Balancing LiFePO4s

[MtB]

I have grave doubts about the viability/capacity of my cell bank now I have the BG-8S. It was sold to me as a 24v 8cell 200ah bank. The cells are assembled up into a plastic case just as it was in the car.

 

On charging, cell 8 gets there before all the others and cell 1 lags behind. When cell 8 was up at 3.599v yesterday (during charging) cell 1 was still at 3.331v and the monitor claimed 61% SoC for the whole bank. I think it just returns the SoC for the lowest cell. This morning after drawing perhaps 10ah (at 24v) overnight, the SoC is reported as 21%. Will post two photos in a sec from my phone.

 

First is last night when I stopped charging, second is now at 9.30am following morning:

 

 

 

 

 

[nicknorman]

Hmmmm slightly worrying, but how accurate is the voltage measurement? I know it reports to 1/1000th of a volt (resolution)  but I bet the accuracy is much less. It is difficult to determine SoC from voltage at the best of times. Stick a decent DVM across each cell and see if it agrees. If the voltage is correct, there does seem to be a huge imbalance.

[MtB]

9 minutes ago, nicknorman said:

Hmmmm slightly worrying, but how accurate is the voltage measurement? 

 

I was staggered to find a device costing £25 that reports to three decimal places too. It seems consistent though so I'm inclined to believe it. the third digit flutters around a bit by 1 unit but moves up or down consistently with what is going on.

 

I have the preceding model the BC-8S too, which reports to two decimal places. I'll plug it in and post a photo for comparison....

 

 

[MtB]

 
Sooo... not very accurate at all!!! One or both...

[Dr Bob]

RULE 1 - IGNORE THE SoC

RULE 2 - IGNORE THE SoC

RULE 3 - RULE 1 APPLIES

 

the SoC on these gauges is absolutely useless. Mine always reads <30% even when fully charged. IGNORE IT.

 

Ok, on the first screen display, you are well out of balance on charging . You have a 250mV delta between top and bottom. I am working on having less than 100mV delta when the top one gets to 3.6V.

First. Get yourself some resistors as above and drain cells 7 and 8 (8 will require more). I can give you an estimate of how many hours it will need to drain if you can give me some voltage info. When you charged it up, Cell 8 was increasing in voltage. How many minutes between cell 8 on 3.38V and 3.6V and what was the charging rate? That gives you the Ahrs put into the bank between those voltages. Divide that number by 4 and that is the Ahrs into cell 8...which is what you want to take out.  Connect up the resistor circuit. Check there is 8-10A going out and leave it on for X hours. I put a fan over my resistors to keep the temp down but I think they are good up to 100°C or more. Repeat for Cell 7 but likely only need to do it for half the time.

That should bring all the cells to within 100mV delta and you can then start to fine tune them.  Looks like 2-6 are reasonably well balanced but as 1-6 are nowhere near the knee, you wont know if any of 1-6 are laggards until you get them out of the 20-90% SoC range. You may find Cell 1 which looks low now, is actually not far off the knee. You cannot tell until voltage per cell gets over 3.4V or even 3.45V.

I would totally ignore the second screen, Cell 3 is almost the same as Cell 8 so look in balance. It is totally incorrect to assume this as you are on the flat part of the discharge curve and the voltages tell porkies. As you charge, cell 8 will zoom away as it is more charged ....and that is what you have to balance. Only use the cell monitor to do balancing when getting to the top knee or knowing when you are right down the bottom ie you will see a big delta occuring as you get down to the bottom 10% SoC.

 

AS BEFORE...IGNORE THE SOC ON THIS GAUGE.

 

I dont think there is an issue on voltage measurement accuracy. Just check it with a multimeter. Your readings do not surprise me at all. This isnt down to accuracy.

Edited November 11, 2019 by Dr Bob

[MtB]

And the other volt meters for comparison

 

[nicknorman]

Erm why divide the number by 4? Is it to do with the number of cells (Mike has 8 ). But surely not, cells in series each receive the total number of AH, not 1/4 or 1/8th the AH each.

 

“AS BEFORE...IGNORE THE SOC ON THIS GAUGE.” ... it’s Smartgauge deja vue all over again for Mike!

Edited November 11, 2019 by nicknorman

[rusty69]

1 minute ago, Mike the Boilerman said:

And the other volt meters for comparison

 

I guess the gap in the middle is where the smartgauge use to live

[MtB]

1 minute ago, rusty69 said:

I guess the gap in the middle is where the smartgauge use to live

 

Correct!!! ?

[Dr Bob]

1 minute ago, nicknorman said:

Erm why divide the number by 4? Is it to do with the number of cells (Mike has 8 ). But surely not, cells in series each receive the total number of AH, not 1/4 or 1/8th the AH each.

Yes sorry, I was thinking a 4s bank. If you have 8 in series divide by  8. It worked on mine. I had to take 50-80Ahrs out of my high cell (480Ahr nominal- but likely 20% less) to get the delta of 50mV which is my target delta.

[MtB]

6 minutes ago, Dr Bob said:

RULE 1 - IGNORE THE SoC

RULE 2 - IGNORE THE SoC

RULE 3 - RULE 1 APPLIES

 

the SoC on these gauges is absolutely useless. Mine always reads <30% even when fully charged. IGNORE IT.

 

Ok, on the first screen display, you are well out of balance on charging . You have a 250mV delta between top and bottom. I am working on having less than 100mV delta when the top one gets to 3.6V.

First. Get yourself some resistors as above and drain cells 7 and 8 (8 will require more). I can give you an estimate of how many hours it will need to drain if you can give me some voltage info. When you charged it up, Cell 8 was increasing in voltage. How many minutes between cell 8 on 3.38V and 3.6V and what was the charging rate? That gives you the Ahrs put into the bank between those voltages. Divide that number by 4 and that is the Ahrs into cell 8...which is what you want to take out.  Connect up the resistor circuit. Check there is 8-10A going out and leave it on for X hours. I put a fan over my resistors to keep the temp down but I think they are good up to 100°C or more. Repeat for Cell 7 but likely only need to do it for half the time.

That should bring all the cells to within 100mV delta and you can then start to fine tune them.  Looks like 2-6 are reasonably well balanced but as 1-6 are nowhere near the knee, you wont know if any of 1-6 are laggards until you get them out of the 20-90% SoC range. You may find Cell 1 which looks low now, is actually not far off the knee. You cannot tell until voltage per cell gets over 3.4V or even 3.45V.

I would totally ignore the second screen, Cell 3 is almost the same as Cell 8 so look in balance. It is totally incorrect to assume this as you are on the flat part of the discharge curve and the voltages tell porkies. As you charge, cell 8 will zoom away as it is more charged ....and that is what you have to balance. Only use the cell monitor to do balancing when getting to the top knee or knowing when you are right down the bottom ie you will see a big delta occuring as you get down to the bottom 10% SoC.

 

AS BEFORE...IGNORE THE SOC ON THIS GAUGE.

 

I dont think there is an issue on voltage measurement accuracy. Just check it with a multimeter. Your readings do not surprise me at all. This isnt down to accuracy.

 

Fanx Dr Bob.  Will have to look properly later as gottagotoworknow. 

 

 

[nicknorman]

9 minutes ago, Dr Bob said:

Yes sorry, I was thinking a 4s bank. If you have 8 in series divide by  8. It worked on mine. I had to take 50-80Ahrs out of my high cell (480Ahr nominal- but likely 20% less) to get the delta of 50mV which is my target delta.

No, this isn’t right. Cells in series all individually receive the total AH a put into the bank. So if you want to balance by discharging a cell to the point where it had been before say 50AH had been added to the battery (say 50A for 1 hour from the charger) then you need to take 50AH out of that cell.

 

Kirchoff’s first law (aka current law). You cannae change the laws of physics, as someone will say in a few hundred years.

 

remember AH is only a current integral, it has no dimensions of voltage and hence no concept of energy/power. If you had calculated the Watt-hours added, it would have been correct to divide by 4 to work out the WH to be removed. But not so with AH.

Edited November 11, 2019 by nicknorman

[Richard10002]

18 hours ago, Mike the Boilerman said:

 

You are a JEANIUS!!

 

Thank you

Here’s my mate making and balancing a 4 cell battery with the gizmo you have, and some resistors to discharge individual cells.

 

 

[peterboat]

Valence has the answer built in BMS

[Richard10002]

32 minutes ago, peterboat said:

Valence has the answer built in BMS

You mentioned one of your 36v Valences that has failed. What are the symptoms, and have you any idea why?

 

Ive never had a Lead Acid Battery “fail”, as such.

 

Interested as I am close to buying the second hand ones, but keep dithering, and now hear that they might fail.

 

(I accept that anything can fail, but second hand has no come back at all).

[peterboat]

2 minutes ago, Richard10002 said:

You mentioned one of your 36v Valences that has failed. What are the symptoms, and have you any idea why?

 

Ive never had a Lead Acid Battery “fail”, as such.

 

Interested as I am close to buying the second hand ones, but keep dithering, and now hear that they might fail.

 

(I accept that anything can fail, but second hand has no come back at all).

Dont worry Richard it didnt fail on me, I bought unseen 30 batteries some had some charge others only 2 volts they were very very cheap so I took a gamble, a couple had flash green lights some flashing red lights and some no lights!

So I charged them up and 27 were ok 3 wernt, they hold some charge but not as good as the rest and they have flashing red lights which means something is wrong.

I will strip them to see what I can find, no doubt I could make my money back selling the cells on ebay so I am not concerned in the slightest, but with luck I will be able to fix two batteries,

Remember these were hard worked batteries from a bus changed on time rather than anything else, some of the batteries were brand new, they were replacements for ones that had failed in service. I have replaced to 3 dodgy ones and I am more than happy with my bulk purchase.

You will be buying flashing green batteries with a known number of cycles, which will not have been worked hard, the Smiths electric vans had a slow charge system, so they will not have been abused

[rusty69]

5 hours ago, Richard10002 said:

Here’s my mate making and balancing a 4 cell battery with the gizmo you have, and some resistors to discharge individual cells.

 

I found him difficult to watch to start with, but after a while, think he is pretty good to watch. Very enthusiastic. Love the Tesla S battery he is using, 24V too, so may suit MtB, though probably not,as the charging will be different.Inbuilt heating/cooling tubes too.

Edited November 11, 2019 by rusty69

[Dr Bob]

4 hours ago, nicknorman said:

No, this isn’t right. Cells in series all individually receive the total AH a put into the bank. So if you want to balance by discharging a cell to the point where it had been before say 50AH had been added to the battery (say 50A for 1 hour from the charger) then you need to take 50AH out of that cell.

 

Kirchoff’s first law (aka current law). You cannae change the laws of physics, as someone will say in a few hundred years.

 

remember AH is only a current integral, it has no dimensions of voltage and hence no concept of energy/power. If you had calculated the Watt-hours added, it would have been correct to divide by 4 to work out the WH to be removed. But not so with AH.

This is starting to get out of my pay grade! It is interesting though. I calculated the Ahrs needed to take out of my high cell using the 'divide by 4' as 40Ahrs. In practice to balance it took 80Ahrs so half way to what you were saying Nick. The most important thing is to get an approximate idea of how much power to take out. The problem when you start to do it is that you are a little constrained when top balancing. I will comment on the video Richard posted about bottom balancing next but if you are charging with an alternator you need to top balance.

If you only have a 30-40W load circuit ie 3.3V 10A then you are only reducing charge in a cell by 10Ahrs in an hour. If you have to drain the cell by 50Ahrs then that is going to take 5 hours. You can only check the voltages when under charge ie you are looking at the delta but it is only valid when there is amps going in and you are in the voltage knee. Therefore it is a bit trial and error to discharge one cell then put the bank on charge, go to nearly 100% and check the delta and then discharge the bank to put the resistors on the high cell. You cant really keep charging for the 5 hours you are bringing your target cell down. The answer then is to have a bigger drain but I couldnt find resistors that would take 200W and 50A. Nick, can you think of something that would take this sort of power? On my system I now know how long I need to drain one cell from all the voltage data but it will be different on Mike's 24V system. Something that drains at more than 10A would be so useful.

 

Just a comment for Mike. Your bank at the moment is out of balance but it is not a drama. It will only be an issue when charging to over 90% so as long as you discharge the bank to 80% it will be fine for weeks. Get the bits together you need and spend a few days doing the balancing. You said you only took out 10Ahrs. Make sure you discharge to 80% soon. Not a good idea to leave them full.

 

Now, Richard's video on bottom balancing. The guy is using them on a solar only system.

This is interesting but not that applicable if you are using an alternator and have an LA in parallel. Once you are down to 3.0V per cell, you are down to 12.0V total so your LA will be severly discharged. Not a good idea if it is your starter battery and it will sulphate if left down there and you dont use 14.4V+ to get it back to 100%...which is then bad for the Li's. Also if the bank is bottom balanced then when you get it near full with the alternator it will be significantly out of balance ...far worse than Mike's case. So my advice is top balance, not bottom balance. I never let my bank go below 12.7V to protect the LA's so that is circa 3.2V minimum.

Now bottom balancing is quite easy as you have plenty of  time to do it....ie discharge to 3.0V each cell and you can easily see which is high or low and the bank can remain there and you can modify the voltage via the resistor pack with the voltages meaningful. You cant do that at the top end as you dont want to hold it with a charge on when the top cell is at 3.6V. In the video, the guy did the coarse adjustment with the resistor pack and then did the fine tune with the cell monitor. You cant do this at the top end as you wouldnt leave it on charge for 6-8hrs to do the final balance on the cell monitor with cells up at 3.6V. Certainly not on a 30A charger. If you had say a 5 or 10A charger it might work.

 

Those cells in the video look to be smallish...ie <100Ahrs and maybe the BG8S will work at the low end...but I doubt if it transfers anywhere near 1A so would take a long time to do 50Ahrs. I must admit I have not tried to balance on my BG8S as I can only balance at the top end under charge.

I did like the bit about internal resistance to point out which cells could be susceptible to inbalances.

 

The important thing to mention is you cannot use cell voltages when between 20-80% charged to indicate how to balance.

 

Edited November 11, 2019 by Dr Bob

[rusty69]

8 minutes ago, Dr Bob said:

Nick, can you think of something that would take this sort of power?

An electric heater.

 

My old dad has Lithiums in his glider, apparently he has a special plug in gadget he bought to get the voltage down when storing them.It may have been a light bulb.

 

Carry on everyone.............

[Dr Bob]

2 minutes ago, rusty69 said:

An electric heater.

 

Well yes. That's what I was thinking. Does a heating element exist that would take 30-100A at 3.3-3.6V? My muppet brain is not configured well enough to work that one out. Or  would anything else that might  be lying around the boat do.. ie Mrs Bobs hairdryer etc etc.

[rusty69]

1 minute ago, Dr Bob said:

Well yes. That's what I was thinking. Does a heating element exist that would take 30-100A at 3.3-3.6V? My muppet brain is not configured well enough to work that one out. Or  would anything else that might  be lying around the boat do.. ie Mrs Bobs hairdryer etc etc.

Don't ask me, I am still trying to recall Kirchoff's laws.It is all long forgotten like laplace transforms and Fleming's left hand rule.

 

Sensible people will be along shortly.......

[Dr Bob]

Just now, rusty69 said:

........ like laplace transforms and Fleming's left hand rule.

 

 

I had a slide rule once!

1 minute ago, rusty69 said:

Sensible people will be along shortly.......

Nah, Mike's gone to work.

[rusty69]

5 minutes ago, Dr Bob said:

I had a slide rule once!

 

You could hook it up to your shunt.You will then have a variable resistor.

[nicknorman]

On the point about the high current shunts, well it’s just a matter if of dissipating the power. 3.3v at 50A is 165 watts. I can’t think where you would find a heating element, powerful light bulb etc that runs on 3.3v.

 

Why not just do as you have done but a bit more - a metal plate or heat sink onto which you bolt some chunky resistors. You need 0.066 ohm to get 50A from 3.3v. So how about two 0.1 ohm in parallel, that gives 0.05 ohm but with a bit of loss in wires etc it will be close. These ones are rated at 100W each with suitable heat sinking. If there was no loss in wires you would be over 200W but here’s a thought, you are sitting on one massive heat sink aka the canal. Make the wires long enough so that you can dangle the contraption in the canal. Don’t burn any fish though!

 

https://uk.rs-online.com/web/p/panel-mount-fixed-resistors/1752437/

 

[peterboat]

1 hour ago, rusty69 said:

An electric heater.

 

My old dad has Lithiums in his glider, apparently he has a special plug in gadget he bought to get the voltage down when storing them.It may have been a light bulb.

 

Carry on everyone.............

I use a light bulb on my Honda Insight batteries to discharge them to 75 volts