BMS for regenerative braking. How do i know ?

Sparfuchs

1 kW
Joined
Nov 22, 2020
Messages
305
Hello EV friends,
i'm about to build a 14s4p battery pack with samsung 21700 40T cells and bought this BMS:

https://www.aliexpress.com/item/1005003606124235.html?spm=a2g0o.order_list.0.0.13481802iw9GWj

(EDIT: i just got told this is a common port bms. So i'll try that one instead:
https://www.aliexpress.com/item/4000875181450.html?spm=a2g0o.productlist.0.0.11ca2c71r868fh&algo_pvid=e1f87802-ea45-4c20-9e31-52588485b3cc&algo_exp_id=e1f87802-ea45-4c20-9e31-52588485b3cc-0&pdp_ext_f=%7B%22sku_id%22%3A%2210000010076203402%22%7D&pdp_pi=-1%3B20.68%3B-1%3B6.6%40salePrice%3BEUR%3Bsearch-mainSearch
Unfortunately i wasn't able to find any 14S,60A, separate port bms with less than 15mm thickness on aliexpress. So if anyone has an alternative please let me know.)


As non of the sellers i asked on aliexpress knows anything about regenerative braking i was hoping someone here could tell me how i know if the BMS supports it ? Is there a way to check ? Or to find out before buying it ?


Thanks for your Help.
Best regards
Sparfuchs
 
It has a common port for power and charging, so it should be able to treat the regen voltage as a regular charging input, provided the regen current doesn't exceed the max charging current.

That's my guess; I've not tested this myself yet. I'll try it in my next build.
 
GalFisk said:
It has a common port for power and charging, so it should be able to treat the regen voltage as a regular charging input, provided the regen current doesn't exceed the max charging current.

That's my guess; I've not tested this myself yet. I'll try it in my next build.
Thanks for your reply GalFisk,
damn.. you are right - it's a common port bms. I wanted to buy a separate port bms but messed up.
But at least i know know, so thanks a lot :)
So i guess i'll use that one (14S, 60A) instead:

https://www.aliexpress.com/item/4000875181450.html?spm=a2g0o.productlist.0.0.11ca2c71r868fh&algo_pvid=e1f87802-ea45-4c20-9e31-52588485b3cc&algo_exp_id=e1f87802-ea45-4c20-9e31-52588485b3cc-0&pdp_ext_f=%7B%22sku_id%22%3A%2210000010076203402%22%7D&pdp_pi=-1%3B20.68%3B-1%3B6.6%40salePrice%3BEUR%3Bsearch-mainSearch
 
Is really no one here who can help me out ?
At the moment it seems like i'll have to try it with the mentioned bms an just hope that it will work and that the bms..and more important the battery won't get damaged ?
 
What do you want the bms to do with regen?
Cut if it thinks the battery get charged to hard?
Make sure that the bms dosent cut during regen?

If you want it to be able to cut it should be one port for both C and P.
If you dont want it to be able to cut get one with separate C and P ports.
 
j bjork said:
What do you want the bms to do with regen?
Cut if it thinks the battery get charged to hard?
Make sure that the bms dosent cut during regen?

If you want it to be able to cut it should be one port for both C and P.
If you dont want it to be able to cut get one with separate C and P ports.
Thanks for your reply j bjork,
i didn't even know what to consider and what are thinks i might want or need. I just wanted a bms that "supports" regenerative braking. I got a little more knowledge now but need someone to confirm my thoughts.
 
Is it right that theoretically almost every available bms (independent of common or separate port) can handle regenerative braking without getting damaged, provided that the braking current isn't higher than the output current of the bms ?

So lets say a separate port bms rated for 100A (peak) output current and 30A charging current can handle up to 100A braking current because it uses for regen brake the same connection and if it can handle 100A going out it could theoretically also handle 100A going in, right ?

But why aren't common port bms rated for same output and input current if they use the same connection ?

I found explanations of the technical difference between common port and separate port but haven't found the advantages or disadvantages of the two options in use of e bikes and regenerative braking.
So separate port won't cut if the battery is fully charged, but as the battery spends much much more energy to power the motor than it gets by regen braking, the only and rare scenario where this would be a problem is when you start with a fully charged battery from a mountain and use regen. But also with a common port bms you can't use regen in that case. So the only difference is that with one you mustn't and with the other you can't use regen braking when the battery is fully charged ?

So lets say this is a advantage of the common port bms... why do separate Port bms even exist and why does almost every e bike battery use them instead of common port ?

As the braking current with a sep. bms goes in "uncontrolled" through the wrong side, is there a difference with the cell balance if you use a sep. or a common bms for regenerative braking ? Or do bms always keep the balance and not just while charging ?

Is there anything else i should think of or is it really that irrelevant what bms you use ?
 
Sparfuchs said:
Is it right that theoretically almost every available bms (independent of common or separate port) can handle regenerative braking without getting damaged, provided that the braking current isn't higher than the output current of the bms ?

Yes, that is my understanding at least. There might be a little difference in what direction the current flows in the mosfets, but I dont think it is enough to be really relevant.

So lets say a separate port bms rated for 100A (peak) output current and 30A charging current can handle up to 100A braking current because it uses for regen brake the same connection and if it can handle 100A going out it could theoretically also handle 100A going in, right ?

Yes,I think so.

But why aren't common port bms rated for same output and input current if they use the same connection ?

I think it is to protect the battery.

I found explanations of the technical difference between common port and separate port but haven't found the advantages or disadvantages of the two options in use of e bikes and regenerative braking.
So separate port won't cut if the battery is fully charged, but as the battery spends much much more energy to power the motor than it gets by regen braking, the only and rare scenario where this would be a problem is when you start with a fully charged battery from a mountain and use regen. But also with a common port bms you can't use regen in that case. So the only difference is that with one you mustn't and with the other you can't use regen braking when the battery is fully charged ?

It is not that it wont cut, it is that it cant. The mosfets conduct in one direction, so the discharge mosfets can only cut discharge and the charge mosfets can only cut charge.

So lets say this is a advantage of the common port bms... why do separate Port bms even exist and why does almost every e bike battery use them instead of common port ?

It takes less mosfets for a separate port bms, it could have 5 discharge mosfets and 2 charge mosfets or something.
A advantage is it can have a little less voltage drop on discharge as the current only has to pass one row of mosfets


As the braking current with a sep. bms goes in "uncontrolled" through the wrong side, is there a difference with the cell balance if you use a sep. or a common bms for regenerative braking ? Or do bms always keep the balance and not just while charging ?

Is there anything else i should think of or is it really that irrelevant what bms you use ?

I think the best solution is a bms that excepts the charging current you get on regen, preferably programmable.
 
Sparfuchs said:
Is it right that theoretically almost every available bms (independent of common or separate port) can handle regenerative braking without getting damaged, provided that the braking current isn't higher than the output current of the bms ?

So lets say a separate port bms rated for 100A (peak) output current and 30A charging current can handle up to 100A braking current because it uses for regen brake the same connection and if it can handle 100A going out it could theoretically also handle 100A going in, right ?

But why aren't common port bms rated for same output and input current if they use the same connection ?

I found explanations of the technical difference between common port and separate port but haven't found the advantages or disadvantages of the two options in use of e bikes and regenerative braking.
So separate port won't cut if the battery is fully charged, but as the battery spends much much more energy to power the motor than it gets by regen braking, the only and rare scenario where this would be a problem is when you start with a fully charged battery from a mountain and use regen. But also with a common port bms you can't use regen in that case. So the only difference is that with one you mustn't and with the other you can't use regen braking when the battery is fully charged ?

So lets say this is a advantage of the common port bms... why do separate Port bms even exist and why does almost every e bike battery use them instead of common port ?

As the braking current with a sep. bms goes in "uncontrolled" through the wrong side, is there a difference with the cell balance if you use a sep. or a common bms for regenerative braking ? Or do bms always keep the balance and not just while charging ?

Is there anything else i should think of or is it really that irrelevant what bms you use ?

Hi Sparfuchs,

I've been working on a design and also struggling with some similar BMS questions. I've recently come to a realization that there a two fundamentally different approaches to BMS design. Anyone that disagrees please correct me because I am still trying to get a solid grip on this stuff. The first approach is what you have been talking about; a BMS that connects to the battery and everything that wants to move power IN or OUT of the battery must go through the BMS. The BMS protects the battery by connecting the battery when operating within safe parameters or disconnecting the battery when the voltage goes too high or low (using programmable limits). The connecting/disconnecting is done by mosfets acting as a solid state switch. These BMSs will have power ratings that correspond to the mosfet's power rating (30amps, 100amp, etc.). In my quick research of the common vs split port BMS issue, it seems that split port were commonly used by solar charging applications but now that MPPT charge controllers have become common in solar installations it seems the split port BMS are becoming less useful in solar applications. Don't know why ebike builders would use them.

The second approach to BMS in CAN bus (Controller Area Network). In a CAN bus BMS the voltage monitoring leads connect the same way but the power does not run through the BMS. All the devices that connect to the battery connect directly to the battery and they are connected to each other via CAN bus wires (two wires). The BMS, motor controller(s), the battery charger connect to CAN bus and talk to each other as needed. The battery charger when charging talks to the BMS to determine how fast it can charge and when to stop charging. The motor controller talks to the BMS to determine the SOC (State Of Charge) this can be used to stop the motors or limit the current draw at the end of a discharge cycle.

In my project, I am planning to use Brushless DC Motors with VESC motor controllers. The VESC software can be used to control how aggressively to apply regenerative braking. This affects how much power you will generate via braking and of course how quickly you will decelerate; very configurable. Dual VESCs can talk over CAN bus to synchronize motor speeds. From my research it seems that the motor controller is really what governs the behavior of your regenerative braking. And if you go with a CAN bus (smart) BMS then the main question left on the BMS is what type of cell balancing do you get (passive or active). Because then your BMS is just monitoring your cell voltages and communicating to other CAN buss enabled devices.
 
One thing to keep in mind is if the BMS disconnects the battery during regen, the voltage on the controller will shoot up and could blow something up. What you really want is a logic signal that will disable the regen if any cells go too high.

As long as the output FETs are on, the direction of the current won't make any difference. The current rating will be the same in both directions (for the BMS).

Unless you live at the top of a hill and fully charge your pack before leaving, it will be unlikely regen will over charge the pack to the point it needs protection.
 
fechter said:
One thing to keep in mind is if the BMS disconnects the battery during regen, the voltage on the controller will shoot up and could blow something up. What you really want is a logic signal that will disable the regen if any cells go too high.

As long as the output FETs are on, the direction of the current won't make any difference. The current rating will be the same in both directions (for the BMS).

Unless you live at the top of a hill and fully charge your pack before leaving, it will be unlikely regen will over charge the pack to the point it needs protection.

If your motor controller speaks CAN buss you can tell it to disable regenerative braking on the condition of a fully charged battery. In any case, if you are trying to maximize your battery life you should never be taking the battery to 100%; so there should always be some head room between your 90% full charge and 100%. In any case it seems like you wouldn't want to change the driving behavior by disabling regen because of the state of charge of the battery; so it seems that the best way to go is to have logic that either sends the regen power to the batteries or shunts the excess power to a resistance that just creates some waste heat.

I haven't tried this but here is an example implementation:
https://spintend.com/collections/frontpage/products/over-voltage-protector-rehostatic-braking-resistor-module
 
j bjork said:
It takes less mosfets for a separate port bms, it could have 5 discharge mosfets and 2 charge mosfets or something.
A advantage is it can have a little less voltage drop on discharge as the current only has to pass one row of mosfets

Thanks a lot for answering all my questions j bjork,

so in your example the current at a sep. bms has to pass 5 "aktive" mosfets on discharge, and 2 "aktive" mosfets on charge. The current on the same bms with common port wold have to pass 7 (5 aktive,2passiv) mosfets on discharge and 7 (2 aktive,5 passive) on discharge right ? That explains the voltage drop, but does a common port bms need more mosfets with the same current ?
 
David Rosales said:
Hi Sparfuchs,

I've been working on a design and also struggling with some similar BMS questions. I've recently come to a realization that there a two fundamentally different approaches to BMS design. Anyone that disagrees please correct me because I am still trying to get a solid grip on this stuff. The first approach is what you have been talking about; a BMS that connects to the battery and everything that wants to move power IN or OUT of the battery must go through the BMS. The BMS protects the battery by connecting the battery when operating within safe parameters or disconnecting the battery when the voltage goes too high or low (using programmable limits). The connecting/disconnecting is done by mosfets acting as a solid state switch. These BMSs will have power ratings that correspond to the mosfet's power rating (30amps, 100amp, etc.). In my quick research of the common vs split port BMS issue, it seems that split port were commonly used by solar charging applications but now that MPPT charge controllers have become common in solar installations it seems the split port BMS are becoming less useful in solar applications. Don't know why ebike builders would use them.

The second approach to BMS in CAN bus (Controller Area Network). In a CAN bus BMS the voltage monitoring leads connect the same way but the power does not run through the BMS. All the devices that connect to the battery connect directly to the battery and they are connected to each other via CAN bus wires (two wires). The BMS, motor controller(s), the battery charger connect to CAN bus and talk to each other as needed. The battery charger when charging talks to the BMS to determine how fast it can charge and when to stop charging. The motor controller talks to the BMS to determine the SOC (State Of Charge) this can be used to stop the motors or limit the current draw at the end of a discharge cycle.

In my project, I am planning to use Brushless DC Motors with VESC motor controllers. The VESC software can be used to control how aggressively to apply regenerative braking. This affects how much power you will generate via braking and of course how quickly you will decelerate; very configurable. Dual VESCs can talk over CAN bus to synchronize motor speeds. From my research it seems that the motor controller is really what governs the behavior of your regenerative braking. And if you go with a CAN bus (smart) BMS then the main question left on the BMS is what type of cell balancing do you get (passive or active). Because then your BMS is just monitoring your cell voltages and communicating to other CAN buss enabled devices.
thanks a lot for your detailed answer David Rosales,
in my case i'm not just limited by my know how, but also by the size of my battery box. Meanwhile i've finished my first build with 21700 14s4p cells in a hailong max box but as it usally has a 40A bms it was impossible to find a 60A bms that fits in the same space. I even had to modify the thinnest bms i found to make it fit. As i checked it seems like the CAN Bus bms takes more space than i've got and i guess my kelly kbs controller won't communicate with the CAN Bus bms.
But nevertheless useful information and good to know more about the options for further projects.
 
fechter said:
One thing to keep in mind is if the BMS disconnects the battery during regen, the voltage on the controller will shoot up and could blow something up. What you really want is a logic signal that will disable the regen if any cells go too high.

As long as the output FETs are on, the direction of the current won't make any difference. The current rating will be the same in both directions (for the BMS).

Unless you live at the top of a hill and fully charge your pack before leaving, it will be unlikely regen will over charge the pack to the point it needs protection.
Thanks a lot for your reply fechter,
So how do i avoid that scenario with the shooting up voltage ?
I guess i've to make sure that the braking current and voltage won't be ever more than the charging rating of the bms ?
Hopefully my kelly kbs48101x offers all the settings i need for a safe use.
 
With a dual port BMS, the discharge port will stay connected even if the cells go over voltage. This is one way to prevent the BMS from tripping during regen. But you also don't have over voltage protection for the cells.

In real life, it's hard to over charge the pack with regen unless you charge to 100% and live at the top of a hill. I've never had it happen to me but I don't live at the top of a hill. If I charge my pack to 100% and ride up a hill and come back down, the losses ensure the pack will not be 100% by the time I get back to the starting point.

Some controller will have a over voltage protect feature that will inhibit regen if the voltage goes too high. This is what you really want. You could suddenly lose regen going down a hill, but that's what brakes are for.
 
Sparfuchs said:
j bjork said:
It takes less mosfets for a separate port bms, it could have 5 discharge mosfets and 2 charge mosfets or something.
A advantage is it can have a little less voltage drop on discharge as the current only has to pass one row of mosfets

Thanks a lot for answering all my questions j bjork,

so in your example the current at a sep. bms has to pass 5 "aktive" mosfets on discharge, and 2 "aktive" mosfets on charge. The current on the same bms with common port wold have to pass 7 (5 aktive,2passiv) mosfets on discharge and 7 (2 aktive,5 passive) on discharge right ? That explains the voltage drop, but does a common port bms need more mosfets with the same current ?
Hmm, I dont think you get it.
You have a group of discharge mosfets that are parallel. They are then series connected to a group of parallel charge mosfets.
In a common port bms they are connected:

Battery-(to)-discharge fet-(to)-charge fet-(to)-discharge/charge port.

So all the current has to pass both the charge and discharge fets.
So both the charge and discharge fets need to handle the max discharge/charge current.

On a separate port bms they are connected:

Battery-(to)-discharge fet-(to)-discharge port and charge fet-(to)-charge port.

So only the discharge fet(s) need to handle the discharge current. The charge and discharge fets need to handle the charge current.
If the max charge current is lower than max discharge current you can get away with less charge fets.

(this is what I have seen at least, there might be other possibilities like the battery might be connected to both charge and discharge mosfets on a separate port bms)
 
Ok newbie on this forum,
So after building a 200 amp hour Prismatic LiFEPO4 system with separate port BMS for our caravan I now I have another project.
I have been gifted a brand new 40 cell 3.2v 32650 LiFePO4 battery, I am intending use it as a complimentary power system for my portable fridge.
I already have a spare Victron 75/15 MPPT, and a small 200/50 amp separate port BMS ready to go.
However, unlike the van setup I have no inverter in the system, so this leaves me wondering about the best configuration for the three devices.
Even though I pointed out that C- is not bidirectional on the Victron forum I received the following advice:

Hi Tex,
The P- is not bidirectional, but when you connect both P- and C- to the same point (BAT- in the MPPT in this case), the path becomes bidirectional.
The reason BMS's come with charge (C) and load(P) connections separately is because sometimes you want to isolate the charger from the load. (I was fully aware of this).
In your example, because the MPPT acts as both charging through solar and discharging through load output, you can connect P- and C- together. Does it make sense?
And you have more versatility in your system.


Frankly no, it doesn't make sense to me, can anyone confirm this is kosher please, and if not the best way to connect the devices please? :?

 

Attachments

  • Fridge battery (Ver3).jpg
    Fridge battery (Ver3).jpg
    317.3 KB · Views: 625
  • Fridge battery system.jpg
    Fridge battery system.jpg
    1.5 MB · Views: 625
You don't want to connect the C- and P- together on the BMS. This defeats the safety cutoffs. If you want to wire like that, you should find a single port BMS.
 
I can happily report that my build (20s17p battery for a 3kW EU moped) is now complete, and that the common port BMS handles regen without any issue at all.

The only annoying thing is that the event log of this smart BMS fills up with "charging" events for every time I brake.

Edit: if you want regen on a separate charging port, you can put a diode between the power and charging ports, which only allows current to flow from power to charging and not the other way around. You'll lose a little bit of energy, but you won't interfere with the functioning fo the BMS.
 
Back
Top