BMS or Controller Cut Out?

[rollingreenhills]

Hi,

I have just built a 48v 500w Bafang BPM (code 13) 20" Wheel Bike, with a ElifeBike Controller (19A) and a home made 16s headway (screw type) battery pack.

I purchased the controller from a reputable retailer, and the cells I purchased secondhand from a local EV manufacturer, I tested all the cells and they hold 3.2 volts a piece.

Cells: http://www.ebay.co.uk/itm/262396300047?_trksid=p2060353.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

Controller:http://www.elifebike.com/peng/iview.asp?KeyID=dtpic-2012-8W-QNK2.073KQ

BMS: http://eclipsebikes.com/lifepo4-576v-cell-30a60a-p-1042.html

Charger: http://eclipsebikes.com/lifepo4-battery-charger-p-1038.html

Going on my first test ride, when I hit a hill the bike cuts out. The power loss is instant, and the LED's on the throttle power gauge go out and will not come back on.

There is a voltage sag (observable on the throttle power gauge) when you run the wheel with it pushed against the ground, and then the power cuts out.

On disconnecting the battery I apply a voltmeter and get 48 or 49 volts, although this rises to 51v overnight charging, and still it cuts out. The charger puts out 57v.

The cut out is less likely to happen if you leave the charger plugged in. . . . . when doing the above test (in bold)

I appreciate the home made pack muddies the water some, but I would be very glad of a few pointers.

 

[amberwolf]

Sounds like the pack is either not fully charged, or not balanced.

What are each of *all 16* of the cell voltages while under the load, listed per-cell?

FWIW, used cells are worth nothing unless first proven to work under load conditions you'll use them under, and still have the capacity you need.

 

[rollingreenhills]

Oh dear amberwolf that is worrying regarding the cells . . . so they could be a load of rubbish.

I am not sure I have the technology to record the voltage from each cell under load . . . the voltage of the pack as a whole drops significantly when under load, as the power gauge on the throttle goes from green to orange (before it cuts out)when under load.

If it were just a few bad cells, would the overall voltage drop to this extent?

 

[izeman]

Headways are known to sag a lot. So if ONE cell in your pack is broken then your bms will shut off really early even if charged.
You should check every single cell when fully charged. Then ride and immediately check voltage of the single cells when the motor/controller cuts off. If one cell is way different then the others then you got the bad one.
If you say your charger outputs 57v, then this is the voltage your pack must reach to be fully charged. As amberwolf said: the pack is not ready charged. It may take DAYS to fully balance the pack. But loooong balancing is an indication for a broken cell. So: connect to charger for some days, then measure cells, then measure when bike stops, then post voltages here.

 

[izeman]

To answer you last question : definately YES.

 

[rollingreenhills]

Thank you izeman.

Will do and report back.

I have checked with the shop where I purchased the BMS, and they have recommended to do the exact same thing that you have.

Thanks

W

 

[dogman dan]

For all the others reading, a bit of clarification.

When the motor stops, then starts back up again as soon as the load is removed, that's a controller LVC cutout. it may tend to porpoise on and off with full throttle, but keep running at a crawl.

When the motor stops, the lcd goes off and does not come back on. That is more likely to be the BMS shutting off the battery. However, when you unplug and plug back in and it runs fine for quite some time, that indicates the cutoff was just a bad or loose connection at the battery plug.

Either one requires you to unplug and replug to get going again, but if it's the bms, you won't go far before the next bms shutdown.

 

[999zip999]

The bms shuts down when it senses one low cell 2.5-2.7 volt. May be a bad signal ( sense wire) or cell low capacity ( can't fill up ). I like to put a set of sense wires on a pack with bms for checking with a cellog like cell monitor. Plus before building pack put all cells in parallel on one charger set to 3.5v - 3.6v for lifepo4. This helps a bms keep the pack in balance. Yes bms can take days to balance. A single cell charger for lifepo4 set to 3.6 volt is a great tool. Plus a 12v car tail light to bring down a high cell. So needs to balance.

 

[izeman]

Good idea to put them all in parallel and charge as one big cell. BUT take care that all cells are about the same voltage lecel. If you connect full and empty cell together you will get some nice spark and maybe even damage the cells.

 

[rollingreenhills]

OK I have (limited) data!

All cells were at 3.2V except one which was at 0.02V !

So I had a spare cell put this one in and all were now at 3.2v.

I went for a quick ride to check everything was working, and after about 1 mile the pack cut out again.

Tested cells on return and a different cell (next in series to the replaced cell) is now showing 1.5v. most of the others are showing 3.2.

So. Bad cells? Incompetent newbie? Badly wired pack?

PLeeeease. About to go mad.

 

[999zip999]

You must fully charge and balance your pack. There's a very big difference between 3.200v and 3.399 volts in capacity. Meaning how much waters in the glass. Cells should cut off by bms at 2.5v,-2.7v which is zero, empty. 3.6 volts is full. Don't do bike capacity test it will greatly unbalance a newest pack. Balance grasshopper. Return shit cells.

 

[izeman]

from a far distance, w/o having more info or testing, i'd say at least this 1.5v cell is bad as well.
i can't imagine how this could happen, as the BMS should have prevented going ANY cell below it's low voltage cutoff (LVC) which is 2.0v (as stated on the webpage you gave). i guess 2.0v is too low for that cell, but can be wrong - i would not go below 2.5v.
if you read more, you'll find out that nowadays (almost) noone uses those cells for a new build. there are WAY better cells to use. they can't deliver high amps, and seem to fail quite often. still it's very easy to build a pack with those cells because of the screw terminals.
if you say they are NEAR 3.2v, this is not helpful at all. to help you we need to have exact info. did you charge ALL of them to exactly 3.6v? +/-0.05v? then measure ALL of them immediately after cut-off? use a marker and give them numbers, and write down all voltages ALL the time. otherwise you won't find the bad ones.
edit: 999zip999 you beat me on that, i guess you can type faster then me
edit2: i googled these cells which seem to be 38120 and they can deliver 10c, so not bad, and seem to have a LVC of 2.0v - which still seems very low to me *ggg*

 

[999zip999]

Pen and paper and yes number cells.
1. 2.85v
2. 3.60v
3. 3.45v

16. X.xx volts.
On paper they are great ebike cells somtimes a little hard to balance, because. Of large one cell format and they are headway. A single cell lifepo4 charger that stops at 3.6v is what a lifepo4 builder loves.

 

[amberwolf]

Tested cells on return and a different cell (next in series to the replaced cell) is now showing 1.5v. most of the others are showing 3.2.

So. Bad cells? Incompetent newbie? Badly wired pack?

PLeeeease. About to go mad.

I can only basically repeat my original post, especially the first phrase, because it's unlikely your pack is actually charged / balanced.

Leave your pack on the charger (assuming the BMS can balance the cells) for a week or two, without using it.

Then do the individual voltage test under load.

You can do this by hooking up your voltmeter leads to the first cell, then apply the load (ride the bike and watch the meter (tape it to the handlebars?). Or hold the mechanical brake engaged while pushing throttle up and watching the meter, while sitting in your test area). THen write that voltage down, and move the meter leads to the next cell, apply load, etc. till you have all cells measured under load.

 

[rollingreenhills]

Thanks to those of you who have helped out with my other questions.

I am building a pack with with some 2nd hand headway cells (screw type) and I have noticed some strange behaviour.

The cells seem have their capacities consumed at massively different rates.

After use (with a BMS) I notice that after a cut out just one cell was down to 0.02v, the rest had hardly changed.

This I repeated and now 3 cells have dropped to zero volts (1 after an initial stint at 1.5volts), whilst the other cells remain at 3.2v.

When linked in series and discharged with a headlamp, I notice that some cells discharge very quickly whilst others just stay the same. Say I have 4 cells in series and and they all start at 3.2 volts, after a time 3 cells remain at 3.2v and one is at 2.6v.

Why is this? Is it normal?

If this is happening when using a BMS it is no wonder its is cutting out, as just a few cells are discharging and when they hit 2 volts "click" no more juice. I guess the BMS can only balance so much, it just can't keep up?

 

[kdog]

first, how did you check the voltages? was it directly on the cell terminals?
assuming you did ( and its not a faulty sense wire) then it sounds like you have a few dead cells.
if they truly are getting to less than 1v quickly then throw them out. LFP cells are inherently more stable and safer than li-ion/lipo but they can still do bad things.
discharging in a series string as you stated should give you an idea of which ones to toss.
as for the BMS:
if it is not shutting down when a cell hits ~2v during discharge then either it is faulty or the sense leads/connections are faulty
trouble shoot this before going further
the BMS only balances during the period at the end of charge whilst the charger is still connected. not during discharge or when pack is idle. balancing currents are usually somewhere around 10-100mA and may take a very long time to balance a pack if it is really out (like days/weeks)
you can manually balance the pack by bring down top cells with a light or similar, or if/when cell V are similar, parallel all cells together and charge with a 1s charger.
A problem with LFP is that it has a very flat discharge curve, so its hard to tell what state of charge the cells are at by voltage alone.
k

 

[rollingreenhills]

Dear kdog,

Thank you so much for you reply.

Yes I did check the voltages across the cell terminals.

From what I have observed, these "bad" cells "collapse" when discharged so even if the BMS cut out at 2v, they then just loose their voltage .. . ?

Does this sound feasible?

For example, I observed the 1.5v cells soon after ride at 1.5v, then on next measurement it was down in 0v (having not been touched) since last measurement.

This is one of the most confusing things I have ever attempted.

 

[dogman dan]

Those cells are certainly done.

 

[flippy]

those cells are gone. just replace them and get some new ones.

 

[rollingreenhills]

Thing is I am at the point now where I am wondering whether it is worth trying to get this to work.

If the cells are all so different in their characteristics will they ever balance? 3 cells out of 16 . . .what next.

Are the cells that are dying bad, or are they being "overworked" because the other cells are bad?

Should I just cut my losses and get new cells?

Edit: I mean all new cells (x 16).

 

[teklektik]

...Say I have 4 cells in series and and they all start at 3.2 volts, after a time 3 cells remain at 3.2v and one is at 2.6v.

Why is this? Is it normal?

If this is happening when using a BMS it is no wonder its is cutting out, as just a few cells are discharging and when they hit 2 volts "click" no more juice. I guess the BMS can only balance so much, it just can't keep up?

• You should be seeing 3.65V on the cells if they are hot off the charger. The BMS isn't going to be doing any balancing at all at your starting voltage of 3.2V.
• LiFe has a very flat discharge curve and it's almost impossible to determine the actual state of charge over 80% of the discharge cycle by simply looking at the open circuit voltage. This is very different than all other types of Li-ion cells. 3.2V is almost meaningless as an an indication of charge state except that it's not full charge.
• Although the BMS may have a cutoff voltage of 2.0V per cell, this is probably more like abuse than protection for common use. 2V is the Headway specification to get every last electron out of the battery but it's really flogging the cell. You probably want to cutoff at 2.5.
• If you want to build a pack with used LiFe, you really need to get a single cell charger (2A is fine) and charge them all individually first. The BMS will take forever. If your BMS has a 60ma balance current and your cells are way out of balance - say 50% - then it will take (50% x 10000mAh)/50ma = 83hours to balance - after the first cells hit full charge, so maybe 90+hours. Minimum. Since some of your cells are clearly fully discharged, you are looking at no less that twice that.

• For your 4s experiment, you should be seeing a HOC voltage of about 4x3.65V = 14.6V. If you are not seeing this and the charger in not cycling, then your charger is NG (or you have some really sick cells).
• If your BMS is running the cells down to 0V, then it's NG (or you have some really sick cells).
• Chances are you have some bad cells there, or have killed some with that testing...
• Get a single cell charger - something like this - so you can charge and evaluate your cells without worrying about your charger and BMS. You want this anyway if you are building a Headway pack, so get it before you get new cells to save them from a bad fate as well...

 

[amberwolf]

I am building a pack with with some 2nd hand headway cells (screw type) and I have noticed some strange behaviour.

We tried to help you in your other thread about this problem, but you did not respond with the requested info.

https://endless-sphere.com/forums/viewtopic.php?f=14&t=80322#p1183972

Based on your original info and the new info, you still have the same problem that the pack is not fully charged and is not balanced. Whether it is caused by the cells being bad or just that you haven't yet charged them up yet is unknown.

I highly recommend you let me merge this new thread for the same problem with the original thread so that those trying to help you can do so better, and have all of the information in one place.

I also recommend that you not start new threads for the same problem; if you keep all of your project in one thread, people won't have to keep asking you for information already discussed in other threads.

 

[rollingreenhills]

Dear Amberwolf,

I am sorry for any incovnience caused.

In my warped sense of reasoning I fathomed this constitutes a seperate problem.

By all means merge. . I bow to your good judgement on that.

All the best, and thanks to you all.

RGH