Can a basic CCCV power source be left connected to a Li-ion battery with BMS?

[No_Shorty]

I am looking to integrate a small 13s2p battery into a rackmount enclosure to act as a UPS. As such I would like to have this battery charging whenever other sources of power are connected.

My basic plan was to use a JBD BMS and then a simple CCCV DCDC converter (RD RK6006) supplying it with power.

I would happily use an off-the-shelf charger, but it needs to fit into a 1RU without taking up much space.

I just wanted to check if this would work or if I am barking up the wrong tree/have overlooked something.

 

[harrisonpatm]

In general you don't typically want to leave a charger attached to a lithium battery, any charger, 24/7. If you're asking, can a CC/CV power supply be used as a charger? Then the answer is yes, because that's what nearly all chargers are anyway.

But it's still good practice to unplug or otherwise disconnect a power supply once the battery reaches full charge/voltage. If you leave it plugged in, and everything is working as it should, then 99% of the time nothing bad will happen. In theory, the CC/CV supply is not able to provide a higher voltage that what its set for, so there's no danger of overcharging a battery, right? Yeah, In theory. But all it takes is for one component to fail and for the supply voltage to raise higher than the battery limits. Potentially causing a fire. This is why it's advised to monitor a battery while its charging. When used as a UPS, you're describing leaving a battery sort of plugged into its charger 24/7, increasing risk of overcharging.

I've described the worst-case scenario. If you were to build a UPS with this method, it's very likely that it will work just fine and nothing will go wrong. I'd even be willing to bet that some lower quality, commercial off-the-shelf UPS's circuitry works exactly this way. I just want to advise you of the risks.

May I ask why you're going with 13s instead of 14s?

 

[amberwolf]

I'll second the above, with these specifics to add:

If your BMS actually works correctly, and completely cuts off all current flow from the charge port as soon as it detects a cell HVC, then it's "safe" to use a charge source that does not turn itself off once it's output current drops below it's minimum (which a normal lithium charger will do; some like the Cycle Satiator from Grin have programmable settings for this, most just have a preset rate).

If the BMS does *not* cut off all current flow once the cells have reached HVC, then the BMS design and the actual current rate determine the results:

If:
--it's a balancing BMS
and
--the balancing shunting current capability is higher than the actual current rate
and
--the BMS has turned on all the cell balancers while this current is flowing
then the cells won't receive any charge current and won't continue to charge, and so won't overcharge.

If any of those are not true, then the cells will continue to charge and have the potential to overcharge, or to be damaged from things like lithium plating processes (which I don't have any info on, but a web search will turn up assorted discussions and info about this) that can alter the cell properties and capabilities, and/or damage the cells over time.

The more perfectly the output voltage of the charge source matches the full charge voltage of the pack at cell HVC, the less current could possibly flow assuming all cells are perfectly matched in characteristics, and thus balanced in voltage.

The lower the output voltage of the charge source relative to the cell full charge limits x number of cells (assuming perfectly matched cells), the less likely any overcharge event could be even if the BMS HVC doesn't work.

The less perfectly matched the cell characteristics are, the less balanced the cells will usually be, and the more likely some cells will still be charging while others are overcharging (depending on balancing shunt capability of the BMS). (cell matching also affects other pack properties and capabilities, but not relevant to this discussion).



Note that any FET-based BMS can't *completely* shut off *all* current flow; FETs are imperfect switches that are just very high resistance when off, and so *some* tiny current could still flow. A contactor-based BMS doesn't have this issue; the air gap (or vacuum or neutral-gas gap) in the contactor is effectively infinite resistance for the charging source voltage, with no measurable leakage current in normal conditions.



Note also that any charging source that doesn't disconnect itself from the cells (via relay, etc) and isn't manually disconnected after charging means that when it is not powered from the wall it's output stage will then have some leakage current back into it from the pack, draining it at some (probably very small) rate continuously. Any LED output indicators will add to this drain, if they are powered at the output itself. In your usage this is unlikely to be an issue because the only times the system will not be powered from the wall it will be being discharged at a much more rapid rate by the load it's supporting until the load can shut itself down (computer, etc). If you find it is a problem, you can add a diode to the output of the charging source to mitigate backflow current.




All that said: I've used a Meanwell LED PSU HLG-600H-54A built into the SB Cruiser trike for charging for years, and it has no shutoff, just CC limiting with a set max voltage lower than the max charge voltage of the 14s2p EIG C020 pack in the trike. It's not terribly uncommon for me to leave it on the charger for up to several hours after it's full, as I'm usually too wiped out from work to go unplug it at the proper time. There could be lithium plating damage to the cells; I don't have a way to determine this; but the cells are so old (probably a decade and a half?) that simple cell aging effects exceed any problems created by this, as the packs that have not been regularly used or charged have about the same properties and capabilities as the one kept on the trike. I don't have a backflow-diode either, so the output LED of the potted Meanwell is always on whenever the trike is on (which connects the pack to the charger and controller, etc; my main disconnect completely isolates the pack).

 

[No_Shorty]

Thanks for the thorough replies.

@harrisonpatm - I am planning to test the BMS when it arrives, set it's upper charge limit and check that if the power supply/charger exceeds this if it will correctly cut the power, I couldn't find any proper documentation for the BMS, so I was assuming this would be a 'standard' function.

13s because it best fits the voltage range of my application. It needs to be above 40v and below 50v, so I am planning on limiting the upper and lower voltage limits in the BMS to this.
The reason for this is that I am using a few ideal diodes to act as source switchers, and to have the UPS battery as the lowest priority source it needs to be below the voltage of the other sources, whilst also being higher than the minimum input voltage of a Meanwell DCAC inverter (NTS-400P).

Here is a basic diagram of the design:



@amberwolf
Thanks, for reference it is the 20a NMC version of this BMS: https://www.aliexpress.com/item/1005006347440227.html?
primarily chosen for its compact size. Although I actually may have room for something larger if it is inadequate. It does claim a balancing function, but the documentation is very much lacking.

As mentioned above, I actually plan to limit the charge voltage to 3.85v per cell, so I guess this should provide me an extra layer of safety as the cells would have to be seriously out of balance to be overcharging from this source.
I am more than happy to buy higher quality or additional components to ensure the safety of the pack when charging, if there is anything I should add beyond what I have already mentioned.
The power supply - the RD RK6006 was chosen as RD seem to be a relatively well respected brand (for a Chinese vendor) in reviews on YT and such, and their DCDC PSUs are very easy to configure. If there was anything comparable from a higher quality brand (like Meanwell, or even other brands like TDK etc) I would be happy to consider. I haven't found anything remotely comparable in terms of size/ease of use.

Re: FET vs Contactor - I have no idea which my BMS is... but if I am making total guesses would I be right to assume that the cheaper brands like JBD use FETs? Can I ass a contactor inline to the BMS or does it need to be controlled by the BMS to allow the charging current?

The system actually would be regularly disconnected from source power (only in use during the day, disconnected overnight). But as you can see from the design above, the system is based around Ideal Diodes, so there should be no back-flow, the only thing that may have any drain would be the DCDC charging circuit, I'll test this when I have it in and add a diode for this as well if it's problematic. Thanks for the tip there!

edit:
Also are there any good resources on the forum regarding expected cell temperature at certain discharge rates?

I'm planning on using Eve 18650 25p cells. Which can support 20a discharge - but should I expect significant heat generation at this discharge? It is unlikely I will actually be drawing this much... but I would like to take it into my design consideration. There will be forced airflow through the UPS, so I can design the pack to take advantage of this.

 

[harrisonpatm]

It sounds like you have a well-thought-out plan, and I personally would feel comfortable implementing it as you have described.

As mentioned above, I actually plan to limit the charge voltage to 3.85v per cell, so I guess this should provide me an extra layer of safety as the cells would have to be seriously out of balance to be overcharging from this source.

In mu opinion, this is going to be your biggest element of safety. It's going to be hard to ever overcharge a battery that doesn't even come close to full voltage. This is a good idea.

I'm planning on using Eve 18650 25p cells. Which can support 20a discharge - but should I expect significant heat generation at this discharge?

Yes, it is possible that they would heat up if you actually take them to this discharge. You said you'd be doing 2p, meaning continuous discharge of 40A. And I found a couple tests online suggesting that these cells can do 30A, but when they do that they get up to 80 degrees C and experience massive voltage sag. Just things to keep in mind.

Also are there any good resources on the forum regarding expected cell temperature at certain discharge rates?

I couldn't find a thorough test on that particular cell online, just some anecdotes about the 80 degrees temp at 30A on Reddit.

Re: FET vs Contactor - I have no idea which my BMS is... but if I am making total guesses would I be right to assume that the cheaper brands like JBD use FETs? Can I ass a contactor inline to the BMS or does it need to be controlled by the BMS to allow the charging current?

You could always add a contactor into the circuit, either controlled by the BMS or otherwise, as long as you're okay with the current draw that an always-on contactor will have.

 

[No_Shorty]

It sounds like you have a well-thought-out plan, and I personally would feel comfortable implementing it as you have described.


In mu opinion, this is going to be your biggest element of safety. It's going to be hard to ever overcharge a battery that doesn't even come close to full voltage. This is a good idea.


Yes, it is possible that they would heat up if you actually take them to this discharge. You said you'd be doing 2p, meaning continuous discharge of 40A. And I found a couple tests online suggesting that these cells can do 30A, but when they do that they get up to 80 degrees C and experience massive voltage sag. Just things to keep in mind.


I couldn't find a thorough test on that particular cell online, just some anecdotes about the 80 degrees temp at 30A on Reddit.


You could always add a contactor into the circuit, either controlled by the BMS or otherwise, as long as you're okay with the current draw that an always-on contactor will have.

Thanks, appreciate the vote of confidence, as I always worry I have missed something obvious!

I have a load tested which can pull 600w on the way from AliE, so I'll test out the batteries at maximum drain. The Eve25p specs actually do say they can provide a max of 30a in the spec sheet, but most websites list their max drain at 20a, which is all I need in any case.

Would a contactor drain current when the system is disconnected from power? If so that may be a problem, as I wouldn't want the UPS battery to be dead if it had been sat without use for a while.

 

[harrisonpatm]

Would a contactor drain current when the system is disconnected from power? If so that may be a problem, as I wouldn't want the UPS battery to be dead if it had been sat without use for a while.

If the contactor is engaged/on/closed, it's draining current. How much current and when it's on is something you'll have to find out for your use case. You can always add a toggle switch on one of the wires leading to the contactor's coil.

 

[dominik h]

You have to test your BMS with an overvoltage scenario. You need to be shure what happens when your DCDC sends 60V to your battery.

 

[No_Shorty]

You have to test your BMS with an overvoltage scenario. You need to be shure what happens when your DCDC sends 60V to your battery.

Yes, certainly planning on it

I think I almost certainly have space for a bigger BMS, so if this one isn't up to it I will choose something that is.

 

[No_Shorty]

@amberwolf (or anyone actually)

Do you happen to know of a source where one can purchase a high quality ideal diode circuit? It's hard to tell the difference between the various ones on AliE, I would love to find someone making them with chips from Analog or TI or similar.

It's also really hard for me to understand the difference between something like this:

https://www.aliexpress.com/item/1005002603185232.html

Which is nice and compact, and very cheap.

And something like this: https://www.aliexpress.com/item/32983160620.html
Which is significantly more expensive. I am tempted to spend more just because it tends to trend with quality on AliE... but not sure if I am just paying for something else I don't need. I can obviously see it has 2 ICs... I guess to increase current capability?

 

[amberwolf]

Thanks, for reference it is the 20a NMC version of this BMS: https://www.aliexpress.com/item/1005006347440227.html?
primarily chosen for its compact size. Although I actually may have room for something larger if it is inadequate. It does claim a balancing function, but the documentation is very much lacking.

JBD should have docs / etc on their website if for some reason they're not on the AE page. Won't guarantee they're very good, as documentation in general regardless of product source or type is not very good:

jiabaidabms.com

Re: FET vs Contactor - I have no idea which my BMS is... but if I am making total guesses would I be right to assume that the cheaper brands like JBD use FETs? Can I ass a contactor inline to the BMS or does it need to be controlled by the BMS to allow the charging current?

JBD makes both kinds.

If it doesn't specify it's contactor based, it's almost certainly not.

But FET based types can be modified to use a contactor; very easily if you don't need to monitor current thru the BMS, less easily if you do. Some info on that is over here on the contactor sale pages of BatteryHookup:

https://batteryhookup.com/products/te-connectivity-ev200aaana-500a-0-900vdc?_pos=1&_sid=5f1be6401&_ss=r

https://batteryhookup.com/products/2x-9-36v-200a-contactor-relay-1x-350a-fuse-module-bmw-4596753-03

Also are there any good resources on the forum regarding expected cell temperature at certain discharge rates?

Depends on the specific cells used, but the spec sheet for the cells should have this listed on it.

Personally I wouldn't want to use cells that increase temperature noticeably when used within their spec sheet limits; it means that their resistance is too high for that application and capacity and power are being wasted as heat instead of doing work.

Many sellers' ads provide ratings for cells (or entire batteries) that are at or above the actual maximum limits of the cells. Sometimes they may just claim a continuous current when the specs show that's only a momentary peak (that is not intended to be repeated frequently or soon).

I don't know how Eve actually rates their cells; if they rate them for room-temperature operation or if they rate them at various temperatures, or if they rate them such that they will generate a lot of heat when used outside the median rates, etc.

 

[amberwolf]

Do you happen to know of a source where one can purchase a high quality ideal diode circuit? It's hard to tell the difference between the various ones on AliE, I would love to find someone making them with chips from Analog or TI or similar.

You can build your own if you like; somewhere here there's at least one thread (started by Tiberius, I think, and possibly another by Methods, and probably one by Fechter), on how to build these with FETs.

Cautionary "rant" below:
I don't know which prebuilt ones would be better, but I can nearly guarantee that whatever you find on Aliexpress is less likely to contain genuine parts from any specific manufacturer, if that's something you need. The pictures on any particular ad may not be of the item you will receive; you can't know what it is until you get it and test it compared to the part manufacturer spec sheets (if you have equipment to do that).

Usually it will at least "work", but it may not handle as much voltage or as much current or as much heat, etc., as the actual parts would have--or the board design may be missing peripheral components that would help reduce noise, or other things that would make the product better-performing. (there might be empty pads to put the parts on, or the design may have left them out entirely from the beginning).

Most things I've gotten off AE (or amazon, ebay, etc) have done more or less what they said they would, but some of them were nothing like what was advertised.

If you *need* guaranteed-genuine parts, you have a much better chance going thru Mouser or Digikey, Farnell, Avnet, or some other source like that--much more expensive both in parts cost and shipping, though. If you do run into a counterfeit problem, at least they might be able to do something about it--the other sources are unlikely to care.

It's also really hard for me to understand the difference between something like this:

https://www.aliexpress.com/item/1005002603185232.html

Which is nice and compact, and very cheap.

And something like this: https://www.aliexpress.com/item/32983160620.html
Which is significantly more expensive. I am tempted to spend more just because it tends to trend with quality on AliE... but not sure if I am just paying for something else I don't need. I can obviously see it has 2 ICs... I guess to increase current capability?

WIthout specific part and design info they don't provide, I couldn't say for sure why one is different than the other--there could be 2 chips for lower resistance, for better current, less heating, or simply because the chips used are less capable than the single chip so they have to use two to do the same job, etc. (either because of lower specs, or counterfeit, etc).

Whether either one is *actually* capable of the specs it claims, you'd have to test until they blow up under the same conditions to find out....

But...if the chips they are using are the same (didn't look), then the one with more chips is spreading the load and heat over more area, if they're just paralleled, and if used at the same rate and conditions the one with more chips would be potentially more reliable. (assumptions...can't say any of it for certain).

 

[No_Shorty]

Thanks, all useful info. Circuit design is a little beyond me, even something relatively simple like this. So off the shelf stuff is much more appealing.

I did find this blog post with freely available designs: MOSFET – IDEAL DIODE

I may head to somewhere like Fiverr to get someone to look over the design and tailor it for me.

It's not so much that I 'need' name brand components. I would just like to use quality components where possible. If I can spend £50 on something I know is good I would sooner do that than spend $20 on AliE. I shop there a lot so I know the ups and downs, and suaully how to separate the good stuff from the bad (to some degree), but my lack of knowledge here is making that difficult!

 

[DogDipstick]

High end, distributed, current sensing, voltage reporting, BMS use chargers controlled by mechanical contactor relays and communications to ensure the BMS has full control of the charger ( 's supply) at all times.

 

[No_Shorty]

High end, distributed, current sensing, voltage reporting, BMS use chargers controlled by mechanical contactor relays and communications to ensure the BMS has full control of the charger ( 's supply) at all times.

Yes, I think they are likely a little beyond the scope and size of my project unfortunately!

 

[calab]

You dont discern quality from junky crap, whether it be the cc-cv or bms?

 

[memy]

I have not read this above in detail, but it the telecom space we do have some systems running 14S "floating" on the exchange voltage of 54V 24/7, usually loaded with the LG 1500ma cells. They need to carry a load for about 3min before the gen kicks in. Though the enginers running the show say lead is still cheaper for this function in the 15min timeframe.
Anyway don't know about the JBD BMS but i'm really likeing the JK BMS. Running one on 17cells of 280ah EVE 200A version for a few years. Bought another 40A version for the bike just because I like that I can set it to active balance at a particular voltage.

In the case of the the 17cell LFP anything above 3.45 and it will start sending power to the other cells. Not that it has got there as system voltage is limited to 58V(and I need more solar). But one day when the cells are showing their age it may come in usefull... If the BMS lasts that long.