[dtbaker]

I'm considering adding at least simple cell level BMS to prevent over-voltage charging... because I have a few cells in my pack that are 'drifting' out of balance after a few charge cycles.

I have pack that is 48 x 130ah CALB LiFePO4 large format prismatic cells.

I'm looking for whatever people have found, and used for a while, that can at least stop a charger when the first cell hits some high-voltage limit like 3.70v. It could even be as in-elegant as a signal to a separate relay that would cut power to the charger.

I did a little poking around the web last night and found that MiniBMS is no longer sold standalone... but there are a number of chinese units that look like they MIGHT do the job except they are all intended for 100amp max thru them and I can't really tell if they would work since my Zilla controller could be pulling 1000amps. My charger is just a little 1500 watt Elcon, so 8amps is about all the charge current the pack sees. I don't really even need active shunting, just a HV sense and power off would be fine.

I found things like this:
https://www.ebay.com/itm/4S-12V-100A-Balance-LiFePo4-LiFe-18650-Battery-Cell-BMS-Protection-PCB-Board/152854161901
but they look like they are intended purely for 12v (4 cell) packs

Can any of the EE types here tell if these units would be capable of watching 4 cells at a time, and sending a signal I could use to trigger a relay if any single cell hits the high-voltage limit? I can't tell if this type of BMS could be used just to watch voltage and not worry about balance shunting or passing full current load.

I am just looking for the cheapest dumbest way to watch 48 cells in series, and turn off the charger as soon as the first one hits 3.70v

 

[john61ct]

Build it off an Arduino?

 

[dtbaker]

Build it off an Arduino?

I am looking for something off the shelf, or with complete parts list and instructions..... I'm mechanical, not an EE.

 

[MattsAwesomeStuff]

because I have a few cells in my pack that are 'drifting' out of balance after a few charge cycles.

Not to beat a dead horse, but, again... you do not have cells drifting out of balance. You have pack with cells that no longer have matched capacity due to damage.

Your only solution is to make the capacities match. No amount of balancing will solve a capacity issue.

I know you keep insisting you don't want to measure the capacities because you don't have time, but, that is your only solution.

You can't "balance" a weak cell to be larger again.

 

[dtbaker]

Not to beat a dead horse, but, again... you do not have cells drifting out of balance. You have pack with cells that no longer have matched capacity due to damage.

Your only solution is to make the capacities match. No amount of balancing will solve a capacity issue.

I know you keep insisting you don't want to measure the capacities because you don't have time, but, that is your only solution.

You can't "balance" a weak cell to be larger again.

I can't change the capacity either. So, at this point it doesn't do me any good to measure either the capacity or the IR as there is nothing I can do to 'fix' the damaged cells. I cannot replace them at this time, so I am exploring ways to live with them and prevent overcharge as the short capacity cells will doubtless 'drift' higher and higher at end of charge.

I can (hopefully) catch the first cell that hits 3.7 and end the charge. Then, periodically top balance and go some more before I have to replace a whole bunch of cells.

The reason I continue to refer this as balance, or drifting, is that is an accurate description of the observed behavior I am trying to live with. I top balance all the cells, and after a couple charge cycles the damaged cells start finishing significantly ahead of the the others... i.e. 'no longer top balanced' pretty much defines the drifting voltage at the end of charge.

Why do you continue to tell me I don't understand it?

Why don't you accept that I am trying to find a way to live with the slightly mismatched cells in situ for now?

 

[john61ct]

Either increase your skillset, reduce your expectations, or spend a lot more money.

This is not a mature area of technology for DIY efforts.

 

[dtbaker]

Either increase your skillset, reduce your expectations, or spend a lot more money.

This is not a mature area of technology for DIY efforts.

I think I am clearly reducing my expectations.... All I want is to know if anyone has a simple inexpensive solution off the shelf that can watch voltages on 48 cells, and send a signal to a relay I can use to cut power to the charge.

If you don't HAVE a solution, why spend your time with a meaningless unhelpful reply?

 

[john61ct]

No one will be more pleased than I, if it turns out that what you are looking for actually exists.

If it turns out my suggestion is the best way to go within your budget, you may find it takes just a couple of enjoyable evenings to figure out how to do it.

I am only motivated by trying to be helpful, and (constantly) climbing my own learning curves.

 

[dtbaker]

No one will be more pleased than I, if it turns out that what you are looking for actually exists.

If it turns out my suggestion is the best way to go within your budget, you may find it takes just a couple of enjoyable evenings to figure out how to do it.

I am only motivated by trying to be helpful, and (constantly) climbing my own learning curves.

I am hoping that one of the more EE oriented people in the forum can see a way to jigger one of these guys:

https://www.ebay.com/itm/4S-12V-100A...d/152854161901

to a simple relay that would shut down the charger when the first cell hits 3.70 .

 

[pm_dawn]

Hi !

rwaudio has his Cellog8 breakout boards for sale again.
That could be an idea.

But be aware of the unbalaced load of the Cellog8.


Regards
/Per Eklund

 

[MattsAwesomeStuff]

I can't change the capacity either.

Sure you can. For quite cheap you could buy some LiFePO4 18650s (or even regular lithiums if they're more available, and just cut your expectations in half) and parallel them to the weakest cells. You could build a little booster box, since presumably you already have some kind of BMS onboard, you'll already have all those sense wires broken out.

I top balance all the cells, and after a couple charge cycles the damaged cells start finishing significantly ahead of the the others.

So, what's happening there is quite likely that the degree to which they are unballancing is directly the amount which their capacity is being murdered every single time they're used. The weak cells are continuously being amputated and permanently losing capacity an an accelerating rate.

This will continue until you add some capacity.

Why don't you accept that I am trying to find a way to live with the slightly mismatched cells in situ for now?

It just seems like you're labeling and attempting to fix the wrong problem the wrong way. I kinda feel shitty for you since your solutions are continuing to harm your batteries.


...

I have ghetto solution you might not like...

1 - Go to Dollar Tree and purchase 48 cell phone chargers for $48.
2 - Buy 50x TP4056 lithium battery chargers ($12.50 total).
3 - Disassemble the phone chargers and gang them up, or get a bunch of powerbars and another $48 worth of micro-USB charging cables.

That will allow blind, independent, electrically isolated charging of each cell in addition to your normal charger, at a rate of about 5 watts per cell. So, slowly, but, it will creep them higher overnight.

Or you could just guess how many are bad and only install boosters for those cells.

Err... except you're using LiFe, not LiIon, so the TP4056 might need modification of a resistor to change the termination voltage.

Well, some kind of solution like that might be what you want.

All the shunt-type BMSs will overcharge your good cells because they can't handle full charging current in the rebalancing (they can't bleed off enough of the power going to the full ones).

 

[dtbaker]

Sure you can. For quite cheap you could buy some LiFePO4 18650s (or even regular lithiums if they're more available, and just cut your expectations in half) and parallel them to the weakest cells. You could build a little booster box, since presumably you already have some kind of BMS onboard, you'll already have all those sense wires broken out.

I do not have room to create independent 'boosters' for weak cells... What I am after is how to use the pack I have a little longer without adding or replacing batteries.

Hence, just a simple sense-and-terminate system capable of sensing the first cell to 3.70 +/- .05 volts and pulling the plug on the charger regardless of the state of charge of the rest of the cells.

All the shunt-type BMSs will overcharge your good cells because they can't handle full charging current in the rebalancing (they can't bleed off enough of the power going to the full ones).

exactly

All I need is something like an external box that I plug my charger into that has a relay inside that does a hard stop if any of the cells signal voltage over 3.70. I am not looking for something smart enough to shunt current around full cells and 'fill' the pack completely. I'm just trying to find the cheapest way to limp along with the pack I have and not murder any more cells from over-voltage at end-of-charge.

 

[dedlast]

dtbaker, I am finding it rather amusing that two people with barely 200 posts between them are lecturing you on what you should do and how to do it. That seems a bit like someone trying to tell Major he doesn't know motors or Tesseract that he really should stop talking about controllers until he learns a thing or two. But that's just my opinion...

 

[dtbaker]

dtbaker, I am finding it rather amusing that two people with barely 200 posts between them are lecturing you on what you should do and how to do it. That seems a bit like someone trying to tell Major he doesn't know motors or Tesseract that he really should stop talking about controllers until he learns a thing or two. But that's just my opinion...

I certainly do not put myself in a class with either Major or Tesseract... but I find the noise to signal ratio high in the responses to my recent posts.

I have built 2 cars, driven more than 30,000 electric miles on LiFePO4 packs and I think I have a decent understanding of their behaviour and at least the external observable actions and reactions of an aging pack that has suffered the occasional inevitable under or over charge. I chose to go without a BMS originally, just as I chose to top-balance, after much consideration... I don't need to a lecture on how those choices were wrong; and AM looking for some informed, credible, and direct input though.

I am now entering the 'how do I best deal with an aging pack' that won't stay top-balanced at end of charge without increased manual checking and tweaking. top-balanced to me is defined as when all cells end up close to the theoretical average volts/cell given your charger end-of-charge setting. 'Not balanced' to me is when one or more cells goes significantly over the pack average before the pack voltage ends the charge.

In the early stages, I found the packs stayed very well balanced for hundreds of charge cycles.... no cells 'drifting' up or down much from the pack average at end of charge. This behavior changed a little after the first under-voltage 'event' that killed one cell under load. And changed more after the second 'event' about a year later. Hence my recent posts as I try to come up with a way to deal with the issue (hopefully avoiding massive expense for at least a while).

I really don't want to argue about semantics of WHY or HOW cells get damaged, lose capacity, or increase internal resistance. All I am interested in at this point is how to keep cycling this pack and driving while minimizing manual intervention and at a minimum time and expense.

My best bet at this time appears to be to find, or build, a simple cheap way to monitor all cells during charge, and stop the charge if any cell goes above 3.70 before the pack voltage triggers a 'normal' end of charge to the charger.

so.......

If anybody has any specific answers to this, let's hear it!

 

[Russco]

Hi !

rwaudio has his Cellog8 breakout boards for sale again.
That could be an idea.

But be aware of the unbalaced load of the Cellog8.


Regards
/Per Eklund

I use the CellLog8M to shut off my charger when the first cell reaches 3.500 volts.

Unfortunately, the CellLogs have been discontinued.

Too bad, they work really well.

Dimitri has moved to California and discontinued his Clean Power BMS. A simplified version would work for you.

Seems like products for the DIY conversion crowd are getting hard to find.

Sorry, Dan, not much to choose from. It takes an electrical smarty to design something. Probably not much market for that.

You could build up 45 little circuit boards with an LM431, a FET opto isolator, a small transistor and 3 resistors. Daisy chain the boards like the Clean Power setup and when any cell hits your 3.7 volts, the opto opens and latches a power relay to shut off the charger. Pretty simple but sort of a PIA to make up 45 units.

That's how I would do it.

 

[Russco]

You could build up 45 little circuit boards with an LM431, a FET opto isolator, a small transistor and 3 resistors. Daisy chain the boards like the Clean Power setup and when any cell hits your 3.7 volts, the opto opens and latches a power relay to shut off the charger. Pretty simple but sort of a PIA to make up 45 units.

That's how I would do it.

Or, take a JLD404, which every conversion should have, and set the alarm for VOLTS and set the number for (3.7) (45) 166.5 volts. Trim the volts as required a little down until the first cell hits the 3.7 volts. The alarm relay contact of the JLD can be wired to latch a power relay to shut off the charger.

And that's another way to do it.

 

[nebster]

Zeva has an 8-cell monitor for $60 AUD. Six of those and you're good to go.

 

[MattsAwesomeStuff]

I am finding it rather amusing that two people with barely 200 posts between them

I find it funny that, there's really nothing lamer than a pissing contest based on the number of posts they've made, or valuing anyone's opinion based on how much they've spoken. Especially on a particular forum with a particular account. Or how old their account is. Or some other clique-y but meaningless measurement. I certainly don't judge anyone that way, seems it's the entirety of how you do.

I care more about how much someone listened than how much they speak.

I don't need to a lecture on how those choices were wrong

I haven't seen anyone do that. I don't think anyone's criticizing you for your choices.

I do not have room to create independent 'boosters' for weak cells.

You just need some empty space... anywhere in the car. They don't have to be side-by-side. You can locate them all in another box and just drag wires back to the pack. But, if you don't want that solution, that's okay.

My best bet at this time appears to be to find, or build, a simple cheap way to monitor all cells during charge, and stop the charge if any cell goes above 3.70 before the pack voltage triggers a 'normal' end of charge to the charger.

I don't think your problem is on the charging side, your problem is on the discharging side. Lithiums, especially LiFes, are decently tolerable to over-charging. I've got some I've regularly thrown into normal lithium chargers, (cutting off at 4.2v instead of 3.6) and it's not doing them any harm.

What murders Lithium is the reverse charging that happens when you drain them past dead. That's what you need to be extra sure you're not doing.

The problem with doing what you want is it's probably as much effort to do that, as to actually fix it.


The easiest thing would be the relay trip, you've already got that figured. There's probably already a contactor in there to do that so just hijack it an insert another switch, or, interrupt the input side.

The difficult part is that you need 48 different voltage readings, and some brains overtop that to control it. That's 49 different wires to the cells. If you're building a circuit, that's 48 different circuits to do things to, and you seemed to indicate earlier you weren't even interested in the effort it would take to do a discharge test to find out which cells were low. So I'm trying to picture what kind of solution is cheaper than the couple hundred 18650s you could buy to fix the capacity issue, and also take less time than doing a simple capacity test. There's not much on the table with those conditions.

But in any case...

Maybe something like 48x 3.7v zener diodes, in series with a NC relay? Then when any of them conduct (when any one of their voltage cross 3.7v), it activates the coil and breaks the relay? It'll pulse and flicker, maybe add the few components to make it a one-shot-stop, like emergency pushbuttons on industrial gear?

On the discharge-side, where I think your actual problem is, would you be okay with a human solution? I.E. If you could just occasionally keep an eye on something and manually know not to keep driving?

In that case, I'd drill an 8x6 grid into some plastic and insert red LEDs. Have each of them wired to their own cell with some kind of resistor that yields barely enough current to light the LED when at your minimum voltage is.

So when you're driving you'll see the LEDs brightly lit when charged, they'll all slowly dim (but barely, it's pretty digital for most of its range), until they get near the bottom of their range and then they'll dim much more rapidly and eventually go out. If you ever see one go out, you're done, stop the car.

 

[Duncan]

Hi
How much of your capacity do you NEED?

If you have got some degree of excess capacity the first thing to do would be to stop charging and discharging a bit earlier

Next - you have been monitoring your cells so you should have a good idea which are the (most) dodgy ones - if you could put the dodgy ones together then you could put something like the old JLD404 across the weak sisters and set it to stop the charge and give you a warning buzzer when they get too low

 

[jhuebner]

Have been using my homebrew monitoring system for 5 years now. http://www.diyelectriccar.com/forums/showthread.php?p=361287#post361287


BMS simply cuts out the charger when first cell reaches 3.65V. No balancing whatsoever.



Not really planning to offer it right now but just posting it as an example that a car can run just fine on a heterogeneous pack.


Some of the 40Ah cells only have 30Ah left in them because they were deep discharged. So what, still 15kWh left. As I'm fine with the range there is no need to swap cells or add booster cells in parallel.