BMS sizing? SmartBMS quality? Group Buy?

I've built a pack to replace my SLAs with some of my UAV Lipos.. (at least untill I get around to building an 18650 pack).
Looking at adding a BMS for piece of mind and reducing my work to checking it every now and then instead of every charge to and ensure per-cell low-voltage protection incase someone turns on my bike at work.

I
My question is about the sizing in A for the continuous discharge. I known its good to be very conservative on C ratings.. but not sure about BMS.. is the Amp rating adding protection to my controller or should I over size it?

My pack is 7s 12ah with a 19C rating for the cells. My bike has a 600W motor @24v nominal.. but I'll will normally be charging to 28.3 most of the time and 29.4 when I feel I need added range and/or to force balancing. My hills are steap and I'll suck every watt going to work.. so should I be looking at at 30A BMS or should I go for the 60A for a few dollars more?


I'm looking at Supower BMS. not too many choices at 7S. I dug around the forms and did not see any reported problems with them.. Anything I missed?
Updated: based on feedback I started looking st bmsbattery's (ecitypower) smart BMS for $25 + shipping.. interested in feedback on that.. and maybe a group buy of them...

no problems reported because nobody uses them. 7S is so small i don't know of anyone who builds to 7S.
you can use a regular 13S model of many lipo BMS and just use the first 7 channels.

I see a bit of a problem from using two different charge voltages. A 7s bms is going to be expecting 4.2v per cell, but your talking about 4.05v per cell charge levels. There is a chance that the bms won't initiate balancing at such a low voltage. The actual board in question would need it's specifications checking to see what the balancing threshold is.

If you buy the 40A $25 smart bms you can have the parameters set however you want, as it uses a microprocessor. You could request a balance threshold of perhaps 3.95v and an emergency cutout (hvc) of 4.28v. Then you could use either charge voltage. I'm unsure what the factory threshold is if you don't specify one. Which is the basis of my concern.

dnmun knows a lot about the bestech bms boards. dnmun, do you know their threshold?

i just did a special order and got them to build a buncha D131 to 4.18V balancing voltage. they still use the normal 2.9-4.28V LVC/HVC.

i wanted to have them balance a little lower because i have this 21S lipo pack in parallel with a 24S lifepo4 pack so wanted the 4.18 so i did not overcharge the lifepo4 too much when i allowed them to charge long enuff to go into balancing.

i am not one of these people who believes that charging to 4.2V is bad either.

i do think it is bad to charge to full voltage and leave the pack charged so i usually allow balancing once in a while and just charge up close to full charge and unplug the charger before it goes to full charge. then i use it up right away.

i have one channel on my big lifepo4 pack that self discharges more than the others so i do have to balance regularly though so that is why i ordered these new D131 built to the 4.18V.

friendly1uk said:

I see a bit of a problem from using two different charge voltages. A 7s bms is going to be expecting 4.2v per cell, but your talking about 4.05v per cell charge levels. There is a chance that the bms won't initiate balancing at such a low voltage. The actual board in question would need it's specifications checking to see what the balancing threshold is.

If you buy the 40A $25 smart bms you can have the parameters set however you want, as it uses a microprocessor. You could request a balance threshold of perhaps 3.95v and an emergency cutout (hvc) of 4.28v. Then you could use either charge voltage. I'm unsure what the factory threshold is if you don't specify one. Which is the basis of my concern.

dnmun knows a lot about the bestech bms boards. dnmun, do you know their threshold?

Click to expand...

Thanks.. the lower voltage is for my daily use.. If I charge to 4.05... commute to work and back (about 5miles with my normal route) I arrive home at 3.9v.. i.e. I don't use much of the pack on a normal day. Given how little I use I wanted to keep the voltage lower when I bulk charge as it increases pack life to stay lower. I want the BMS to protect individual cells incase they are out of balance and go to 4.2 while some runt cells are at 3.8 or something, and also to protect for discharge if someone turns/leaves it on or if I take much longer trips.

I plan to charge to an overall 4.2 every now and then to force full back balance. I'll check them by hand and if I see the gap growing I'll force the pack to balance by using a higher charge voltage. Will also do that on days when I need extended range for longer trips.

After reviewing the concepts I decided going for 4.2 on the overvolatage protection and balance is fine.. if I don't hit it the cells are safe and if I do it provides protection. But allowing 4.2 when I want it allows for increased range on those days I want/need it. (Though its not much added range). A programable one would be cool, but I thought they were too much $$ for the value.

Is this the $25 smart BMS you are suggesting?
http://www.bmsbattery.com/smart/330-lifepo4lithium-ion-smart-bms-for-513-cells-in-series.html
how is their reliability? Shipping is very high ($33 -- more than the unit) and shipping back to china for problems is painful/expensive. Are there any US sellers?

I saw that but when they said you need to pay $199 (or buy 1000) to get the programer did not think it relevant.. Has anyone hacked their own reprogrammer?

Thanks dnmun. That was enlightening.

Dr.I If you set both hvc and balance to 4.2v you can't charge and balance at the same time. It works better if there is a gap between them. Then as one cell reach's 4.2v first, the balance circuits can start discharging it while the rest of cells can carry on charging. The perfect situation is where balance currants and charge currants are also well balanced to keep it all ticking over nicely. 4.28v is commonly claimed as max voltage by manufacturers, Which is a reasonable hvc level when your target is 4.2v per cell.


The Smart is the one you linked to, and the computer stuff generally on offer at half price. The chip is the oz890 and has it's own thread, all about hacking. If I wanted a one-off custom version I would talk to crossfire here on es. He likely has one acquired as part of a larger order, and can set them however you please. You will see him on the oz890 thread.

Well looked at the smart BMS from bmsbattery.. interesting but the quality of ecitypower seems questionable (many problems). (http://endless-sphere.com/forums/viewtopic.php?f=14&t=54156&hilit=smartBMS and http://endless-sphere.com/forums/viewtopic.php?f=14&t=20971)
(And frendly1uk.. did you mean crossbreak? I contacted him and is no longer working on them... doing his own board).

So I updated the thread title .. for those using SmartBMS would you buy it again if you had to use it out of the box without the programmer?
How functional are the Opensource programming tools like the atmega http://www.pedelecforum.de/forum/index.php?threads/diy-smartbms-reader-mit-atmega.22585/
(I'd perfer command line, but it was not really clear if it can change all I need to reprogram/calibrate?)

Anyone interest in a group buy of $25 SmartBMS to help reduce the crazy shipping costs (seems even if we order 5 the shipping is the same $30).. I could handle reshipping in the states?
(Not sure if this is allowed here or if I need to repost in the new-sales section.. )

the quality of the ecitypower products is not in question. some people complain about the service at BMSbattery but the EMC chargers are pretty good chargers for the money.

nobody is gonna pursue buying the smart BMS because of the need to program it. they are made for battery manufacturers to use and in that case the engineers at the battery manufacturer would write up the code. UK guy just figured out how to do it on his own. he and crossbk were able to make a programming cable i think so that was how they got there.

it is just easier to buy a BMS already set up since the BMS makers already have a better handle on the correct charging parameters.

don't do it. There are several more expensive BMS'es, that are working. This "SmartBMS" does not work as expected.

Oh yes, crossfire was an 80s tabletop game. I still have it in fact

I like the smart bms. I must of had half dozen of them. Programmed by suppliers. I have heard the concerns of others, but they just seemed like user error. Setting the charger and hvc to the same voltage being quite a common one. Leading to complaints about not balancing once hvc is reached. Which they shouldn't. I think that was your complaint wasn't it circuit?

friendly1uk said:

Oh yes, crossfire was an 80s tabletop game. I still have it in fact

I like the smart bms. I must of had half dozen of them. Programmed by suppliers. I have heard the concerns of others, but they just seemed like user error. Setting the charger and hvc to the same voltage being quite a common one. Leading to complaints about not balancing once hvc is reached. Which they shouldn't. I think that was your complaint wasn't it circuit?

Click to expand...

When you don't know much about BMS, it looks all good to you. It may look OK when all cells are in perfect balance and no balancing is actually needed. If imbalance develops, lower charger voltage will not help, you will still hit HVC and end up in "user error" condition. The less you know, the better it looks.
That bug was brought up by several members on several forums, not just me.

circuit said:

If imbalance develops, lower charger voltage will not help, you will still hit HVC

Click to expand...

Yes, if you specify the hvc the same as charge voltage this will happen, and many have done it. Then lowering charger voltage is too late, as the cells that are ahead are still going to hit hvc if the condition got out of hand enough for someone to actually notice.

I have set them the same and watched the behaviour you speak of. hvc is reached every time, just like I had specified it to do by asking for the numbers to match. Then the unit shuts off. Just as it should do in a fault condition. Your actual gripe is that you want it to carry on working in this hvc state. That isn't right though. hvc should never be reached to begin with. The board can start balancing from any voltage you specify and it's not a bad balance current. There is no reason to over shoot as far as hvc unless you set the hvc too low or you have a fault worthy of shutting the charge process off completely. I have seen hvc specified as battery voltage enough times to of initially thought it how to set them up myself. I had to step back and think about how I would do it, then it all makes sense.

friendly1uk said:

Oh yes, crossfire was an 80s tabletop game. I still have it in fact

I like the smart bms. I must of had half dozen of them. Programmed by suppliers. I have heard the concerns of others, but they just seemed like user error. Setting the charger and hvc to the same voltage being quite a common one. Leading to complaints about not balancing once hvc is reached. Which they shouldn't. I think that was your complaint wasn't it circuit?

Click to expand...

Your signature suggests that bmsbattery.. who is the supplier of the smartBMS.. sent you broken stuff... as I recall from other threads and your suggestion above you had to find someone with a programmer to correct it.. is that true?

When one orders on bms battery one specifies
How many cells in series(5-13)*
Over voltage protection*
Under voltage protection*
Cells type(LiFePO4,Lipo)

But does not specify the balance.. is that what leads to problems they ship with a bad config?

Yes, bmsb are a lottery, and have sent a board that got broke in transit. Other boards not finished to the manufacturers spec but working, and one I couldn't get to work. I sent that to crossbreak and he changed it a bit and thought it seemed fine. 12S though or I would send it you. bmsb are a proper lottery and I would certainly buy two. Then pass one on, if both are fine.

Yeah... reading that you do have to wonder 'why bother' but they are still the cheapest generally. You won't get another processor controlled board for this money, but some analogue one's might come in a bit cheaper from aliexpress. Who are another shit storm in a shit pan of shit providers.


You don't specify a balance voltage. If you say lipo, they will decide from what voltage onwards that balancing will occur. Probably around 4v as lipo is usually charged to 4.2v. It does not need to know about the 4.2v though, that is the chargers job.

What will happen is that above 4v the board will try to keep all cells level. Your charger can be set anywhere from 4.01v per cell to 4.2v per cell and it won't matter. It just needs the charger to be over the balance threshold and under your hvc.

If you look at the oz890 data sheet you can see a whole heap of settings. bmsb are just asking the basics. I bet you could request more outlandish value's if you wanted though.

If crossbreak won't set it up for you, then who is the guy in the UK? I can send him a board to change to 7s for you, but it would need another round of testing. I found it to be broke, then crossbreak says it works. A third opinion would be good. I also have a 13s board I blew the fets on. It could be fixed with the fets from this 12s board. Both are up for grabs, if you can find your man in the UK. I'm in the UK so getting the boards to him should be the cheap bit of all this.

I do like the core design of the Oz890-based systems.. With OZ890 smartBMS, I can actually watch the voltage easily (I'll build a bluetooth interface ala
http://www.pedelecforum.de/forum/index.php?threads/diy-smartbms-reader-mit-atmega.22585/ and then can check it while charging every day).
Plus if I switch to a different pack/chem I can reprogram it for the next pack (e.g. lifepo4) or if I go 18650 i can continue with it as originally programmed. still a bit worried about quality but most of what I saw was about not balancing. I'm really looking to buy a BMS to ensure the battery is not drained and to ensure no cell goes above 4.2 during bulk charging. I'll probably keep my pack between 3.7 and 4.05 volts at almost all time so 4.2 is way high for me. And I'm not terribly worried about balance (what seems to be the most common design/software issue). I can balance by hand if needed (and if I'm monitoring daily I'll know). I've not had to do hand balancing on the bike since I started using my UAV Lipos -- 2 weeks and the max difference is .02v which is about what it was when I started.

Luckily, I may have found a way to program the smartBMS just incase I cannot get the existing software talking to it with one of my existing i2c adapters.. a german site
http://www.pedelecforum.de/forum/index.php?threads/smart-bms-5-13s-programmierbar-meine-ersten-erfahrungen.9396/page-18
pointed to software from russians developers (including an linux iSO so hopefully self-contained package) and so I'm feeling a bit better that even if its its programmed wrong from BMSbattery I might be able to fix it even if I cannot get the BMSbattery software working with my i2c devices..

Anyone here tried the russian ebike_kit.iso ?

Would you like the 13s with broken fets then? You can practice upon it. I think I can have it there next week for $5
I could send the 'should be' working 12s board, but it puts me in a position. Crossbreak didn't really do anything to repair it. Just tweeked the software first, then it tested ok. I know it was broke though. leaving me a little hesitant to warranty it.

Go on, how about the price of a board for a board. If it don't work you have a second that should if you move the fets around. So to recap, that's $25 for 2 boards, delivered to you, next week.

friendly1uk said:

Would you like the 13s with broken fets then? You can practice upon it. I think I can have it there next week for $5
I could send the 'should be' working 12s board, but it puts me in a position. Crossbreak didn't really do anything to repair it. Just tweeked the software first, then it tested ok. I know it was broke though. leaving me a little hesitant to warranty it.

Go on, how about the price of a board for a board. If it don't work you have a second that should if you move the fets around. So to recap, that's $25 for 2 boards, delivered to you, next week.

Click to expand...

Thanks for the offer.. Debugging unknown software on potentially broken hardware is not generally a good strategy. Later in the week I'll play with the software and get a feel for what I might be able to do with it. I already tried a simple test of changing stuff using the online webpage and it generates errors when saving and when restoring from that saved file its not the same.. so it may not be a viable solution for me without digging deep into the documentation/code..

Wouldn't your 24v motor be grateful getting 36? 10s is going to be a lot more common. Though a bestech bms can usually be under populated. So ordering a 10s and then using it 7s is acceptable.

friendly1uk said:

Wouldn't your 24v motor be grateful getting 36? 10s is going to be a lot more common. Though a bestech bms can usually be under populated. So ordering a 10s and then using it 7s is acceptable.

Click to expand...

Nope.. tested it and the controller shuts down at 30.5 .. and I don't really want to get too close to that so 7s is all I'm willing to run. 36v might be nice but not on this old bike (its a 2006-ish LashOut with 600w motor).

But on reflection a test board at $5 might be interesting, there is not much to loose. Tell me what happened to break the FETs..
Does the chip talk i2c to the software (i.e. is this a board you sent to crossbreak and he said the fets were fried but the board was okay?) if all that happened was fried FET's I can easily replace those..

I'm in the same boat as you, with somewhat same goals (want to extend life by not charging fully to 4.2, conservative LVC, balancing, etc), and considered pulling the trigger on the SmartBMS from BSMB just last night, but didn't because I was worried the general basic parameters you enter on their form might not get me what i wanted since they were non-standard, and then I'd be stuck with an ill-programmed BMS. Sounds like you might be onto a solution for the programming part with the Russian ISO boot environment to program over I2C.

Anyway, me know if we can gang up resrouces to save on shipping/whatnot if you go that route.

Thanks,

-=dave