Expected BMS Behaviour

Get all your technical information about electric bikes here.
Post Reply
Lurkin   100 kW

100 kW
Posts: 1428
Joined: Jan 18 2015 4:31pm
Location: Melbourne, VIC, AUS

Expected BMS Behaviour

Post by Lurkin » Jan 26 2019 10:14pm

I have finished the spot welding on a 13S8P pack and I am bench testing a BMS with a 13S1P pack to test and make sure it actually works and will manage the cell groups during charging.

I have the BMS Connected to the pack based on the instructions for it (nothing unusual there) and have the BMS connected to a Yihua DC Power Supply (YH-605D). I've set the power supply to 54.6v as that is what the pack should charge to at 1A.

But the charging behavior is not what I thought it would be. The power supply is constantly switching between CC and CV faster than the display updates the current and and voltage. During this time, I have measured the voltage of individual cells at +4.2v - which is higher than I want them to charge to. It is supposed to have switched to balance charging at 4.2v. Eventually it stops drawing any current at all - presumably it should then bleed the cells and charge again?

- should it be constantly switching between CC and CV (to check the pack voltage?)
- wouldn't it start drawing current again when bleed down is complete and the charging to balance must occur?

The BMS is a 086-13S-40A. Any ides would be much appreciated.

User avatar
amberwolf   100 GW

100 GW
Posts: 27854
Joined: Aug 17 2009 6:43am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: Expected BMS Behaviour

Post by amberwolf » Jan 26 2019 10:24pm

If the charge current during balancing (on full cells that trigger it) exceeds the tiny (few dozen mA or less most of the time) balancing shunt current capability on each cell, then the cell will continue to charge beyond the trigger point, while balancing is still going on, so they could reach higher than what you'd expect.

If the BMS has a simple function of "shut off charge input whenever any cell begins balancing" then that won't happen.

But if it has the function of "leave charge input on when cells begin balancing, but shut off charge input when any cell exceeds HVC", then it can.

I don't know what specs the manufacturer of your BMS provides; many I've looked around at over the years don't seem to have complete info on exactly what they do, but I"ve never bought one (mostly because of cost, but partly because if I don't know exactly what it's going to do under all the possible conditions the pack will experience, I don't want to use it).


I haven't used any power supplies that switch modes, so I don't know what yours is doing (or why). Chargers designed as chargers do this, if their charging profile specifies it. The adjustable power supplies I've used for charging (Sorenson, Meanwell) just output a voltage, and have a current limit that when reached causes the voltage to drop to maintain that limit under the given load.

Lurkin   100 kW

100 kW
Posts: 1428
Joined: Jan 18 2015 4:31pm
Location: Melbourne, VIC, AUS

Re: Expected BMS Behaviour

Post by Lurkin » Jan 26 2019 10:49pm

Hi AW!!! Happy New Year! awesome to see you posting again last time I was on here you were taking a hiatus but its just great to see you again!

There isnt really any information about how it works. It does have some parameters but its not a description of how the charge process works. I'd paste it here but the table is not copying properly as it loses its format. it can be found here:

https://www.aliexpress.com/item/Free-Sh ... 57890.html

I put a multi meter on DC across the highest voltage cell. When the power supply stops registering a current draw (voltage = 54.6, current = 0.00) the cell slowly decreases in voltage.

It turns out I am too impatient for this process - eventually it stops decreasing the voltage on the cell and starts charging again (voltage 52.00, current = ~1a). then proceeds to follow the same process.

Sigh.

I really like the set and forget style of limitations imposed by the BMS but I cannot stand how slow the charge process is.. this is ridiculous.

I recall 'balance chargers' being sold on here and seeing others with suitcases full of plug back chargers - I remember thinking that's crazy and impractical... but now I get it. Charging via the balance leads directly to each cell and charging them all to a specific voltage regardless of how out of balance they are seems like the go - do you have any suggestions on this? I recall there being issues about 'isolation' but this is a hobby for me and I don't fully understand the issue. Presumably it has something to do with individually charging the parallel groups whilst they are still connected to each other in series?

User avatar
amberwolf   100 GW

100 GW
Posts: 27854
Joined: Aug 17 2009 6:43am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: Expected BMS Behaviour

Post by amberwolf » Jan 26 2019 11:21pm

FWIW, if the cells are well-matched, then you won't need to do balancing.

As they age, they'll grow less well-matched, and begin to become imbalanced a bit over time as they get closer to empty or full (but will still be balanced enough in the middle of the SoC).


If the cells are brand new and have never been charged before, then the first charge and balance might take a while, but if they are well-matched then you won't have to deal with it again.

If they're not well-matched, then you may have to deal with this every charge cycle, depending on how badly mismatched they are, and how deeply you discharge them.


If your BMS is programmable, you can raise the LVC and lower the HVC so it only uses the middle 80% or so of the SoC curve of whatever cells are in it, and then even with imperfectly-matched cells you won't see much of a balancing issue, at least until age sets in.


I don't use a BMS at all, and my EIG NMC cells remain balanced within a few hundredths of a volt or less, from dead to full, even after I ran out of juice and had to keep going, draining the top cell down to around 2.6-something volts, IIRC (way below dead). I did not rebalance the cells, just recharged normally, bulk charge, though I did it at lower current (a few amps) for the first part of the charge, to stress that low cell as little as I could in the time I had.

User avatar
amberwolf   100 GW

100 GW
Posts: 27854
Joined: Aug 17 2009 6:43am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: Expected BMS Behaviour

Post by amberwolf » Jan 26 2019 11:45pm

Lurkin wrote:
Jan 26 2019 10:49pm
Hi AW!!! Happy New Year! awesome to see you posting again last time I was on here you were taking a hiatus but its just great to see you again!
Oh, I've stayed on here, posting my own stuff and helping others where I can, just not moderating anymore (been a long time since then).

It turns out I am too impatient for this process - eventually it stops decreasing the voltage on the cell and starts charging again (voltage 52.00, current = ~1a). then proceeds to follow the same process.
If you can decrease that final stage charge current to be less than or equal to the minimum balancing shunt current, then it won't have to shut off to drain the cells--they'll maintain whatever charge state they have, while the low cells simply slowly catch up.

If the balancing current is very low, in the few-dozen-mA range like most, then it'll still take a long time, but the charger won't shut off so at least you'll know it's still going. ;)
Presumably it has something to do with individually charging the parallel groups whilst they are still connected to each other in series?
Yes, that's why each charger has to have it's output isolated from it's input. If it doesn't (like chargers that connect DC ground to AC ground, or AC neutral, etc), then the more positive charger will short out the more negative charger and cell group, because the grounds (negatives) of both chargers would be tied together. POOF. :/


My best recommendation if you have a pack that always needs to be balanced is to find out *why* it needs to be balanced, which is usually that a cell in a group (or a cell in each of several groups) is significantly different in Ri or capacity from the rest of the group, or is actively leaking internally (draining other cells).

A second reason is that a balancing shunt can be stuck on, so that it is always draining that group. It's rare, but it happens. Easy test for this is leave the BMS connected after charging and balancing, and if after several days one of the groups is lower than others, then recharge and rebalance, and then disconnect the BMS from the pack, and if after the same time period none of the groups have changed, then the BMS is at fault. If the same groups change to the same voltages, then the cell groups are at fault.


YOu can always keep using a mismatched-cell or defective pack, letting the BMS balance every time, or you can always bruteforce balance it with individual group chargers, but finding and fixing the problem is a better solution. :)

Lurkin   100 kW

100 kW
Posts: 1428
Joined: Jan 18 2015 4:31pm
Location: Melbourne, VIC, AUS

Re: Expected BMS Behaviour

Post by Lurkin » Jan 27 2019 12:18am

This battery was originally intended for a bike which would be used for commuting. Periodically I work crazy hours and its very easy to make mistakes when I am that tired. Equally any time I spend error checking or being less confident is less charge time.

My previous setup was a em3ev battery with BMS and pre made em3ev charger. Slow to charge but very, very user friendly and never had any mixups/errors.

The battery build is using Sanyo 18650ag cells and they are remarkably consistent. However, I always try to plan for poor conditions so if they eventuate I am ok. Reducing the charge process to - swtich off bike, plug in charger (plug to match and cannot mis plug in) is what this needs to end up as.

I've been reading about the bc168 which others have raved about which was a balance lead charging RC charger. Yes, the isolation on grounds seems to be the problem. Not sure how to get around this.

I assume then this issue is the same, when using one BC168 - because its charging 6x parallel groups from the same source of charge and when a user is charging using two/three BC168 chargers (plugged into a multiboard) because it would be 12 or 13 parallel groups charging at the same time, from the same group source... hmm

These chargers appear to have 6 'channels' which are charging from the same ultimate supply. How can one separate them/ensure isolation? or is this impossible as it would need to be done at manufacture before the circuit board for the charger is finalised?

Lurkin   100 kW

100 kW
Posts: 1428
Joined: Jan 18 2015 4:31pm
Location: Melbourne, VIC, AUS

Re: Expected BMS Behaviour

Post by Lurkin » Jan 27 2019 12:23am

I guess just thinking out loud to avoid the isolation issue. Get a plug back with an appropriate voltage (i.e. 4v is fine for me, not fussed about the .2) which can charge at 1a ish and attach to each group, charging them individually by connecting it to each parallel group then moving onto the next once charged?

sounds terrible compared to charging them all at once with a balance charger but it would be faster than whats happening how - the balance current is a miserable 45ma... its small enough not to even show on the power supply ammeter!

Lurkin   100 kW

100 kW
Posts: 1428
Joined: Jan 18 2015 4:31pm
Location: Melbourne, VIC, AUS

Re: Expected BMS Behaviour

Post by Lurkin » Jan 27 2019 12:30am

Here we go, looks like this has been covered before. I'd be wanting to do something like this (thank you mistercrash):

https://endless-sphere.com/forums/viewt ... 81#p748655

Image

User avatar
amberwolf   100 GW

100 GW
Posts: 27854
Joined: Aug 17 2009 6:43am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: Expected BMS Behaviour

Post by amberwolf » Jan 27 2019 1:20am

As long as you use a separate, isolated, power supply for each balancing charger (BC168 or whatever) then they'll all be isolated from each other, and using several of them simultaneiously to charge the entire pack at once is possible.

But that gets large and expensive. :/

Can also be made "idiot proof", by making a single plug at the battery end that everything is wired to, keyed so it can't be plugged in wrong.



Bulk charging via BMS is the simplest, cheapest method, at it's root, but is usually slower.

If the transistors on the BMS balance channels can handle it, you could parallel more resistors onto the existing ones, to double or triple the balance current. But I have no idea what transistors they used, or what capabilities they mght have. May also require increasing the airflow around the BMS, because more balance current means more heat produced. If it doesn't get warm now, no worries, but if it gets hot, you'd want to keep it cooler.


However...again, it is still better to fix a pack that has to *be* balanced all the time than to make a better balancer for it. ;)

Lurkin   100 kW

100 kW
Posts: 1428
Joined: Jan 18 2015 4:31pm
Location: Melbourne, VIC, AUS

Re: Expected BMS Behaviour

Post by Lurkin » Jan 27 2019 1:42am

I understand you are saying that two of these cannot be used unless their source is isolated. How do I do this?

Major is commenting in another thread that the negative is common to the first channel. Does this mean that even if the sources are isolated, the six channels are not? if so, what can I do about this?
It is difficult to separate the EnerDel 12S module in two. I think I'll take a couple of 6S A123 modules I have and hook them up. That way I can measure potential between the two. I really don't want to pop open the BC168. Having 6 separate chargers in the BC168 I was surprised and disappointed to find channel one negative is common to the power supply negative. That sucks and is likely the root of my problem. I thought having isolated power supplies to each BC168 would overcome that.
https://endless-sphere.com/forums/viewt ... 65#p852760

Yes, I get it - I don't expect this to be a major issue given the cells are actually pretty good at staying balanced/ they do not drift. But I'd still like a recipe to follow for a quickly balanced pack all the same.

User avatar
amberwolf   100 GW

100 GW
Posts: 27854
Joined: Aug 17 2009 6:43am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: Expected BMS Behaviour

Post by amberwolf » Jan 29 2019 4:07am

Lurkin wrote:
Jan 27 2019 1:42am
I understand you are saying that two of these cannot be used unless their source is isolated. How do I do this?
You would have to use independent source supplies for each. Each of the source supplies would have to be made so that it's output side (DC) is electrically isolated from it's input side (usually AC).

Some of them can be modified to be isolated if tehy are not already, but it can be a PITA to do.

Major is commenting in another thread that the negative is common to the first channel. Does this mean that even if the sources are isolated, the six channels are not? if so, what can I do about this?
The individual channels would *have* to be isolated from each other as-built. If they are not, then to be used on a series-connected set of cells, all of the series connections would have to be disconnected before you could even plug the channels into the cells. (or it would short them out). Since I don't imagine anyone would be doing that with their packs, I don't imagine it's actually designed so taht all channels ahve a common negative.

It is of course possible taht the first channel has it's negative connected to the ground of the source input. But as long as the source is completely independent from one unit to the next, and completley isolated from input to output, then there should not be a problem using mutiple units in series without dsconnecting the secitons of the pack from each other.

Having 6 separate chargers in the BC168 I was surprised and disappointed to find channel one negative is common to the power supply negative. That sucks and is likely the root of my problem. I thought having isolated power supplies to each BC168 would overcome that.
It should. If it does not, then there is still a connection between them somewhere. Find that and remove it, and shouldn't be a problem.

Lurkin   100 kW

100 kW
Posts: 1428
Joined: Jan 18 2015 4:31pm
Location: Melbourne, VIC, AUS

Re: Expected BMS Behaviour

Post by Lurkin » Jan 31 2019 7:26pm

What test do I conduct/ how do I know that the power supply ground is isolated from the charger?

If I use the bench ac to DC supply would it be testing continuity of GND on the AC wall plug to positive or negative of the DC output?

User avatar
amberwolf   100 GW

100 GW
Posts: 27854
Joined: Aug 17 2009 6:43am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: Expected BMS Behaviour

Post by amberwolf » Feb 13 2019 3:39pm

Lurkin wrote:
Jan 31 2019 7:26pm
What test do I conduct/ how do I know that the power supply ground is isolated from the charger?
Sorry I missed the question. :oops:

With everything unplugged from power, you can use an ohmmeter to measure continuity from each output wire to each input wire.

If I use the bench ac to DC supply would it be testing continuity of GND on the AC wall plug to positive or negative of the DC output?

Normally the only thing that *might* be connected is the AC ground to the DC ground, but occasionally I've had AC neutral connected to DC ground on PSUs that don't have a ground pin. :/ So testing all the inputs to all the outputs ensures you shouldn't have any worries on that score.

Lurkin   100 kW

100 kW
Posts: 1428
Joined: Jan 18 2015 4:31pm
Location: Melbourne, VIC, AUS

Re: Expected BMS Behaviour

Post by Lurkin » Feb 17 2019 11:38pm

All good, thanks AW, just need to find one to purchase now! 8)

Post Reply