmable zener starts to conduct, which then drives a transistor/power resistor shunt. When the shunt first starts to bypass current, the green portion of a bicolor LED starts to come on. When the shunt is close to maxing out (about 150-200mA...), the red portion of the same LED comes on. This makes the color a yellowish-orange hue. Wed the red portion comes on, so does the opto output for this channel, which then pulls down the FET gate, via the gate driver, which then cuts the charger current. The cell voltage starts to drop when this happens, which turns off the opto, turning the FET back on. This oscillation continues, just keeping the high cells at the cutoff, while the rest catch up. Once all the cells have their shunts working (i.e. -- all the LEDs are on and "yellowish"...), the current will start to drop. When a cell can't accept anymore current than what the shunt can handle, its red portion of the LED and its opto go off, leaving just the green portion of the LED lit. Once all the LEDs are green, the cells are about as full as they are going to get.

Richard is still playing with voltages and the bypass current, which is why there might be a couple of tweaks, here and there. We want to make this as bulletproof as possible. Once the tweaks are all done, I will post the schematic. At that point, the kits should be ready to go as well. I've already got all the parts.

-- Gary

 

[Deepkimchi]

I still can't understand why you put so much effort into something to sell and then post the schematic. Why not just PM it to the few individuals that might have a valid interest (and actually understand).

If someone buys the board, like me then you could include the schematic with a description of how it works, critical voltage points, as an "owners manual".

I really think you should protect your investment and hard work.

 

[GGoodrum]

I still can't understand why you put so much effort into something to sell and then post the schematic. Why not just PM it to the few individuals that might have a valid interest (and actually understand).

If someone buys the board, like me then you could include the schematic with a description of how it works, critical voltage points, as an "owners manual".

I really think you should protect your investment and hard work.

This all started as a sort of "group effort", holy grail sort of quest to do a cost-effective, barebones-type of BMS, one that would do two basic things, and that's it: low-voltage protection, and some sort of charging management, in order to allow each cell to get as full as it can. This effort has also served as a learning process, for many, including myself, so the philosophy has always been to keep the details as public as possible, which is why this thread is a "sticky".

I can tell you, that just having a schematic is not enough. The real effort, for me at least, has been what is required to do the board layouts, and then gtting the boards made. If somebody really wants to do their own board design for one of these, I say have at it. Doing small quantities (2-4...) means they will cost $40-$50 each, or more. Same thing with the parts. Buying small quantities (less than 100-500...) adds as much as 30-40% to the cost. When we do get the "final" version done, I will get at least 100 boards made, which gets the per board cost to a minimum. As I said, I've already got all the parts. most of them I ordered in quantities in the thousands, in order to get the best price possible. This keeps the total costs down as low as possible, so we can then offer these kits as cheap as possible, and still make a few dollars that can then go towards recouping my early investment. Richard also gets a few ducats, to offset all the effort he's put into this, which is a lot.

-- Gary

 

[Ypedal]

And i commend you both on that !!! THUMBS UP GUYS !!!!!!!

there is more to be gained by opening it up to public scrutiny, guys like me simply don't understand the electronics well enough to help, but will buy from you to support the effort.. Heck, once in a while , as we have seen happen already, someone with more knowledge happens to drop in and provide a boost of knowledge that helps the project improve fruther.. win win..

Keep up the great work ! Should i take a number ?

 

[rkosiorek]

for example "randomly" raised some interesting ideas and questions earlier in this thread.

i wonder what ever happened to him?

rick

 

[maxwell652000]

can someone let me know if i purchase a soneil 4808 SRF charger can i charge my 48v/20ah lifepo4 pack (16cells)without a bms ,i could manually check the voltages for each cell to make sure they stay balanced until gary has the bms ready to offer us mere mortals,,or is there another charger i should be considering? i know the soneil 4808srf will remove loose sulfation for sla's and i wonder if that represents a problem for lifepo4's? btw how does one go about purchasing a bms from gary when it becomes available?does he have a sight?

 

[GGoodrum]

for example "randomly" raised some interesting ideas and questions earlier in this thread.

i wonder what ever happened to him?

rick

Yes, he did, and we are using a similar idea to one he first proposed which is to keep the cells right at the threshold, by pulsing the charge current. The way we are doing it is pretty different, but the same idea, nonetheless.

I think he's around, as he posted recently in one of the threads (PIC BMS?), but I think he just has a limited amount of time to participate here.

-- Gary

 

[GGoodrum]

can someone let me know if i purchase a soneil 4808 SRF charger can i charge my 48v/20ah lifepo4 pack (16cells)without a bms ,i could manually check the voltages for each cell to make sure they stay balanced until gary has the bms ready to offer us mere mortals,,or is there another charger i should be considering? i know the soneil 4808srf will remove loose sulfation for sla's and i wonder if that represents a problem for lifepo4's? btw how does one go about purchasing a bms from gary when it becomes available?does he have a sight?

Yep, this works fine with LiFePO4 packs. I use one all the time to charge both a123 and LiFeBatt-based packs, even without a BMS. What it does is it will overshoot the cutoff, by about 1.25V, reset itself, reduce the charge current to 1.0A and then hold the voltage at the cutoff. That will cause the current to slowly drop down. During this phase, the green LED is on. What I do is use a WattsUp meter, so I can monitor the current. When it gets a bit under 100mA, I shut off the charger, as the cells are about as full as they are going to get.

-- Gary

 

[Randomly]

for example "randomly" raised some interesting ideas and questions earlier in this thread.

i wonder what ever happened to him?

rick

I drop by occasionally and look around, but as gary mentioned I'm currently spending all my time working on my zero-point energy super suit and my plans for world domination. Finding good minions is so hard these days....

 

[docnjoj]

U gotta do a better job than whats going on now in the world!
otherDoc

 

[maxwell652000]

thanks Gary for that informative response,i justed checked all 16 individual cells in my pack most are between 3.23v to 3.25v but one particular cell is only 2.83v (this is after being discharged on a ride and since apx 2weeks ago my charger fried and i have not been able to recharge )so what can i do to charge that one individual cell to balance it with the rest of the pack? im guessing to order a single cell charger from voltphreak,,,https://s.p10.hostingprod.com/@www.voltphreaks.com/ssl/catalog/product_info.php?cPath=3&products_id=28 ,,,hopefully that cell will come back to match the rest of the pack,,,any input will be appreciated

 

[PJD]

maxwell,

You might also consider a cheap bench top power supply to charge the individual cell. (try here: http://www.mpja.com/products.asp?dept=2&main=1 ) then you will have something that is useful for other things too.

Gary,

I understand that your proposed balancer board will cut the charging current down to 150 to 200ma when the first cell reached 3.65 to 3.7 volts. However, since most SLA battery chargers have a "charge-complete" shutoff current setting of 300 to 400 ma, will this cause the charger to prematurely shut off? Thanks.

 

[GGoodrum]

maxwell,

You might also consider a cheap bench top power supply to charge the individual cell. (try here: http://www.mpja.com/products.asp?dept=2&main=1 ) then you will have something that is useful for other things too.

I'd agree with this, but the VoltPhreaks 2A charger would workfine as well. The bigger issue is why one cell got so far out of balance. You may have a bad cell, or one that is pretty severely damaged.

Gary,

I understand that your proposed balancer board will cut the charging current down to 150 to 200ma when the first cell reached 3.65 to 3.7 volts. However, since most SLA battery chargers have a "charge-complete" shutoff current setting of 300 to 400 ma, will this cause the charger to prematurely shut off? Thanks.

I don't think so, as long as the charger cutoff voltage is set slightly higher than the sum of each channel's cutoff. When a cell first gets to the cutoff (i.e.--around 3.65-3.70V...), the charge current is still up there. The shunt then starts to conduct, bypassing a small amount of current, but the cell is still absorbing most of it. The shunt current will start to climb, and the cell current starts to drop. Somewhere just befoe the shunt reaches its 150-200mA limit, the rest of the circuit will "interrupt" the charge current, via the FET. This will cause the cell voltag to drop slightly, resetting the FET circuit, which turns the charge current back on. When it comes back on, it will be at a higher current. Basically what hapens is that the cell gets hit with series of higher current pulses until it gets full enough that the shunt can handle bypassing whatever current the cell can't absorb, and not get swamped. When that happens, the FET control logic for that channel won't try and turn off the FET. When that happens, this portion of the circuit won't turn on the red portion of the LED either, so the LED will be a full green. Once the last channel reaches this condition, the FET logic will stop trying to pulse a higher current, and the total current passing through the pack will then be in the 150-200mA range, If the charger shuts off at this point, that's okay. The worst that should happen is that the charger shuts everything off, just as the last LED turns green.

I can tell you, however, that for my Sonewil SRF and for my Zivan NG1 chargers, they don't automatically shutoff. the seem to want to stay in some sort of trickle-mode.

-- Gary

 

[PJD]

Gary,

what my made-for-the-scooter mfg. Chinese chargers do is, when the charge current in the CV stage has reached it's "charge finished" setting, the charger shuts off. Then it waits for the excess voltage to decline from the 59.5 volt CV set-point to about 55 volts, then is switches to float maintenance (trickle) charging holding the voltage at about 55 to 55.5 volts.

For SLA, this "wait" is from a few seconds to sometimes 30 seconds. Since the lifePO4's hold their surface charge voltage much, much longer, I've never seen it switch to float. I'm not sure if it would hurt the LiFePO4 pack if it did.

The current setting is adjustable, via a pot on the control board of the charger, but I think the the lowest it can be adjusted down to is about 250 mv.

 

[GGoodrum]

Even still, it shouldn't matter. It won't get to that point until the last cell stops having the FET pulse current, which is going to be higher than 300-400mA. It won't "see" under 200mA unil the last cell's shunt current is lower than the point that the FET logic kicks in.