Cheap LiFePO4 BMS?

[Slow and Steady]

1 hour ago, nicknorman said:

On the travel power question, the whole issue of alternator overheating disappears if you are charging by Travelpower and a charger (assuming the leisure 12v alternator is disconnected or otherwise disabled). What may not be so easily addressed is control over the charger. Depending on the model of charger it may have a lithium mode but if not you would need to disable voltage temperature compensation. You would probably want to be able to adjust the charging voltages to suit. Of course it depends on how long you intend to cruise for but it might be an issue that a long cruise would hold the batteries at a high voltage for longer than desirable. On the other hand you might be able to configure the charger to go into a suitable float voltage that doesn’t push any more current into the batteries once they are charged. So many variables! But overall yes I think you would be starting from a better place than someone who just has a 12v alternator.

Thanks Nick, that's encouraging. It's a Victron Pheonix MultiPlus 12/3000/120 Charger/inverter "VE.Bus enabled" 

It seems like quite a thing. I downloaded the control software which allows you to "fake" a target rather than connect it up to the computer which I did so I could look around the program. The beast itself is 55ft away at the exact opposite end of the boat though I have a laptop I might be able to persuade to work for this when I do it in anger.

Then I went to the charger setting page and yes it's rather good. At a glance I can put in any values for anything plus it does list LiFePo4 batteries which it thinks should be charged like the following picture. I can also create a new battery type and stick in any numbers I like. It would charge at up to 200A [edit - no, of course not - 120A, but hey, that's pretty good!] if the TravelPower could cope but I suspect it couldn't deliver that. All in all it seems quite promising, even has a low temp cut off! The bumph also claims it's exceptionally good as a solar controller though I somehow doubt it could do both at the same time, that would be just TOO lucky., though there could be an option to somehow switch between the two if I make config files for both... pure conjecture at this point.

Edited January 10, 2022 by Slow and Steady

[Slow and Steady]

Ref the above, I now noticed it's given me config options for a more posh piece of kit. I need to find a fake file for my actual inverter charger which maybe (probably) won't be as configurable or capable - that would be more my luck. Oh well, got excited there for a minute!

[nicknorman]

11 hours ago, Slow and Steady said:

Ref the above, I now noticed it's given me config options for a more posh piece of kit. I need to find a fake file for my actual inverter charger which maybe (probably) won't be as configurable or capable - that would be more my luck. Oh well, got excited there for a minute!

I think your Combi doesn’t have a lithium mode, but you could still adjust the voltages to be suitable. The remaining problem is temperature compensation of charge voltage. Maybe you can set that to zero but I don’t think so. You can disconnect the temperature sensor but then you’d need another method of monitoring battery temperature and avoiding charging at low temperatures.

 

Oh and also, 120A at say 14v is about 2kw allowing for charger efficiency. A Travelpower can supply this at idle, but running the engine slowly with that heavy load on the crank pulley and belt is a very bad idea, so you’d want to run the engine a bit faster maybe around 1200 rpm minimum.

Edited January 11, 2022 by nicknorman

[nicknorman]

Well now … my lovely Chinese lithium cells (all 600Ah of them) which were just perfect, all controlled by my smart charging system … are no longer perfect 😱😢. On this trip I’ve suddenly found that when I take the cells up to near 100% to synchronise the BMV, one bank of cells hits the “knee” way before the rest, like about 5% SoC before which is 30Ah. The automatic balancing has been kicking in but since it only takes about 2A for 30 mins or so, it hasn’t dented the imbalance. But oddly, if I repeat the 100% charge cycle, after several eventually all the cells go up the knee together just like they should do.🤯
 

Apparently one is suppose to deep cycle the cells every 6 months or so to wake them up. So I disconnected the alternator for a whole day, tried to run the batteries down by various means, lots of electric kettle, toaster, and even, at 30% SoC, doing a hot wash which still left 10% so tumble drier on the 1kw setting, and FINALLY managed to get down to about 3% SoC at which point the cells were dipping down to about 2.8v - nicely balanced, incidentally. That’s low enough, so charging back on. It takes a while to put in 600Ah!

 

But this evening the rogue cells hit 3.65v at about 90% SoC, even worse than before, with 100% SoC target set, charging terminates at that point. 😱😢

 

Much fretting about which of the 3 200Ah cells in the group it might be, so after tying up this evening I poked them, talked to them, waved a spanner at them but they weren’t having it even after I tightened their nuts. Even checked that my BMS was reading correctly, which it was.

 

After engine shut down for 30 mins, all the cells are nicely balanced again. Started the engine again, set to 100% SoC, voltage on the rogue set once again rising towards 3.6v (75A charge) so out with the meter to probe the actual terminal posts of the 3 cells. 3.6v, 3.6v, 3.4v. Hmmmm, how can that be, the are all connected in parallel with fat juicy copper bus bars? Well, they should be anyway. The bolt which threads into the post has the higher voltage on it, there is nearly 200mV differ nice between the bolt and the actual terminal. After switching off, the voltage difference persisted for a while and then gradually fell away to zero as the cells equalised themselves in slow time. So it is just a bad connection. Phew! But I did take a lot of care to clean up the connecting surfaces and the bolt is tight. I’m about to take it to bits but hopefully, cleaning it all up will fix the problem. Just goes to show that having an accurate (to 3 decimal places of a volt) readout of cell voltages, reveals problems before they become much of a problem.

 

Picture shows the problem area, the voltage on the post (black arrow) is not the same as the voltage on the bolt or copper bus bars (red arrow) during charging. Which ain’t right!

 

 

 

 

 

 

[BEngo]

Soinds like one of more of the terminal bolts are at the Lower limits of size. (Bolt and female thread  machines tend to wear so they produce undersized threads.).   Might be worth talking to your friendly local fastener supplier to see if  they can locate either some close tolerance bolts or some plated bolts, both of which will be nearer maximum size and so make better thread to hole contact.   I cannot remember the metric terminology  and my books of  words is  at home  but I am sure a good fastener stockist will know what you are asking for.

 

As an interim you could try some conducting grease on the threads- perhaps Coppaslip or similar.

N

 

 

 

 

 

[Craig Shelley]

Get polishing!

Hey, we've had a similar issue. It was especially visible on the temperature readout. From memory, the CALB manual does say that the terminals need to be polished each time the pack is reassembled. I think our issue started when we striped the pack down to relocate it. The terminals looked good enough so we just put it back together.

The best method we found in the end was fine wet and dry paper backed with a block of wood to ensure the polished surface is flat.

The copper at the centre of the terminals becomes a mirror finish. Same goes for the straps.

[Tony1]

 

I must admit, I've never considered cycling my lithiums periodically down to 2.8v (say 5% SoC). 

With all the solar I get in, I usually cycle them between 45 and 75%, with the odd run down to 20% and sometimes up to 100% to synchronise- so they'll be practically comatose by now (although to be fair, they do seem to carry on performing exactly the same, day and day out). 

But on a more technical note, looking at your diagram, I think its clear the electrical pixie density on the post is lower than it should be, but that is entirely predictable. 

You've not sacrificed a single gerbil since you got the cells.  

What's needed here is practical solutions, so my recommendation for this issue is to slaughter a brace of gerbils on the next full moon - and the more grisly the killing, the better they like it. 

 

[nicknorman]

Slightly more complicated than I thought, I’ve just removed the connections from that post and realised that although they are the same colour, the hex bit is actually threaded onto the central bit. The voltage difference was still detectable between the central bit (where the bolt threads in) and the outer hex bit. Presumably it is the latter which is connected to the internal gubbins. Which explains how the bolt was at a different voltage from the cell terminal outer hex bit.

 

I’m sure I read somewhere warning instructions “do not attempt to tighten the hex nut” but anyway, I nipped it up a fraction and now all is sweetness and light … well except of course that the frantic attempts by my BMS to top equalise the cells means that this bank of cells is now a bit low, so the system will have to sap charge out of the other 3 cell banks. Never mind, it will do it eventually!

 

So I’m a happy bunny but the moral is that not only is a bad connection possible, but also it might be in the cell, not in the wiring. If I get a sudden unexplained cell imbalance next time, I’ll know what to look for.

[Craig Shelley]

I just noticed that you have some additional lugs for your instrumentation on the 'dodgy' terminal. That's something we considered doing, but didn't like the idea of because it jacks the bolt out of the battery terminal, and causes the mechanical pressure of the bolt to be applied to a small area on the strap.

In the end we ended up going for crocodile clips for the instrument wires, which certainly aren't great.

[nicknorman]

2 minutes ago, Tony1 said:

 

I must admit, I've never considered cycling my lithiums periodically down to 2.8v (say 5% SoC). 

With all the solar I get in, I usually cycle them between 45 and 75%, with the odd run down to 20% and sometimes up to 100% to synchronise- so they'll be practically comatose by now (although to be fair, they do seem to carry on performing exactly the same, day and day out). 

But on a more technical note, looking at your diagram, I think its clear the electrical pixie density on the post is lower than it should be, but that is entirely predictable. 

You've not sacrificed a single gerbil since you got the cells.  

What's needed here is practical solutions, so my recommendation for this issue is to slaughter a brace of gerbils on the next full moon - and the more grisly the killing, the better they like it. 

 

They do say that some capacity is temporarily lost if the batteries are not exercised. But then again, if you have more capacity than is needed, it probably doesn’t matter. I mostly did it because I couldn’t understand how one bank of cells was suddenly topping out before the others, but I now know it was nothing to do with “lazy cells”.

[nicknorman]

1 minute ago, Craig Shelley said:

I just noticed that you have some additional lugs for your instrumentation on the 'dodgy' terminal. That's something we considered doing, but didn't like the idea of because it jacks the bolt out of the battery terminal, and causes the mechanical pressure of the bolt to be applied to a small area on the strap.

In the end we ended up going for crocodile clips for the instrument wires, which certainly aren't great.

There are longer bolts on the terminals for this reason. I measured it all carefully to ensure they didn’t quite bottom out. The extra wires are one for the BMS and one for the balancing board. But yes you are right about the second point. However, there was no detectable voltage difference between the straps and the bolt even at 200A so I think it is a problem only in theory. And it could be argued that the smaller area = higher point force which could be a good thing.

 

[Tony1]

52 minutes ago, nicknorman said:

They do say that some capacity is temporarily lost if the batteries are not exercised. But then again, if you have more capacity than is needed, it probably doesn’t matter. I mostly did it because I couldn’t understand how one bank of cells was suddenly topping out before the others, but I now know it was nothing to do with “lazy cells”.

 

I shall plan to cycle mine down to about 11.4v some time in the next week. I dont think they've been below 12.7v for many months.

I think many lithium users are now au fait with the idea of cycling between say 20 and 80% SoC (or similar) as a routine practice, but I'm not sure how many have picked up on the idea of improving longevity by periodically doing almost a full cycle, say 5% to 100%. 

 

For a liveaboard there are some practical considerations to work around. One of the constraints I have involves the time of day that the batteries reach their lowest SoC.  I dont want the SoC or the voltage to get so low during the night that it sets off the low voltage alarm on the BMV712 during at say 4 am (or even worse, to get so low that it triggers the emergency disconnect of all the loads). 

 

So I set my trigger values very conservatively. 

The normal cycle is that the MPPTs will charge the batteries up to say 75%, and then the BMV 712 will switch them off. The next thing that should happen in an ideal world is they should discharge down to say 25% before the MPPTs switch back on again and start charging.

But if I were to get down to that 25% SoC level at say 7pm (by which time the solar charge is pretty much over for the day), it could be that by the next morning when the solar starts charging again, I'll be down to 10% SoC, which is lower than I really want to cycle them on a day to day basis. 

 

So what I tend to do is to set the MPPTs to switch on again at about 50%, because even if the SoC gets down to that trigger setting late in the day when there's no more solar, I know there will still be enough charge to get through the night. 

The other consideration is whether I'm going to need hot water early in the day. If I know that I need to wash myself, or perhaps some clothes, first thing in the morning, I'll adjust the BMV712 relay settings so that I allow more charge into the batteries during the afternoon before, and maybe let them get up to 85% SoC.

What that means is I'll have enough charge in the batteries that I can switch on the immersion heater first thing, and have hottish water within 40 minutes.  

But if I've let the batteries get down to 20% overnight (in the interests of better cycling), and the morning solar isnt great, then I cant use the immersion heater when I need to. 

 

So there are these day to day considerations that affect how you manage the solar charging process, and how you set the limits so that you dont overstress the batteries. 

 

Edited May 25, 2022 by Tony1

[Craig Shelley]

49 minutes ago, nicknorman said:

The voltage difference was still detectable between the central bit (where the bolt threads in) and the outer hex bit.

That's really strange.

I always assumed the hex part was just a nut to stabilise the terminal against the case of the cell.

So you saw a voltage drop between the two points arrowed below?

[peterboat]

Whilst we are all admitting our accidents or mistakes, I was under the steps where my 6 batteries live. I decided to check why my battery meter had been jumping around between 26.2 and 27.6 volts occasionally, anyway it transpired that one of my bolts securing the link leads was a fraction longer than the others and was bottoming out, without really pulling the link leads fully down...........oops. Now these are the factory bolts, however I only added these 2 batteries a couple of years ago so maybe they were for a different application originally?

[nicknorman]

1 hour ago, Craig Shelley said:

That's really strange.

I always assumed the hex part was just a nut to stabilise the terminal against the case of the cell.

So you saw a voltage drop between the two points arrowed below?

Yes. Although it was the + terminal where the material looks the same. You’d thing that the central terminal was the battery connection and the nut was just holding it all in place,  but it seems like it’s the opposite. To be specific, when charging the nut was at a lower voltage than bolt threaded into the central bit, and also of course at a lower voltage than the equivalent on the other 2 cells in the pack. Which surely means the nut had a good connection to the actual cell whereas the central bit didn’t?

[rusty69]

13 hours ago, nicknorman said:

Well now … my lovely Chinese lithium cells (all 600Ah of them) which were just perfect, all controlled by my smart charging system … are no longer perfect 😱😢. On this trip I’ve suddenly found that when I take the cells up to near 100% to synchronise the BMV, one bank of cells hits the “knee” way before the rest, like about 5% SoC before which is 30Ah. The automatic balancing has been kicking in but since it only takes about 2A for 30 mins or so, it hasn’t dented the imbalance. But oddly, if I repeat the 100% charge cycle, after several eventually all the cells go up the knee together just like they should do.🤯
 

 

 

 

 

Any chance this has been caused by vibration loosening the battery connections due to close proximity of the cells to the engine?

[nicknorman]

46 minutes ago, rusty69 said:

 

Any chance this has been caused by vibration loosening the battery connections due to close proximity of the cells to the engine?

The actual battery connections were tight, and the problem persisted after I slackened off the connections, wiggled them a bit, and retightened. If you read on you’ll see it was a battery issue, not an external connection issue. But whether the battery problem was caused by vibration, who knows? The engine is fairly smooth and the batteries are sitting on a thin layer of stiff foam, but of course there is still some vibration. Since the problem started after a long period of non- use (couple of months) I am more inclined to think it was corrosion on the threads between the cell’s nut and centre post.

[rusty69]

44 minutes ago, nicknorman said:

The actual battery connections were tight, and the problem persisted after I slackened off the connections, wiggled them a bit, and retightened. If you read on you’ll see it was a battery issue, not an external connection issue. But whether the battery problem was caused by vibration, who knows? The engine is fairly smooth and the batteries are sitting on a thin layer of stiff foam, but of course there is still some vibration. Since the problem started after a long period of non- use (couple of months) I am more inclined to think it was corrosion on the threads between the cell’s nut and centre post.

Yes, my poor wording. I meant the cell connection, not any connection made to the battery. Still a bit worrying if it is corrosion that the others won't do the same thing given enough time.

[nicknorman]

59 minutes ago, rusty69 said:

Yes, my poor wording. I meant the cell connection, not any connection made to the battery. Still a bit worrying if it is corrosion that the others won't do the same thing given enough time.

I don’t know really. They were (relatively) cheap cells bought direct from China so who knows what the quality control was like. The nut turned slightly with only moderate force so maybe it wasn’t tightened adequately at manufacture? Who knows!?

 

All fine and dandy this morning and at least I will know what the problem is if it happens again!

[cuthound]

Sounds like it was a slightly loose hex nut. Glad it is now sorted 

 

When I was working I always used to specify the use of a jointing compound for busbar joints & connections and then measure the milli-volt drop across them under design load during commissioning.

 

https://etechcomponents.com/products/prysmian-bicon-bx1-225-electrical-joint-compound-u7086401/

Edited May 26, 2022 by cuthound
To add a missing worm

[Craig Shelley]

19 hours ago, nicknorman said:

You’d thing that the central terminal was the battery connection and the nut was just holding it all in place,  but it seems like it’s the opposite.

Yes, it doesn't seem logical, unless the terminal post is fractured within the nut. Or your DMM is acting up. Contaminated probe tips?

I've read that the nuts can sometimes have thread locking compound applied, and therefore shouldn't be relied upon for an electrical connection. E.g when probing with a dmm.

It would be interesting to see a cut away diagram of the inside of one of these cells. It might shed some light on the mystery.

Thanks for posting... We'll bear it in mind in case anything similar happens to ours.

Regards,

Craig

[Craig Shelley]

5 minutes ago, Craig Shelley said:

It would be interesting to see a cut away diagram of the inside of one of these cells.

 

Interesting teardown photos and description:

https://liionbms.com/php/prismatic_cells.php

 

Youtube video CALB Autopsy:

 

 

Youtube video of CALB Resuscitation!

 

[MtB]

2 hours ago, cuthound said:

When I was working I always used to specify the use of a jointing compound for busbar joints & connections and then measure the milli-volt drop across them under design load during commissioning.

 

I was wondering if was due to Nick's failure to slather the battery connections in non-conducting Vaseline, as often recommended on here!

 

🤣 🤣 🤣

[nicknorman]

2 hours ago, Craig Shelley said:

Yes, it doesn't seem logical, unless the terminal post is fractured within the nut. Or your DMM is acting up. Contaminated probe tips?

I've read that the nuts can sometimes have thread locking compound applied, and therefore shouldn't be relied upon for an electrical connection. E.g when probing with a dmm.

It would be interesting to see a cut away diagram of the inside of one of these cells. It might shed some light on the mystery.

Thanks for posting... We'll bear it in mind in case anything similar happens to ours.

Regards,

Craig

The whole situation supports the hypothesis that the bolt and central bit weren’t well connected to the nut, and the nut was better  connected to the actual cell than the central bit (that latter point being the weird bit). I repeated the test several times, got zero (well about 2mV) difference with all the other cell posts, and 200mV with that one, with around 80A flowing into the group. So it’s not a big additional resistance but certainly enough to throw an imbalance when charging or discharging. When the cells were rested the whole thing slowly sorted itself out of course. And having rotated the nut about 10 degrees, all is now fine.

Edited May 26, 2022 by nicknorman

[nicknorman]

3 hours ago, MtB said:

 

I was wondering if was due to Nick's failure to slather the battery connections in non-conducting Vaseline, as often recommended on here!

 

🤣 🤣 🤣

Actually I did put a very thin wipe of Vaseline on the mating surfaces. But the mating surfaces weren’t the problem!

3 hours ago, Craig Shelley said:

 

Interesting teardown photos and description:

https://liionbms.com/php/prismatic_cells.php

 

Youtube video CALB Autopsy:

 

 

Youtube video of CALB Resuscitation!

 

hmmm, really hard to see how what I experienced could be valid. But it was!