Cheap LiFePO4 BMS?

[MtB]

Some data on my own lithium battery now if Dr Bob is interested in analysing it...

 

On 16th Nov after a lot of monitored charging I finally found the upper knee and accidentally allowed the voltage on cell 8 to rise to about 4.3v. This happened in just minutes after it got to 3.6v! I stopped charging and turned on every load in the boat I could find to drag it down to a sensible level, at which point I took this photo:

 

 

 

 

 

 

So here I have the bank about as fully charged as its gonna get until I do some top balancing.

 

So I checked the BMV-702 intending to set the AH reading to zero but it had already done this for itself. Now I can start AH counting until the bank is fully discharged to establish the real capacity of my supposed 200ah 24v lithium bank from a vandalised car. 

 

After a few days of the AH reading see-sawing up and down as a bit of solar charging happened (to my surprise), last night the bank finally got low enough for me to know I'd found the lower knee. One cell dropped below 3.0v as in the following photo:

 

 

 

 

The AH counter on the BMV-702 shows -67AH as below, so it appears I was probably conned when I bought the batteries! Beware the bloke in Cobham who converts ICE cars to electric and sells lithiums as a sideline.

 

 

 

 

And here is the bank after ten minutes of charging at 22A with the Whispergen, showing the voltage of the lowest cell 2 beginning to creep up:

 

 

 

And here is the bank after a couple of hours of charging at about 25A

 

 

 

[WotEver]

My first thought is that with a cell imbalance of 669mV you were probably a fair way off actually being full.  

[MtB]

Just now, WotEver said:

My first thought is that with a cell imbalance of 669mV you were probably a fair way off actually being full.  

 

I was thinking that one cell is bound to get over the knee first, unless you really do have them all accurately top balanced. Cell imbalance on the plateau is closer to 30mv. It varies widely but only ever got that large as cell 8 ran away with overcharge first.

 

Mine cells do not appear well matched in capacity and at the lower end, a different cell gets empty first.

 

The interesting thing is that in the plateau, the relative voltages of all the cells are constantly swapping around. The 'shape' of the histogram is constantly changing. 

 

 

[WotEver]

2 minutes ago, Mike the Boilerman said:

Cell imbalance on the plateau is closer to 30mv

Ahh ok, fair enough. So it does appear that your battery is down to not much more than a third of its original capacity. 

[MtB]

6 minutes ago, WotEver said:

Ahh ok, fair enough. So it does appear that your battery is down to not much more than a third of its original capacity. 

 

I don't know enough about lithiums to guess at badge capacity just by looking at them. The cells bear no data or labels, names, badges or anything. The 200AH was what the seller said they are, and I stoopidly believed him. He may well have been guessing himself and/or relying on buyer ignorance too.  

 

The whole bank of 8 cells is about the size and shape of a 15 year old desktop computer, and is astonishingly heavy. I'd guess it is 40kG.

 

 

 

Edited November 22, 2019 by Mike the Boilerman

[Dr Bob]

Dont worry about capacity at this stage. Your cell 8 (and to extent cell 7) are way out. When cell 8 was full, maybe the rest were only half full!!!! Sort out the cell balance first. Cells 1 and 2 were also very low so this compounds the problem. Once balanced you can then redo the capacity check and I wouldnt be surprised if it is doubled. Also, what are the temperatures? You may be loosing 10% of capacity if say 10°C rather than at 20°C which is what the capacity will be rated at.

First thing to do is get a resistor circuit onto cell 8 and then cell 7 and I would take say 10Ahr out of 8 and 5 Ahr out of 7 and recharge until the top cell gets to 3.5V and see how the balance is. I think you may need to repeat this 2 or 3 times but do it carefully.

You then need to repeat but charging cells 1 and 2 by say 10A each (using a 12V charger but with some thin wire - lets cover that once you have sorted cells 8 and 7).

You need to try and get the so that the delta is <100mV so low cell is 3.5V and high cell is 3.6V ....but when charging watch the tail current.

If you charged cell 8 to 4.2V you may have damaged the cell. You must be watching full time once cells get to 3.5V as they shoot up very fast. The voltages shown after the 4.2V look fairly good so fingers crossed you havent damaged anything.

......so ignore capacity for now. Sort out the balancing.

The variation of voltage in the plateau is very typical - you must ignore these differences - just look at the deltas when you get near the top and bottom.

 

[cuthound]

11 minutes ago, Dr Bob said:

Dont worry about capacity at this stage. Your cell 8 (and to extent cell 7) are way out. When cell 8 was full, maybe the rest were only half full!!!! Sort out the cell balance first. Cells 1 and 2 were also very low so this compounds the problem. Once balanced you can then redo the capacity check and I wouldnt be surprised if it is doubled. Also, what are the temperatures? You may be loosing 10% of capacity if say 10°C rather than at 20°C which is what the capacity will be rated at.

First thing to do is get a resistor circuit onto cell 8 and then cell 7 and I would take say 10Ahr out of 8 and 5 Ahr out of 7 and recharge until the top cell gets to 3.5V and see how the balance is. I think you may need to repeat this 2 or 3 times but do it carefully.

You then need to repeat but charging cells 1 and 2 by say 10A each (using a 12V charger but with some thin wire - lets cover that once you have sorted cells 8 and 7).

You need to try and get the so that the delta is <100mV so low cell is 3.5V and high cell is 3.6V ....but when charging watch the tail current.

If you charged cell 8 to 4.2V you may have damaged the cell. You must be watching full time once cells get to 3.5V as they shoot up very fast. The voltages shown after the 4.2V look fairly good so fingers crossed you havent damaged anything.

......so ignore capacity for now. Sort out the balancing.

The variation of voltage in the plateau is very typical - you must ignore these differences - just look at the deltas when you get near the top and bottom.

 

 

Doesn't quite seem like the "fit and forget" battery that some people claim Li batteries are!

 

I think I will wait until a drop in system is available at a reasonable price before adopting this new technology.

[MtB]

18 minutes ago, cuthound said:

 

Doesn't quite seem like the "fit and forget" battery that some people claim Li batteries are!

 

I think the 'drop in' type Peter and Richard have cell balancing hardware built in to do it automatically. All the discussion you are commenting on is because I'm doing it the hard way, manually using bare cells.

 

A world of difference. Very much NOT fit and forget until you understand in fine detail how to design and manage your system. But when you do, and you get the design and configuration right, they seem to run with less attention than LAs.

 

With the 'drop in' type (a misnomer in my opinion) the cells look after themselves and *all* you have to do is arrange high and low voltage cut-off, appropriate charge voltages, and high and low temperature protection. Some I suspect, have the high and low temperature protection internally too but I'm not sure.

 

 

 

 

 

 

Edited November 22, 2019 by Mike the Boilerman

[nicknorman]

Talking of “drop in system”, has this article been linked to before?

 

https://pbase.com/mainecruising/mastervolt_lifepo4_autopsy

 

[Dr Bob]

39 minutes ago, cuthound said:

 

Doesn't quite seem like the "fit and forget" battery that some people claim Li batteries are!

 

I think I will wait until a drop in system is available at a reasonable price before adopting this new technology.

If you are installing a DIY Li setup, then it is fit, configure and forget. The configure is the DIY bit and there is some faffing required to sort it out - to balance and to work out your charging voltages per different charging source. The configuring is not difficult but the user needs to learn about the Li's which is different to LAs. Mike is in the configuring phase and has found his cells way out of balance - its not difficult to sort but will take  a week of faffing to get it right. Once done though you can forget about them for 3 months. Worth re-checking after this time to re-sync the amp hour counter and take them to full again.

 

I dont think a full commercial system will ever come down to a low enough price for most boaters. Go to victron and you will pay £5K for a plug and play system which will be fit and forget. However, the use of 2nd hand batteries ex electric vehicles means that you can do it for £1000 which is then viable....but you need to DIY the management system ......hence it will never be fit and forget. Each to his own! The first month of ownership was a steep learning curve for me but with the help of Tom, peter and MP, it was very painless (just a bit time consuming) and now it is really really easy. As more peeps put them in, the experience base will grow and it will be easier for others to understand. Mike is just suffering because he got a set of cells that were not balanced. Had this been spotted straight away, the supplier would likely have taken them back to balance them.....but it is not difficult to sort.

[nicknorman]

Have people done the physically constraining the cells from swelling thing?

[Dr Bob]

49 minutes ago, nicknorman said:

Have people done the physically constraining the cells from swelling thing?

YES!

The batteries that Tom, MP and myself bought from Jeremy have aluminium holding straps. Very important to get them assembled this way.

[peterboat]

1 hour ago, nicknorman said:

Talking of “drop in system”, has this article been linked to before?

 

https://pbase.com/mainecruising/mastervolt_lifepo4_autopsy

 

I think someone posted it when I bought my batteries, I think my solar charging is much more gentle than a charger or alternator, and the whispergen seems to do a good job as well? If I am using the whispergen generally their has been little sun and the batteries are down to 13.1 volts, my settings are for 1.5 hours running so their no chance of a full charge, so again they are protected by design. I haven't taken a 12 volt valence battery apart but they look well made on tinernet strip downs  time will tell if that's true 

[Dr Bob]

1 hour ago, nicknorman said:

Talking of “drop in system”, has this article been linked to before?

 

https://pbase.com/mainecruising/mastervolt_lifepo4_autopsy

 

Not seen that one before. Thanks for posting.

So much for plug and play!

If you build your own system based on bare cells then this should not happen.....but it does. Just look at what Mike has done. If you want to protect your batteries you must not overcharge. That is exactly what Mike did by taking one cell up to 4.3V. You cannot go off and make a pot of tea when you cells get to 3.5V!!!! You really need audible alarms and auto disconnects on total bank voltage and on cell voltages for basic protection.....but when configuring for the first time and taking them up to full, you cannot set total bank voltage as you dont know where that is....but you MUST look at cell voltages. Mike should have stopped charging imediately his first cell reached 3.5V with the lowest (and most) down at 3.33V and sorted out the balance issue.  If you are 170mV delta with the top one at 3.5V, you can bet your life you will be 300+mV out when the top one gets to 3.6V. The outcome is Mike might have wrecked cell 8. I hope not but in Mike's case it is avoidable:

- build your own bms so you know a cell cant go over 3.8V max

- if you dont, watch the cell voltages every second when you first take them up to full

- scope out your voltages so you know where your 100% termination point is (ie tail current around 15-20A for a 400Ahr bank)

- set your termination voltage lower ie 0.1V or 0.2V lower so that you terminate at 80% ish full - work out what voltage is needed for 80% for a given charge current.

 

The above works if you are charging at a lowish current ie 30-50A. If you are going to charge at 100A then a different termination strategy is needed and more focus on the tail current rather than a termination voltage. I will leave that to MP and Nick to advise.

 

As for the mastervolt link that Nick posted, it just shows that the commercial systems can get it wrong.....but a plug and play system that fits everyone is very very difficult to engineer. An individual doing his own DIY system should be able to do a lot better. It is not rocket science. Dont charge to high or discharge too low. Dont charge/discharge out side of a  temperature range. It is pretty simple. With audible alarms, auto disconnects and charge sources that cut off charging before the alarms sound,  you should be in reasonable shape. In normal operation you are well away from low or high SoC so you never need to think about it......but as soon as an alarm sounds then respond! That mastervolt system was just bad.

 

What is confusing to the new user is the difference in charging with different power levels. I wasnt really aware of what happend when you charged at 100A. All my experience is down at the bottom end 20A-50A from my data and seeing a lot of MPs data. At those charge rates then termination voltages are low ie 13.8-14.1V and little time is spent in the CV phase.

It is easy to terminate at 80% SoC and then to manually go up to 100% if needed. At charge rates over 100A then you are looking at termination voltages up to 14.6V with longer in the CV phase (although I have no experience of this - MP/Nick discussed this earlier in the thread). A commercial BMS would have to cope with both types of termination and I am not sure they are capable of doing that. My victron IP22 has a LiFePO4 setting but at 30A it does not turn off when the tail current shows we are at 100% full.(ETA - If I left my IP22 charging, it is likely it would have wrecked my bank!). It is far easier in DIYing your own system to tune it to the currents YOU are using and avoid all this overcharging.

Edited November 22, 2019 by Dr Bob

[peterboat]

22 minutes ago, Dr Bob said:

Not seen that one before. Thanks for posting.

So much for plug and play!

If you build your own system based on bare cells then this should not happen.....but it does. Just look at what Mike has done. If you want to protect your batteries you must not overcharge. That is exactly what Mike did by taking one cell up to 4.3V. You cannot go off and make a pot of tea when you cells get to 3.5V!!!! You really need audible alarms and auto disconnects on total bank voltage and on cell voltages for basic protection.....but when configuring for the first time and taking them up to full, you cannot set total bank voltage as you dont know where that is....but you MUST look at cell voltages. Mike should have stopped charging imediately his first cell reached 3.5V with the lowest (and most) down at 3.33V and sorted out the balance issue.  If you are 170mV delta with the top one at 3.5V, you can bet your life you will be 300+mV out when the top one gets to 3.6V. The outcome is Mike might have wrecked cell 8. I hope not but in Mike's case it is avoidable:

- build your own bms so you know a cell cant go over 3.8V max

- if you dont, watch the cell voltages every second when you first take them up to full

- scope out your voltages so you know where your 100% termination point is (ie tail current around 15-20A for a 400Ahr bank)

- set your termination voltage lower ie 0.1V or 0.2V lower so that you terminate at 80% ish full - work out what voltage is needed for 80% for a given charge current.

 

The above works if you are charging at a lowish current ie 30-50A. If you are going to charge at 100A then a different termination strategy is needed and more focus on the tail current rather than a termination voltage. I will leave that to MP and Nick to advise.

 

As for the mastervolt link that Nick posted, it just shows that the commercial systems can get it wrong.....but a plug and play system that fits everyone is very very difficult to engineer. An individual doing his own DIY system should be able to do a lot better. It is not rocket science. Dont charge to high or discharge too low. Dont charge/discharge out side of a  temperature range. It is pretty simple. With audible alarms, auto disconnects and charge sources that cut off charging before the alarms sound,  you should be in reasonable shape. In normal operation you are well away from low or high SoC so you never need to think about it......but as soon as an alarm sounds then respond! That mastervolt system was just bad.

 

What is confusing to the new user is the difference in charging with different power levels. I wasnt really aware of what happend when you charged at 100A. All my experience is down at the bottom end 20A-50A from my data and seeing a lot of MPs data. At those charge rates then termination voltages are low ie 13.8-14.1V and little time is spent in the CV phase.

It is easy to terminate at 80% SoC and then to manually go up to 100% if needed. At charge rates over 100A then you are looking at termination voltages up to 14.6V with longer in the CV phase (although I have no experience of this - MP/Nick discussed this earlier in the thread). A commercial BMS would have to cope with both types of termination and I am not sure they are capable of doing that. My victron IP22 has a LiFePO4 setting but at 30A it does not turn off when the tail current shows we are at 100% full.(ETA - If I left my IP22 charging, it is likely it would have wrecked my bank!). It is far easier in DIYing your own system to tune it to the currents YOU are using and avoid all this overcharging.

It does seem that high amps are the enemy for us LI users, not at the beginning but definitely beyond 75% it could easily be a disaster waiting to happen.  I suppose it's why I like solar so much I can see the amps dropping as it approaches nearly 80% its gone into absorb which is just a couple of amps. Turn on the kettle and its 50 amps again......but not for long ?

[cuthound]

2 hours ago, nicknorman said:

Talking of “drop in system”, has this article been linked to before?

 

https://pbase.com/mainecruising/mastervolt_lifepo4_autopsy

 

 

What would worry me when buying second-hand cells is not knowing if they have been over or under discharged in their previous life.

 

Does the internal BMC record highest and lowest voltages for Individual cells?

[cuthound]

32 minutes ago, Dr Bob said:

Not seen that one before. Thanks for posting.

So much for plug and play!

If you build your own system based on bare cells then this should not happen.....but it does. Just look at what Mike has done. If you want to protect your batteries you must not overcharge. That is exactly what Mike did by taking one cell up to 4.3V. You cannot go off and make a pot of tea when you cells get to 3.5V!!!! You really need audible alarms and auto disconnects on total bank voltage and on cell voltages for basic protection.....but when configuring for the first time and taking them up to full, you cannot set total bank voltage as you dont know where that is....but you MUST look at cell voltages. Mike should have stopped charging imediately his first cell reached 3.5V with the lowest (and most) down at 3.33V and sorted out the balance issue.  If you are 170mV delta with the top one at 3.5V, you can bet your life you will be 300+mV out when the top one gets to 3.6V. The outcome is Mike might have wrecked cell 8. I hope not but in Mike's case it is avoidable:

- build your own bms so you know a cell cant go over 3.8V max

- if you dont, watch the cell voltages every second when you first take them up to full

- scope out your voltages so you know where your 100% termination point is (ie tail current around 15-20A for a 400Ahr bank)

- set your termination voltage lower ie 0.1V or 0.2V lower so that you terminate at 80% ish full - work out what voltage is needed for 80% for a given charge current.

 

The above works if you are charging at a lowish current ie 30-50A. If you are going to charge at 100A then a different termination strategy is needed and more focus on the tail current rather than a termination voltage. I will leave that to MP and Nick to advise.

 

As for the mastervolt link that Nick posted, it just shows that the commercial systems can get it wrong.....but a plug and play system that fits everyone is very very difficult to engineer. An individual doing his own DIY system should be able to do a lot better. It is not rocket science. Dont charge to high or discharge too low. Dont charge/discharge out side of a  temperature range. It is pretty simple. With audible alarms, auto disconnects and charge sources that cut off charging before the alarms sound,  you should be in reasonable shape. In normal operation you are well away from low or high SoC so you never need to think about it......but as soon as an alarm sounds then respond! That mastervolt system was just bad.

 

What is confusing to the new user is the difference in charging with different power levels. I wasnt really aware of what happend when you charged at 100A. All my experience is down at the bottom end 20A-50A from my data and seeing a lot of MPs data. At those charge rates then termination voltages are low ie 13.8-14.1V and little time is spent in the CV phase.

It is easy to terminate at 80% SoC and then to manually go up to 100% if needed. At charge rates over 100A then you are looking at termination voltages up to 14.6V with longer in the CV phase (although I have no experience of this - MP/Nick discussed this earlier in the thread). A commercial BMS would have to cope with both types of termination and I am not sure they are capable of doing that. My victron IP22 has a LiFePO4 setting but at 30A it does not turn off when the tail current shows we are at 100% full. It is far easier in DIYing your own system to tune it to the currents YOU are using and avoid all this overcharging.

 

The batteries in Nick's link were over discharged by the internal BMS's parasitic load, despite an undervoltage disconnect contactor being fitted to disconnect the actual load.

 

Hopefully this has been addressed on newer Li battery BMS designs, but how would a novice buyer know?

[peterboat]

7 minutes ago, cuthound said:

 

What would worry me when buying second-hand cells is not knowing if they have been over or under discharged in their previous life.

 

Does the internal BMC record highest and lowest voltages for Individual cells?

On Richard's and my batteries yes is the answer, we can check both with the puter and by the flashing LED on the top of the battery. Valence batteries are 2.5k new buy one with 100 cycles on it for 300 squids makes it a bargain especially as you can check it and allegedly it memory can't be wiped 

[cuthound]

7 minutes ago, peterboat said:

On Richard's and my batteries yes is the answer, we can check both with the puter and by the flashing LED on the top of the battery. Valence batteries are 2.5k new buy one with 100 cycles on it for 300 squids makes it a bargain especially as you can check it and allegedly it memory can't be wiped 

 

Thanks Peter, if I take the plunge I will look for secondhand Valence batteries.

[MtB]

1 hour ago, Dr Bob said:

You cannot go off and make a pot of tea when you cells get to 3.5V!!!!

 

Oh yes you can, and look what happened! DAMHIK. 

 

I suspect I've got away with it though, as following the overcharge I took the bank down until the lowest cell was fully discharged, and that cell was NOT cell 8 that got to the bottom first. And on re-charging, cell 8 seems to be accepting a charge just the same as before. I'm not sure however, exactly how a damaged cell is gonna behave. 

 

Shifting back on topic, I've measured the resistances on this big beautiful ceramic core wire wound ballast resistor. The sectors left to right are 128.5, 31.6, 16.0 and 62.6 Ohms.  Must be some useful resistances in amongst that lot!

 

 

 

 

 

 

[MoominPapa]

8 minutes ago, Mike the Boilerman said:

 

Oh yes you can, and look what happened! DAMHIK. 

 

I suspect I've got away with it though, as following the overcharge I took the bank down until the lowest cell was fully discharged, and that cell was NOT cell 8 that got to the bottom first. And on re-charging, cell 8 seems to be accepting a charge just the same as before. I'm not sure however, exactly how a damaged cell is gonna behave. 

 

Shifting back on topic, I've measured the resistances on this big beautiful ceramic core wire wound ballast resistor. The sectors left to right are 128.5, 31.6, 16.0 and 62.6 Ohms.  Must be some useful resistances in amongst that lot!

 

 

 

 

 

 

Re your tested capacity, the value you got ain't really valid as the cells are so unbalanced. Balance then capacity test, the capacity number can only go up.

 

Your resistor values are all nice powers of two, but really rather large. If you put them all in parallel you'll get an 8.5 ohm resistor which will pull about 400mA from a 3.3v cell, and dissipate just over 1W. You'll be there for weeks trying to balance that pack with that.

 

MP.

 

[WotEver]

1 hour ago, Mike the Boilerman said:

Shifting back on topic, I've measured the resistances on this big beautiful ceramic core wire wound ballast resistor...

There are some 1 ohm 100W resistors here pretty cheap. They’d need a decent heat sink:

https://rover.ebay.co.uk/rover/0/0/99?loc=https%3A%2F%2Fwww.ebay.co.uk%2Fp%2F1443198623%3Fiid%3D223728969512

cheap heat sink:

https://rover.ebay.com/rover/0/0/0?mpre=https%3A%2F%2Fwww.ebay.co.uk%2Fulk%2Fitm%2F183509032586

[MtB]

10 minutes ago, WotEver said:

There are some 1 ohm 100W resistors here pretty cheap. They’d need a decent heat sink:

https://rover.ebay.co.uk/rover/0/0/99?loc=https%3A%2F%2Fwww.ebay.co.uk%2Fp%2F1443198623%3Fiid%3D223728969512

cheap heat sink:

https://rover.ebay.com/rover/0/0/0?mpre=https%3A%2F%2Fwww.ebay.co.uk%2Fulk%2Fitm%2F183509032586

 

Thanks, but coincidentally I found and bought some of those half an hour ago!

 

 

[TheBiscuits]

1 hour ago, Mike the Boilerman said:

Must be some useful resistances in amongst that lot!

Nah.  Resistance is futile as any fule noes.

[nicknorman]

Well anyway I have now ordered the AD7280 battery monitoring chip, and the AR6000 alternator regulator chip. I’ll design the PCBs to suit shortly. Each chip / PCB requires a microcontroller to talk to /control it, and the AD7280 talks to its microcontroller over SPI (never used it!), whilst the AR6000 talks over LIN (never used it). So a fair bit of development work to get the communications sorted out, before actual application development and  trials. But the comms can be sorted out at home over winter.

Edited November 22, 2019 by nicknorman