Bottom Balancing? Top Balancing? Charging and BMS

[V-spec]

Hi All

Need some advice for my first EV electric motorbike, the main concern for me is how to build a good reliable battery pack.
So far its been difficult to find all the info i need so i am hoping someone with experience can help.

I currently have 24 Calb 40ah lifep04 cells. They are just sitting there doing nothing until i figure out what to do with them, i want a suitable BMS and charging system.

I need to get 48-50v from the pack, i just want to know how to start setting them up, can i use all the batteries to do this? it would be a mixture of putting them in series and parrallel.

I was thinking of buying this:

http://www.ebay.co.uk/itm/iMAX-B6-AC-B6AC-Lipo-NiMH-Nicad-LiFe-Battery-Balance-Charger-/170913517290?pt=UK_ToysGames_RadioControlled_JN&hash=item27cb3d4eea

or similar?

I also have:

http://www.cleanpowerauto.com/MiniBMS.html

All i need now is a charger and instructions to get it going.

Thanks

V

 

[dnmun]

do you already have the motor and the el moto frame assembled? with 24 cells you would either build a 12S or 24S pack in order to use the cells. why did you think it would be 16S?

rather than using the little short circuit thingies, you could use a the headway 24S BMS and it will provide up to 100A of juice which is about all those cells will make in any case. jmho.

or you could bypass the BMS on discharge and use an LED on the console to signal LVC.

 

[Jeremy Harris]

There's no such thing as "Bottom Balancing", I'm afraid, it's a myth perpetuated by some bloke who makes internet videos and doesn't understand what he's talking about (can't recall his name, but no doubt someone here will remind me).

All cells vary in capacity to some degree, due to manufacturing tolerances and things like slight differences in temperature from cell to cell in the pack. For longest life the depth of discharge needs to be minimised - if you never completely discharge cells they will have a longer cycle life.

Unfortunately, the mythical "bottom balancing" process involves discharging the pack down to near zero remaining capacity, then putting exactly the same charge into each cell. This is pretty fraught, because apart from deep discharge being an unwise thing to do, each cell will have a slightly different charge acceptance efficiency. The result is that this process doesn't balance the cells in a pack at all, all it does it potentially reduce their cycle life.

Balancing cells in a pack is all about maximising the usable capacity and minimising the risk of premature cell failure. The only way to do this is to charge every cell until it reaches the selected ful charge terminal voltage. The time taken for each cell to reach full charge will vary, so the BMS has to allow the first cells to reach full charge to stop accepting charge (usually by shunting the current around each fully charged cell) and yet still allow the other cells in the pack to keep on charging.

The MiniBMS you linked to looks as if it may be OK. Not sure about the idea of using an iMax, as you can only charge a max of 6 cells in series with that. The Headway BMS dnmun mentioned would also be a good choice, and probably easier to use.

If you want to maintain your cells in good condition in the mean time, then you could connect them all up as one big bank in parallel and hook them up to a constant voltage, current limited, power supply, set to around 3.5V (you can charge them to 3.65V, but 3.5V is better for a maintenance/storage charge). Before doing this, it would be a good idea to check the voltage of each cell to make sure they are all still OK (they should be at around 3.2V or so each).

 

[arkmundi]

Balancing cells in a pack is all about maximising the usable capacity and minimising the risk of premature cell failure. The only way to do this is to charge every cell until it reaches the selected ful charge terminal voltage. The time taken for each cell to reach full charge will vary, so the BMS has to allow the first cells to reach full charge to stop accepting charge (usually by shunting the current around each fully charged cell) and yet still allow the other cells in the pack to keep on charging.

Thanks for a simple statement on what it took me a while to figure out. So, I still have as yet to accept a BMS for my pack, having initially choosen the "manual" method. I purposely built a pack with excess capacity - 20ah - of which I use between 25% and 50%. Hence a BMS during ride time I feel is unnecessary. Controlling charge in such a way as to achieve cell balance, however, is another thing. I currently use a batteryspace smart charger to bulk charge all 12 cells. I then periodially (once every 4th or 5th ride) test the voltage on each individual cell. Every now & again, there may be a low voltage cell, one that is more than 0.1v to 0.2v from the others. I then single cell charge it, bringing it back in line. This method is working I believe. What do you think? Thanks.

 

[Jeremy Harris]

It's probably good enough, TBH.

I don't use a BMS on any of my ebike packs and just periodically unhook the cells and charge them on an RC balance charger as smaller sub-packs. On the latest 15S pack I've used a plug-in link to automatically switch the three 5S sub-packs from series to parallel. This means I can charge a 15S pack using a 5S balance charger at every charge, which I'm hoping will work OK in the long term.

 

[V-spec]

do you already have the motor and the el moto frame assembled? with 24 cells you would either build a 12S or 24S pack in order to use the cells. why did you think it would be 16S?

rather than using the little short circuit thingies, you could use a the headway 24S BMS and it will provide up to 100A of juice which is about all those cells will make in any case. jmho.

or you could bypass the BMS on discharge and use an LED on the console to signal LVC.

Hi Dnmun, i just bought a chassis, building a cafe racer type bike. I also have two lynch motors which are rated at 24volts (30 volts max) 4000rpm and 200amps peak, again i have to find a suitable controller. Cedric Lynch advised me to use the motors in series needing 48volts. Its about 150rpm per volt.

I originally bought 24cells to power one motor, then decided to do a dual motor project. I'm thinking i should of just bought one 95r agni motor. I was hoping to get 150-200amps out of these batteries!

 

[V-spec]

There's no such thing as "Bottom Balancing", I'm afraid, it's a myth perpetuated by some bloke who makes internet videos and doesn't understand what he's talking about (can't recall his name, but no doubt someone here will remind me).

All cells vary in capacity to some degree, due to manufacturing tolerances and things like slight differences in temperature from cell to cell in the pack. For longest life the depth of discharge needs to be minimised - if you never completely discharge cells they will have a longer cycle life.

Unfortunately, the mythical "bottom balancing" process involves discharging the pack down to near zero remaining capacity, then putting exactly the same charge into each cell. This is pretty fraught, because apart from deep discharge being an unwise thing to do, each cell will have a slightly different charge acceptance efficiency. The result is that this process doesn't balance the cells in a pack at all, all it does it potentially reduce their cycle life.

Balancing cells in a pack is all about maximising the usable capacity and minimising the risk of premature cell failure. The only way to do this is to charge every cell until it reaches the selected ful charge terminal voltage. The time taken for each cell to reach full charge will vary, so the BMS has to allow the first cells to reach full charge to stop accepting charge (usually by shunting the current around each fully charged cell) and yet still allow the other cells in the pack to keep on charging.

The MiniBMS you linked to looks as if it may be OK. Not sure about the idea of using an iMax, as you can only charge a max of 6 cells in series with that. The Headway BMS dnmun mentioned would also be a good choice, and probably easier to use.

If you want to maintain your cells in good condition in the mean time, then you could connect them all up as one big bank in parallel and hook them up to a constant voltage, current limited, power supply, set to around 3.5V (you can charge them to 3.65V, but 3.5V is better for a maintenance/storage charge). Before doing this, it would be a good idea to check the voltage of each cell to make sure they are all still OK (they should be at around 3.2V or so each).

Thanks Jeremy that sound like an obvious solution, so i can put all the batteries in parallel, therefore all the voltages will (in theory) balance and charge themselves, and i dont need a BMS? Ill be sure to put a multimeter across each battery to check there voltages. Ill also check out the headway BMS system.

BMS seems a headache free way to go, ill keep you posted on how i do!

 

[dogman dan]

Jeremy is the guy to learn from here. I just try to put what he says in a simpler to understand form for some that need it mostly.

But a few points didn't get covered yet.

An Imax B6 can be a handy little multipurpose charger for small batteries. But it will be slow, REALLY slow at 50w. 67 hours to charge your 24s 40 ah pack. So clearly you will need a faster bulk charger later, once you have definitely settled on a voltage. But once you have bulk charged your pack, the b6 might still be usefull for adding that tiny bit more charge needed on one or two cells to get completely balanced.

What I mean is, you could bulk charge, then use the imax in single cell mode to bring up the lowest cells one by one.

As pointed out above, you have an odd number of cells for a 48v system. A 48v lifepo4 pack typically has 16 cells. So you'd need more batteries to make a 80 ah 48v pack, or you will have some extras left if you make a 40 ah 48v pack. This could end up being a good thing. If you find yourself always having to balance cell 5 for example, swap in a different one. Eventually you should have a pack sorted that need very little balancing when you don't discharge deep. Then you might sell off the "different but still good" leftover cells.

When you do discharge 100% or close to it, that is when you can expect to need to balance the pack.

For the higher amp applications, I tend to agree that a "human bms" can be the way to go. Get two cellog 8's, and you can have a readout of each cells voltage. Then a Cycleanalyst in the higher power version will make a nice display of full pack voltage, and all the other wattmeter functions.

 

[dogman dan]

Some will NOT recomend this company. I personally have not done buisness with them, so I can't say thumbs up or down myself.

But something like this might be a good affordable starting point for your charger, particularly for the bulk charges when balancing is not needed. A little over 2 hrs to charge 40 ah of 48v.
http://www.bmsbattery.com/alloy-shell/457-alloy-shell-900w-lifepo4li-ionlead-acid-battery-ev-charger.html

 

[V-spec]

Sounds like a plan Dogman, i want to be able to go at least 40miles per charge and want a fast charger, those cellogs look ideal and very handy for me i think ill order a couple to help keep an eye on the cells.

i was thinking i would put 15 cells in series to make 54volts (if each cell was at 3.6v) means 27volts through each motor. This would leave me with 10 cells left to play with, unless i just buy another 5 and make another pack and parrallel them?

this is probaly naive thinking but couldnt i do this:- (see pic) ?

only thing im concerned about it now is weight, 24 cells is already 50kg.

 

[agniusm]

do you already have the motor and the el moto frame assembled? with 24 cells you would either build a 12S or 24S pack in order to use the cells. why did you think it would be 16S?

rather than using the little short circuit thingies, you could use a the headway 24S BMS and it will provide up to 100A of juice which is about all those cells will make in any case. jmho.

or you could bypass the BMS on discharge and use an LED on the console to signal LVC.

Could you enplane how to bypass discharge fets on BMS if possible? John in CR started new thread.

There is new member working on hes own distributed BMS system. He is thinking it would cost 110USD for the main unit and 12USD for slaves. Might be worth waiting.
CALB's should do 3C constant so on 2P pack you can draw 240A with 400A bursts.
You can get something for onboard charging like this: http://hztiecheng.en.alibaba.com/pr...Lead_acid_and_LiFePO4_battery_EV_charger.html

 

[agniusm]

Sounds like a plan Dogman, i want to be able to go at least 40miles per charge and want a fast charger, those cellogs look ideal and very handy for me i think ill order a couple to help keep an eye on the cells.

i was thinking i would put 15 cells in series to make 54volts (if each cell was at 3.6v) means 27volts through each motor. This would leave me with 10 cells left to play with, unless i just buy another 5 and make another pack and parrallel them?

this is probaly naive thinking but couldnt i do this:- (see pic) ?

only thing im concerned about it now is weight, 24 cells is already 50kg.

Not good. Get few more cells, balance them, put the pack together, stick on BMS, forget about it, come back to check in a while. I think none of centralized BMS will do 200amps.

 

[heathyoung]

Bottom balancing is a crock of horse puckey. Jack Rickard I think is the name of evtv fame. I've watched some of his programs, he probably has been resposible for the demise of many a cell.

I'm working on a largeish pack (44S) of CALB 40AH cells at the moment - They were purchased second-hand (warts and all, but most read at 38AH, and still decenct IR). There are a lot of cell-top BMS units for these - some are good, some are shit.

Mine came with some shit ones that I modified so they work well (they shunt at the correct voltage, but don't throw a HVC signal till about 4.2V - good way to overcharge). The main controller unit (charge control) was also poorly designed, the resistor values were chosen wrongly, so at HVC, the mosfets didn't turn off the charge, they just dropped it a smidge and dumped loads of heat. Ugh. Brand was GBS Battery systems. Fortunatly, they are easily fixed with the adittion of 1 SMD resistor (and an LED so you can tell they are shunting), but I digress...

CALB cells, properly balanced to begin with (ie hook them all in parallel for a week - if they are the same voltage and they will balance) usually stay in balance pretty well if you don't push them for all they are worth. If you get a runt, or a cell with a lot more or less capacity - you are asking for trouble.

All BMS units top balance - with large cells, you want a reasonable balancing current, as the larger the difference in capacity = larger time spent balancing. The output current ratings of some of these cheap BMS units (like the BMS battery ones) are a nice work of fantasy, and tend to result in FET's melting off the board or shunts becoming fuses (personal experience). You can use them, but do yourself a favour and get a lower current one, and use it for charge control (watch the rating on the charge fet - you want to ideally charge the 40AH cells with 0.3C = ~12A) and the discharge control, use that to run a relay / contactor for your controller negative or, use a little one for your controller ignition line.

 

[dnmun]

i like the way the headway balances. it seems they all end up exactly at 3.65-3.66V depending on what the charger voltage is. when they manufacture them they test the shunt transistor to see at what voltage it turns on and then the tester will add a few extra resistors in parallel to the ones on the resistor divider that turns on the shunt transistor so it is so exact it cannot be matched imo.

plus the shunt transistor they use has a higher transconductance than the shunt transistor that ping uses it think. it seems like the transistor turns on fully within a few tens of millivolts and it takes a hundred millivolts it seems for the ping signalab.

about the idea of discharging without the current from the pack flowing thorugh the output mosfets:

a, instead of connecting the controller to the P- lead on the BMS, just connect directly to B- on the battery, skipping the BMS altogether. the BMS is only in the negative lead of the battery so the positive doesn't come into play at all.

b, since you will want to know if the pack reaches LVC in operation, you can either drive the led on the dashboard with the gate drive that normally drives the mosfet gates, or you can take a high voltage lead to the led through current limiting resistor and then the return from the led would go to the P- terminal on the BMS so that as long as the mosfets were turned on, then the led would be lit. when the output mosfet turns off then the led would go off. so you could use a green led as an operating indicator of LVC status and when it goes out then you could shut down the operation of the controller manually. no need for relays, just wait for the green led to turn off. most people know when they have reached the end of the pack, either from loss of power or seat of the pants, or from the cycle analyst or wattmeter readings.

c, since there is no LVC to turn off the current draining the pack, then if you leave the controller or the DC-DC converter or lights on then the pack can exhaust itself to nothing in the interim when you forget. that is the big risk and why using the mosfets to control the current is useful. but not essential since most people do not leave stuff plugged in when they shut down. but it happens, it happens a lot.

but the balancing function is why i consider the BMS so valuable. when people talk about manually balancing a pack, i know they have never had to do it. i have had to fight to bring new built packs from old parts into balance and it can take days of tweaking with power resistors draining some sections and single cell or 4 cell charging to push the others up to match, and it is never as exact as what the BMS does with no effort electronically. the last pack i built took 4 days to balance. 2 days were spent just getting the packs within 30% of each other with the power resistors and then when the BMS would remain on and balance it took another 2 days.

but for this situation where he is gonna run the two motors in series, then he could get another 6 of the CALB cells to make 30 total and make his pack 15S2P and if there is 80Ah in parallel then he should be able to pull 200-250A. but not through the headway, it would have to bypass and go directly to the controller.

 

[999zip999]

Are these cells new or used ? If new or used I would number all cells and make a list of the voltages before charging.

 

[neptronix]

Bottom balancing: best way to get the least amount of watt-hours out of a pack..

Not a fan of Jack Rickard at all.. :|

 

[Stevesie]

...

CALB cells, properly balanced to begin with (ie hook them all in parallel for a week - if they are the same voltage and they will balance) usually stay in balance pretty well if you don't push them for all they are worth. If you get a runt, or a cell with a lot more or less capacity - you are asking for trouble.
....

Yep, they do need to be well matched cells when purchased as I've found, or they are a bit of a headache. I bought into the bottom balancing idea for a while as a solution, but not so much now....

 

[amberwolf]

I was hoping to get 150-200amps out of these batteries!

I don't know the specs for those exact cells, but typically that sort of cell is good for 1-2c without hefty voltage sag. So you'd need to parallel at least two sets of cells to get the current you want out of them without pushing them way too hard. At 40Ah, 2C is only 80A. Even two in parallel is only 160A at 2C.

It is also possible you will need some heavier currents at launch from a stop, so you might want to prepare for the possiblity of either pushing the cells even harder or of adding another parallel set of cells.

 

[Jeremy Harris]

Thanks for reminding me of the guys name. I watched a couple of Jack Rickard's videos once and was impressed by just wrong the guy could be. He presents wholly incorrect stuff so authoritatively, though, that I bet many still believe it. I concluded he was just a bloke with too much money and spare time on his hands who liked the sound of his own voice. Being technically correct wasn't something he seemed to ever be bothered about.

 

[agniusm]

If it is SE version of calb battery then standard discharge rate is 0.3C so for 40A its 12A, 30s - 4C, 5ms - 12C. I know many are using these cells as 3C constant capable.

 

[V-spec]

Thanks everyone this is all really helpful, so plan if action is as follows:

1. Buy 6 more 40ah cells (yes they are the se range SE40AHA) http://www.ev-power.eu/CALB-40Ah-400Ah/SE40AHA-Lithium-Cell-LiFePO4-3-2V-40Ah.html
2. Stick all of them in parallel for about a week ad let them do there thing.
3. Number and check each cell for voltages/ah and any which are not inline separately charge them till they are correct.
4. Don't use my Mini BMS cell modules (if i do check they shunt at the right voltage)
5. Ultimately buy two Headway XJI 48V BMS 16 CELLS 50A/100A or just call them up and tell them what i need.
6. Buy i decent charger either a small one with constant voltage and current for the meanwhile to charge them in parallel, then a bigger one once everything is good to go.
7. Dont bottom balance but top balance!
As to wiring it to the controller and how it shuts off at LVC/HVC this is another area i will have to workout when i get there, i want to power the motors with a constant 150amps and peak it at 200amps!
8. Regen programmable controller is probably a good idea.

8. Surely there must be a computer controlled bms/charging system that does this all for you? You can just hook up each cell and it measures it individually and charges it and balances.

Well its all starting to make sense now, the batteries are thankfully new and still have warranty!

 

[V-spec]

http://www.ebay.co.uk/sch/i.html?_trksid=p5197.m570.l1313&_nkw=li-fe+charger&_sacat=0&_from=R40

will any of these actually help me?

 

[dnmun]

if you are going to be using just 15S then the 16S BMS will not work since you will have a channel empty and that will set off the LVC control signal. i can hack the ping v1 and v2.5 signalab BMS to 15S but not the new 16S digital BMS from headway.

i think i can hack the older 16S BMS from headway but it depends on how much of the black goop they pour on the control lines. you have to get access to the transistor that turns on the LVC signal line and unsolder the resistor so it just depends on each of those old BMSs and if you can reach the resistor with the soldering iron.

the ping is easy, and you are gonna have to discharge outside the mosfets too if you are gonna pour this much power through them.

 

[V-spec]

if you are going to be using just 15S then the 16S BMS will not work since you will have a channel empty and that will set off the LVC control signal. i can hack the ping v1 and v2.5 signalab BMS to 15S but not the new 16S digital BMS from headway.

i think i can hack the older 16S BMS from headway but it depends on how much of the black goop they pour on the control lines. you have to get access to the transistor that turns on the LVC signal line and unsolder the resistor so it just depends on each of those old BMSs and if you can reach the resistor with the soldering iron.

the ping is easy, and you are gonna have to discharge outside the mosfets too if you are gonna pour this much power through them.

Maaaaaaannnnn, what a pain :? think i just best email Headway and tell them what i need see if what they can do for me? Surely they will be able to do that all for me.

 

[dogman dan]

If you chose to use 14s voltage, you could charge with a hyperion 14s RC charger.

Or make the pack into two 8s sections, then charge with two 8s RC chargers. Two 200- 300w 8s chargers would get you charging and balancing pretty good, not taking forever for non balancing charges. They can be set for single cell charging, so you can get your paralell them all and charge to 3.65v out of them as well.

Then it's just a matter of monitoring your voltages, so you don't overdischarge. The charging bms is built into the chargers, which charge anything, lead, lipo, lifepo4, nicad, nimh. You are the bms on the discharge. Just pay attention to your voltmeters and wattmeters. When you've used 80% of your capacity you should stop.

Out on the road, you could still plug in for awhile here and there, using a cheap 48v lifepo4 charger with a low end of charge setting. Something like the 5 amp kingpan chargers.

I don't know where to get this stuff in EU, but this link shows the type of 8 s charger I mean.

The RC chargers mostly require a DC power suppy.


http://epbuddy.com/index.php?main_page=product_info&cPath=17_20&products_id=170

The china source, but they tend to only have the cheaper stuff in stock. http://www.hobbyking.com/hobbyking/store/__9005__iCharger_208B_350W_8s_Balance_Charger.html

This one has actually worked ok for me for two years, but only 150w so kinda slow for your size of battery, 80 ah.

http://www.hobbyking.com/hobbyking/store/__7523__Turnigy_Accucel_8_150W_7A_Balancer_Charger.html