Balancing LiFePO4s

[Dr Bob]

10 hours ago, nicknorman said:

But projects that try to do all things at the outset, are doomed to never get finished!

 

What is certainly needed at the moment is an alternator controller that can manage current out and termination.....and if it is to be available to others....be possible to attach it to a range of common alternators in use on boats. The simplest controller could just be worked on total bank voltage to an 80-90% charge but with inputs to terminate if say a relay closed (or opened) on  the BMV or the cell monitoring module. Maybe then there is a switch to override the voltage for the user to manually take it up to 100% but still managing the current output so you dont overstress the alternator. If you have a 100% setting built in then you would need to get into how to input voltage and tail current which then increases complexity, development time and maybe other modules to input this data.

Yes it would be wonderful to come up with a full BMS but that then would take a lot of effort to make sure it was plug and play and idiot proof and price is then an issue. The controller T&B have was $100 but now it is commercial it is $500...I think you were in that discussion earlier in the year.

One other thing that is missing for 'off the shelf' solutions to a DIY setup is the circuit to activate the Tyco type relays. Yes there are solutions but they add £100 onto the cost of system. Your project is to control the alternator but there will be other charge sources and the need to disconnect on deep discharge or high/low temp .....so perhaps your box could also have a circuit in it to take the signal from the BMV relay or cell monitoring relay and convert it to a pulse that the latching rely will undestand. Kill two birds with one stone. MP was threatening to build one. Perhaps you could join forces?

A few other inputs - if you will pardon the intrusion:

It would be very useful to have easy control over how much current the alterntor would give. Obviously this is tied to alternator temperature and it should be easy to do  that control automatically but does that actually work? You will have a good idea how you are going to handle this but I would really feel comfortable having a control to switch into 'winter charging mode' to pull the power up ie in summer you are cuising for hours and the alternator gets hot (high ambient temps). In the winter I run the engine for 1 hour to warm the water and with low ambients, the current can be set much higher. Not much chance of damaging the alternator in an hour. Maybe it is all done automatically but my Sterling AtoB's temperature control doesnt work in the temperature range I think is bad.

That raises another issue though and that is how you set the temperature limits on the alternator. MP and Tom have both burnt out their domestic alternators  this year by allowing them to work at full output. They have alternators that are cheap to replace. Mine isnt. I have therefore been operating at a conservatively safe level.....but not a clue what is safe or not. It would be great if you can keep us posted on how you will handle 'what is a suitable output'.

[nicknorman]

1 hour ago, Dr Bob said:

What is certainly needed at the moment is an alternator controller that can manage current out and termination.....and if it is to be available to others....be possible to attach it to a range of common alternators in use on boats.

I don't think this is ever going to be a retail project. It will always be necessary to do some surgery on the alternator to disconnect the regulator and replace it with the new regulator. Some retail devices seem to leave the original regulator in as a backup, because the new regulator's aim is solely to increase output voltage. But of course the aim of the new regulator would be to reduce the output voltage, so the old regulator has to be disconnected. It is necessary to connect to D+, B+, a phase (ie the W terminal) and ground,  all of which are available on the back of most alternators. However it is also necessary to connect directly to the slip ring brushes, so that is where surgery is required. Not a big deal for someone able to wield a soldering iron, but it's never going to be plug and play!

 

The simplest controller could just be worked on total bank voltage to an 80-90% charge but with inputs to terminate if say a relay closed (or opened) on  the BMV or the cell monitoring module. Maybe then there is a switch to override the voltage for the user to manually take it up to 100% but still managing the current output so you dont overstress the alternator. If you have a 100% setting built in then you would need to get into how to input voltage and tail current which then increases complexity, development time and maybe other modules to input this data.

Yes but not very difficult. I have an idea to use a hall effect sensor heat-shrinked onto the alternator and or battery cable to measure current. These are cheap and cheerful but not particularly accurate - however if you want to limit current to 50A or whatever, it doesn't matter too much if it is actually 45A or 55A. Accurate current measurement is only necessary for eg amp-hour counting.

 

Yes it would be wonderful to come up with a full BMS but that then would take a lot of effort to make sure it was plug and play and idiot proof and price is then an issue. The controller T&B have was $100 but now it is commercial it is $500...I think you were in that discussion earlier in the year.

 

One other thing that is missing for 'off the shelf' solutions to a DIY setup is the circuit to activate the Tyco type relays. Yes there are solutions but they add £100 onto the cost of system. Your project is to control the alternator but there will be other charge sources and the need to disconnect on deep discharge or high/low temp .....so perhaps your box could also have a circuit in it to take the signal from the BMV relay or cell monitoring relay and convert it to a pulse that the latching rely will undestand. Kill two birds with one stone. MP was threatening to build one. Perhaps you could join forces?

 

No my project is for more of an integrated system. Once you have the sensor data and the microprocessor you can do a lot of stuff quite easily. I would prefer not to rely on external equipment such as a BMV. As I mentioned, whilst the system could easily operate isolation relays in the event of over/under voltage etc, I'd be a bit worried about having all eggs in one basket. A protection relay only has to operate on high or low voltage and should ever normally operate - emergency / failure situation only. So the plan would be for a separate device that monitored for such a situation and shut off only in exceptional circumstances. Like all such devices, it would need a test button to make sure it worked! A question though - does such a device need to monitor individual cell voltages and cut off if any one is under/over voltage, or is total battery voltage an adequate measure?

 

A few other inputs - if you will pardon the intrusion:

It would be very useful to have easy control over how much current the alterntor would give. Obviously this is tied to alternator temperature and it should be easy to do  that control automatically but does that actually work? You will have a good idea how you are going to handle this but I would really feel comfortable having a control to switch into 'winter charging mode' to pull the power up ie in summer you are cuising for hours and the alternator gets hot (high ambient temps). In the winter I run the engine for 1 hour to warm the water and with low ambients, the current can be set much higher. Not much chance of damaging the alternator in an hour. Maybe it is all done automatically but my Sterling AtoB's temperature control doesnt work in the temperature range I think is bad.

I think it should be automatic temperature control. Apart from the seasonal variations you allude to, the temperature of an alternator depends a lot not just on output current, but on the combination of output current and rpm. At higher rpm there needs to be less field current so less ohmic heating from that, but more importantly the fan is running faster so the cooling is better. These LIN controller alternator regulators have a facility for setting the maximum field current, that might be quite an easy way to do it because that would mean less output current at low rpm and more at higher rpm when cooling is better. But its something to experiment with. I'd use a serial digital temperature sensor like a DS1822 so that the microprocessor will absolutely know whether the sensor is working / connected. In the event of no communication with the sensor, it could go into a fall back mode with reduced/conservative output.

That raises another issue though and that is how you set the temperature limits on the alternator. MP and Tom have both burnt out their domestic alternators  this year by allowing them to work at full output. They have alternators that are cheap to replace. Mine isnt. I have therefore been operating at a conservatively safe level.....but not a clue what is safe or not. It would be great if you can keep us posted on how you will handle 'what is a suitable output'. We have an Iskra 175A alternator which I'm sure is expensive! A bit more research needed into what is a safe maximum temperature but I'd guess around 90C might be a suitable limit.  They do run pretty hot! Whatever it would be, it would just be one line of code to edit to change it.

And of course it depends on where exactly on the alternator you measure the temperature. Cooling air is pulled in past the diodes/regulator so that is the cooler end - by the time its passed through the windings it will be exiting very hot!

Anyway, best not to get too excited, this project will be a while in the making and I'm not in a position to start it any time soon, because our Trojans are fine at the moment. That said, I think it has been bad for the Trojans to get an initial charge at 175A - even though this tails off fairly quickly. Trojan recommends a charge current of 10-15% IIRC, which would be ~45-70A. It would be better to limit the charging current to a more conservative value whilst retaining the maximum 175A output to be available to feed the inverter etc. Hence the need to monitor  battery current as opposed to just alternator output.

Comments added in red

 

Edit: It typed this, posted it successfully, then a few minutes later CWDF seemed to have a massive crash and on recovery, it had disappeared! Fortunately it was still in the buffer. Phew!

Edited November 14, 2019 by nicknorman

[Dr Bob]

21 minutes ago, nicknorman said:

Comments added in red

 

A question though - does such a device need to monitor individual cell voltages and cut off if any one is under/over voltage, or is total battery voltage an adequate measure?

The simplest way is to just do it on total voltage to 80% and voltage plus tail current to 100%. Tom and I both have the Aussie BMM8 v2 (IIRC) cell monitor which triggers the disconnect on cell voltage too high or too low.....but that cell voltage is not configurable. I have that module and it is separate from my BMV voltage relay so it is an extra layer of protection ie if high voltages, the BMV triggers the disconnect first, but if that were to fail then the BMM8 would do the job when cells get too high. I would prefer to have the separate modules so all my eggs are not in one basket.

Total voltage is certainly fine up to 90% as long as you can set it to 13.7, 13.8 13.9V etc. Everyone will have their own setting dependent on how much current is going in. If you are charging with a 175A alternator then maybe you set it at 14.1V compared to my 90A alternator. If you are aiming at 100% (or lets say near 100%) then you would work on voltage (above a given value) plus a current dropping from the normal charge rate to 3-4% capacity - so mine would be there with a voltage over 13.9V with the current dropping from 30A to say 15A. MP seems to have got his system working well.  If you are aiming for 100% though, it may be easier just to terminate at 3.6V on the highest cell and not bother with total voltage. That would be independent of charge rate which has a big effect on the voltage termination so there are benefits either way. Maybe your system could do total voltage to 80% and cell voltage to 100%.

The nervous among us would still have the aussie BMM8 as a backup to overvoltage.

I am not sure the you could use cell voltage if you are charging up to 80% as you are still on the plateau and so difficult to pick a voltage to control on.

[nicknorman]

13 minutes ago, Dr Bob said:

The simplest way is to just do it on total voltage to 80% and voltage plus tail current to 100%. Tom and I both have the Aussie BMM8 v2 (IIRC) cell monitor which triggers the disconnect on cell voltage too high or too low.....but that cell voltage is not configurable. I have that module and it is separate from my BMV voltage relay so it is an extra layer of protection ie if high voltages, the BMV triggers the disconnect first, but if that were to fail then the BMM8 would do the job when cells get too high. I would prefer to have the separate modules so all my eggs are not in one basket.

Total voltage is certainly fine up to 90% as long as you can set it to 13.7, 13.8 13.9V etc. Everyone will have their own setting dependent on how much current is going in. If you are charging with a 175A alternator then maybe you set it at 14.1V compared to my 90A alternator. If you are aiming at 100% (or lets say near 100%) then you would work on voltage (above a given value) plus a current dropping from the normal charge rate to 3-4% capacity - so mine would be there with a voltage over 13.9V with the current dropping from 30A to say 15A. MP seems to have got his system working well.  If you are aiming for 100% though, it may be easier just to terminate at 3.6V on the highest cell and not bother with total voltage. That would be independent of charge rate which has a big effect on the voltage termination so there are benefits either way. Maybe your system could do total voltage to 80% and cell voltage to 100%.

The nervous among us would still have the aussie BMM8 as a backup to overvoltage.

I am not sure the you could use cell voltage if you are charging up to 80% as you are still on the plateau and so difficult to pick a voltage to control on.

Yes so for an 80% charge it could be done on total voltage and for ~100% it would need to be cell voltage (and current) but I reality it would probably be easier to derive total voltage from the sum of cell voltages, rather than having a separate means of measuring total voltage. Just thinking how I might avoid having to have 2 separate cell voltage measurement chips (one for general control and one for emergency shut off) but I suppose if one wants a reliable protection system, it has to be fully independent from the control system. The chip is about £8 so not too bad, it just means more wires!

 

Oh and I found an article about alternator temperatures. They do seem to be able to run very hot!

https://www.electronics-cooling.com/2002/05/thermal-design-challenges-in-automotive-alternator-power-electronics/

 

Edited November 14, 2019 by nicknorman

[Dr Bob]

18 minutes ago, nicknorman said:

Yes so for an 80% charge it could be done on total voltage and for ~100% it would need to be cell voltage (and current) but I reality it would probably be easier to derive total voltage from the sum of cell voltages, rather than having a separate means of measuring total voltage.

 

Yes, get total voltage from the sum of cell voltage. Nice!

For 100% you shouldnt need current. Just terminate when one cell gets to 3.6V If the bank is not balanced then you may not be at 100% but you shouldnt go over 3.6V anyway. This sounds too easy and I wonder why MP didnt just terminate on this? Am I missing something? @MoominPapa what do you think? Why did you not just have your 100% charge termination based on the first cell hitting 3.6V rather than a total voltage/tail current value?

There is one thing that concerns me and that is the behaviour of these cells as they get to 100%. It is not clear at all from the literature and I have never pushed the cells to this point. Bear with me as I explain. If the 4 cells are perfectly in balance and all cells hit 3.6V at the same time then that is 14.4V. That sounds high. Say then the charge source is at 14.38V. That would not trip the 3.6V limit. However charge would still go into the cells and the current would start to drop as it approached 100%. At this voltage the cells would go to 100% and maybe over without ever hitting the voltage limit. All the literature says you can go to 3.6V per cell without damage but is that right. We have discussed the issue with regard to charging with solar (Peter has mentioned this a lot) but if we charge with 13.8V can you wreck the cells? The percieved wisdom is yes....dont have a 13.8V float. The 3.6V is therefore OK when charging normally but 3.6V as a float voltage is not. Maybe the answer is 3.6V or 3.55 /3.50V but with a time limit after you get over the knee?

 

[nicknorman]

32 minutes ago, Dr Bob said:

Yes, get total voltage from the sum of cell voltage. Nice!

For 100% you shouldnt need current. Just terminate when one cell gets to 3.6V If the bank is not balanced then you may not be at 100% but you shouldnt go over 3.6V anyway. This sounds too easy and I wonder why MP didnt just terminate on this? Am I missing something? @MoominPapa what do you think? Why did you not just have your 100% charge termination based on the first cell hitting 3.6V rather than a total voltage/tail current value?

There is one thing that concerns me and that is the behaviour of these cells as they get to 100%. It is not clear at all from the literature and I have never pushed the cells to this point. Bear with me as I explain. If the 4 cells are perfectly in balance and all cells hit 3.6V at the same time then that is 14.4V. That sounds high. Say then the charge source is at 14.38V. That would not trip the 3.6V limit. However charge would still go into the cells and the current would start to drop as it approached 100%. At this voltage the cells would go to 100% and maybe over without ever hitting the voltage limit. All the literature says you can go to 3.6V per cell without damage but is that right. We have discussed the issue with regard to charging with solar (Peter has mentioned this a lot) but if we charge with 13.8V can you wreck the cells? The percieved wisdom is yes....dont have a 13.8V float. The 3.6V is therefore OK when charging normally but 3.6V as a float voltage is not. Maybe the answer is 3.6V or 3.55 /3.50V but with a time limit after you get over the knee?

 

For the 100% charge as I understand it you would allow the voltage to rise to 3.6v then hold until the current tails off to some % of capacity. Not sure what what %, maybe 5-10%. Then reduce the voltage to perhaps 13.4 so that no more current flows into the batteries. If you just stop as soon as you reach 3.6v, you will be quite a way off 100%, depending on alternator output vs battery capacity. In the case of our 175A alternator, a long way off.

 

I think the point is that being at 3.6v or even a little higher, is not detrimental for a while. What is to be avoided is holding it at 3.6v all day. Rapid damage doesn’t seem to occur until 4.2v or so, for LiPOFe4

Edited November 14, 2019 by nicknorman

[Dr Bob]

15 minutes ago, nicknorman said:

For the 100% charge as I understand it you would allow the voltage to rise to 3.6v then hold until the current tails off to some % of capacity.

No I think we are all doing it on total voltage to allow to rise to say 14.0V then hold until current tails off to 4% of capacity. 3.6V for a cell is fully charged as I understand it.

 

 

Then reduce the voltage to perhaps 13.4 so that no more current flows into the batteries.

Yes, below 13.4V there is no danger to the cells.

 

If you just stop as soon as you reach 3.6v, you will be quite a way off 100%, depending on alternator output vs battery capacity.

No. Again as I understand it at 3.6V a cell is fully charged and I am not sure of the  dependcy on alternator output as my charge current is reducing down to the 4% tail current well before any of the 4 cells gets to 3.6V.

 

I think the point is that being at 3.6v or even a little higher, is not detrimental for a while. What is to be avoided is holding it at 3.6v all day. Rapid damage doesn’t seem to occur until 4.2v or so, for LiPOFe4

Well maybe not. Not sure holding at 3.6V is safe for any length of time.

All the literature warns not to go into float but terminate charge straight away when full (or when full enough!). I am following MPs guidance and taking full as a voltage (say 14.0V) and then when the tail current decays to 4%. In the example before if you terminate at 3.6V by cell voltage but the charge source only gets to 14.3V max then you may never see one cell at 3.6V so the charge doesnt terminate, current drops to below 4% and then on to zero and you knacker your cells. If you therefore drop your cell target to 3.5V and have a big inbalance, then you will not be at 100%. I think this is why MP uses total voltage and tail current as the only way to get to 100%. I therefore need to backtrack on what I said earlier and tail current is the only way to know you are at 100%.

The thing is here, that you dont need to get to 100% and you really can get there manually if you want to. On my system, on the 3 occasions I have taken mine to 100%, I have gone up to 3.5V max and then kept a very close eye on voltage. Mine is 95% on 3.5V on the top cell although that does depend on balance. Have a look at the record photos I posted earlier and you can see what the cell voltages are doing in the final 5%. I think you could use cell voltage with some time function to know you are over 95% but not 100%.

Hope that helps.

 

[nicknorman]

5 minutes ago, Dr Bob said:

 

Hmmm, for some reason it’s not quoting your post. Never mind.

I think it is important to consider that not all installations are the same as yours. I am in particular thinking of your alternator output vs battery bank size. Without trawling back through past posts, I seem to recall that your alternator is set to 50A and you bank capacity is around 400Ah? Something like that. In other words, you maximum alternator output is only about 10% of capacity. And the current drops off as you approach regulated voltage. This is why when you reach 3.6v you can consider the cells fully charged. But with a much bigger alternator, say our 175A one, the cell voltage will be elevated by ohmic resistance and reaction time. So stopping charging at 175A and 3.6v would be well below 100%. My system would hold the voltage at 3.6v, initially 175A (well maybe 150A for temperature reasons!) and the current would then gradually tail off asymptotically. Only when it got to 3.6v AND a lowish current, could it be considered fully charged. The regulated voltage range of the regulator chip is 10.6v - 16v so there should be plenty of headroom to keep the alternator at max output (allowing for temperature considerations) up to a cell voltage of 3.6v.

 

The point of top balancing is surely that all the cells reach 3.6v more or less at the same time. This is why it would be a good idea to have some balancing algorithm built in to the system - after all, there is support for it in the chip. Even if the rate at which balance was achieved was quite slow (due to heat dissipation issues), once the cells are top balanced at the outset, surely they shouldn’t go out very quickly and so only mild balancing would be required on a regular basis.

[Dr Bob]

1 hour ago, nicknorman said:

 

I think it is important to consider that not all installations are the same as yours. I am in particular thinking of your alternator output vs battery bank size. Without trawling back through past posts, I seem to recall that your alternator is set to 50A and you bank capacity is around 400Ah? Something like that. In other words, you maximum alternator output is only about 10% of capacity. And the current drops off as you approach regulated voltage. This is why when you reach 3.6v you can consider the cells fully charged. But with a much bigger alternator, say our 175A one, the cell voltage will be elevated by ohmic resistance and reaction time. So stopping charging at 175A and 3.6v would be well below 100%. My system would hold the voltage at 3.6v, initially 175A (well maybe 150A for temperature reasons!) and the current would then gradually tail off asymptotically. Only when it got to 3.6v AND a lowish current, could it be considered fully charged.

Yes on my bank size etc.

I have only ever taken mine to 100% at 30A charging rate. I see the tail current decrease starting around 13.9V.

MP usually charges higher - say 40-60A and his voltage IIRC is around 14.0V as the current starts decreasing.

From what I have read, if you go up to 70A or then 100A charging, you will be seeing voltages around 14.1 or 14.2V as the tail current shows you are full.

I have not read anything about currents of 150A but everything I have read says dont charge above 3.6V. If balanced , 3.6V would give 14.4V for the full bank (optimistic as it wont be fully balanced) so say 14.3V. The advantage of having a bank of batteries is that you can test them and understand how these voltages vary.  I am guessing that charging at say 150A would need a voltage of around 14.4V to be full but the charge will be rising very fast and 3.6V would be the limit. Be good to get @MoominPapa s input here. Once you get your batteries installed you can try it out....but I will go back to my last post, that you can guaruntee you are at 100% if you monitor the drop in tail current but you should be able to get near to full with a cell voltage limit.

 

2 hours ago, nicknorman said:

 

The point of top balancing is surely that all the cells reach 3.6v more or less at the same time. This is why it would be a good idea to have some balancing algorithm built in to the system - after all, there is support for it in the chip. Even if the rate at which balance was achieved was quite slow (due to heat dissipation issues), once the cells are top balanced at the outset, surely they shouldn’t go out very quickly and so only mild balancing would be required on a regular basis.

Absolutely worth having a balancing algorithim but then we are back to the discussion of a few days ago of only using cell voltage when you get near the top or if you get to the bottom. Problem then is if you are always between 30-80%, you are on the plateau and you wont be balancing much or balancing incorrectly. Over 3 months my system went from a 50mV inbalance to a 150mV at charge termination on 100%. Not a clue why. I think Tom has suffered the same which is why he bought the balancer module. They dont need balancing if you are only going to 80%. 

[nicknorman]

1 hour ago, Dr Bob said:

Yes on my bank size etc.

I have only ever taken mine to 100% at 30A charging rate. I see the tail current decrease starting around 13.9V.

MP usually charges higher - say 40-60A and his voltage IIRC is around 14.0V as the current starts decreasing.

From what I have read, if you go up to 70A or then 100A charging, you will be seeing voltages around 14.1 or 14.2V as the tail current shows you are full.

I have not read anything about currents of 150A but everything I have read says dont charge above 3.6V. If balanced , 3.6V would give 14.4V for the full bank (optimistic as it wont be fully balanced) so say 14.3V. The advantage of having a bank of batteries is that you can test them and understand how these voltages vary.  I am guessing that charging at say 150A would need a voltage of around 14.4V to be full but the charge will be rising very fast and 3.6V would be the limit. Be good to get @MoominPapa s input here. Once you get your batteries installed you can try it out....but I will go back to my last post, that you can guaruntee you are at 100% if you monitor the drop in tail current but you should be able to get near to full with a cell voltage limit.

 

Absolutely worth having a balancing algorithim but then we are back to the discussion of a few days ago of only using cell voltage when you get near the top or if you get to the bottom. Problem then is if you are always between 30-80%, you are on the plateau and you wont be balancing much or balancing incorrectly. Over 3 months my system went from a 50mV inbalance to a 150mV at charge termination on 100%. Not a clue why. I think Tom has suffered the same which is why he bought the balancer module. They dont need balancing if you are only going to 80%. 

When you see the current decreasing starting at about 13.9v, this is a feature of the alternator regulator (or in your case, is it the A2B?), not the batteries. In order to minimise charge time you want the voltage held precisely at 14.4v (3.6v/cell) whilst the current slowly decreases to the “fully charged” tail value. In other words, CC followed by CV charging. Linear alternator regulators can’t do this easily because any linear control loop has what I call static droop (that might be a helicopter governor expression). In essence this is why the current tails off before the regulated voltage is reached. The amount of droop is a function of the gain of the control loop, but the trouble is if you increase the gain too much you get instability/oscillations. This is why all alternator regulators seem to have a very soft knee between CC and CV modes, they just slide into each other in a fairly vague way.

So when fast charging to 100% I would use a cleverer control loop in software -PID which stands for proportional, integral, derivative. Proportional is a control input related to the error between actual and desired voltage and what a conventional regulator uses. Derivative (or is it differential) is the rate of change term, designed to give rapid response to a sudden change in demand. Integral is the integral wrt time of the error. So over time, this term reduces the error to zero. Ie precise CC turning at an instant into CV once the target voltage is reached.

 

Put simply the target voltage would be significantly higher than 14.4v whilst the alternator was in CC mode (and thus unable to actually reach that voltage) but as soon as the voltage reached 14.4v, the regulator’s target voltage would be ramped down slowly towards 14.4v so that the actual voltage remained at 14.4v.

 

When I say 14.4, of course it would actually have to look at the highest cell’s voltage and keep it at 3.6v whilst operating the balancing MOSFET. Yes I get that balancing can only be done at the top (or bottom) knee. But I also think it worth bearing in mind that the primary reason why you and others only charge to 80% is because that is easy to do safely with low tech. If the system could reliably, safely and automatically take the batteries up to near 100% whilst ensuring that no cell went over 3.6v, why not just charge to 100% routinely? Yes the battery life might be slightly shortened but with such long battery lives does that matter? You would take the usable charge from 60% of capacity to 80%, ie 33% more.

 

Oh and by the way, the switch would have a third position to charge only to ~50% or whatever, when it was intended to leave the boat depowered for a while - ie storage mode.

 

ps it does occur to me that the steel hull of a narrowboat represents a great heatsink for the balancing resistors. Just not quite sure how to attach those wire wound resistors to the hull without drilling holes in it - which might be a bad idea!

 

Finally, just thinking what a balancing algorithm would be like but actually it’s quite simple - any cell hitting 3.6v would have its MOSFET/resistor turned on to partially bypass it. Gradually, more and more cells would hit 3.6v. Once all cells hit 3.6v, turn off all the MOSFETS, job done.

Edited November 14, 2019 by nicknorman

[Dr Bob]

4 minutes ago, nicknorman said:

When you see the current decreasing starting at about 13.9v, this is a feature of the alternator regulator (or in your case, is it the A2B?), not the batteries. In order to minimise charge time you want the voltage held precisely at 14.4v (3.6v/cell) whilst the current slowly decreases to the “fully charged” tail value.

No, I have  only used my IP22 charger (which puts out a constant 30A) to get to 100%. I am not sure of what the voltage is it is putting out but the battery voltage rises all the time. I assume the voltage the battery charger is trying to get to is higher than the target voltage. See attached charge curve. If you charged at higher current then the lines would be steeper.

All of the data I have presented has not come from the alternator as that backs off its charge (via the AtoB) at 13.8V.

 

10 minutes ago, nicknorman said:

 

Put simply the target voltage would be significantly higher than 14.4v whilst the alternator was in CC mode (and thus unable to actually reach that voltage) but as soon as the voltage reached 14.4v, the regulator’s target voltage would be ramped down slowly towards 14.4v so that the actual voltage remained at 14.4v.

 

 

I am not sure about the constant voltage stage. I understand what you are saying but if you reproduced these curves with 150A charge, you can simply pick the target voltage and tail current termination point. I am finding the data is almost perfectly reproducible with the data the same each time I charge to full. Once you do this graph, you can then look at the cell voltages at that termination point and see if that voltage is good. You may well have a better and more robust idea. Worth trying out.

[MtB]

I'm beginning to think I need a hefty cable connected to each interlink, eight ammeters and volt meters permanently wired in, then I can easily charge or discharge each cell individually every once in a while to keep them balanced, whenever I deems it necessary.

 

And I thought desulphating a set of LAs was a load of mucking about!

 

 

[MoominPapa]

1 hour ago, Dr Bob said:

Yes on my bank size etc.

I have only ever taken mine to 100% at 30A charging rate. I see the tail current decrease starting around 13.9V.

MP usually charges higher - say 40-60A and his voltage IIRC is around 14.0V as the current starts decreasing.

From what I have read, if you go up to 70A or then 100A charging, you will be seeing voltages around 14.1 or 14.2V as the tail current shows you are full.

I have not read anything about currents of 150A but everything I have read says dont charge above 3.6V. If balanced , 3.6V would give 14.4V for the full bank (optimistic as it wont be fully balanced) so say 14.3V. The advantage of having a bank of batteries is that you can test them and understand how these voltages vary.  I am guessing that charging at say 150A would need a voltage of around 14.4V to be full but the charge will be rising very fast and 3.6V would be the limit. Be good to get @MoominPapa s input here. Once you get your batteries installed you can try it out....but I will go back to my last post, that you can guaruntee you are at 100% if you monitor the drop in tail current but you should be able to get near to full with a cell voltage limit.

 

Absolutely worth having a balancing algorithim but then we are back to the discussion of a few days ago of only using cell voltage when you get near the top or if you get to the bottom. Problem then is if you are always between 30-80%, you are on the plateau and you wont be balancing much or balancing incorrectly. Over 3 months my system went from a 50mV inbalance to a 150mV at charge termination on 100%. Not a clue why. I think Tom has suffered the same which is why he bought the balancer module. They dont need balancing if you are only going to 80%. 

To understand this, it's really worth reviewing CC/CV charging. The principle is that the batteries are charged first at constant current (actually, normally, the maximum current avaliable from the alternator). In this mode the current is constant and the voltage increases as the SoC increases. Once the voltage reaches the threshold, the voltage stops increasing: this is not a function of the batteries, but a function of the charger: It reduces current to limit the voltage. In this mode, the current drops as the SoC increases.  That's which charge termination is a function of charge current: in CV mode: the voltage gives you no information about SoC. (actually, this is an over-simplification, the voltage must increase a _bit_ during CV charging to provide the information needed to reduce the current.)

 

What you're proposing is essentially to miss out the CV phase: once the system reaches CV state: you declare the charge complete and terminate.

 

So, what is the effect of eliding the CV phase? The answer answer is that it depends on the magnitude of the CC current. If the current is small, then the SoC will be nearly 100% when the CV voltage is reached, and CV phase would be very short anyway, so there is very little difference. If the CC current is large, then the SoC when the CV voltage is reached will be far from 100%, and missing out the CV  phase will ignore a large portion of the battery capacity. This is why your oft-repeated statement that 13.8 corresponds with 80% SoC is nonsense. It may be true for your charging system, but it's not universal: it depends on the maximum available charge current.

 

As an observation: in a true CC/CV system, the actual final SoC  is relatively weakly related to value of the transistion voltage. It's more strongly related to the cut-off current. The time taken to reach final SoC is somewhat related to the transition voltage: the higher the voltage, the quicker final SoC is reached, especially on high-current systems.

 

MP.

 

[nicknorman]

41 minutes ago, Dr Bob said:

No, I have  only used my IP22 charger (which puts out a constant 30A) to get to 100%. I am not sure of what the voltage is it is putting out but the battery voltage rises all the time. I assume the voltage the battery charger is trying to get to is higher than the target voltage. See attached charge curve. If you charged at higher current then the lines would be steeper.

All of the data I have presented has not come from the alternator as that backs off its charge (via the AtoB) at 13.8V.

 

I am not sure about the constant voltage stage. I understand what you are saying but if you reproduced these curves with 150A charge, you can simply pick the target voltage and tail current termination point. I am finding the data is almost perfectly reproducible with the data the same each time I charge to full. Once you do this graph, you can then look at the cell voltages at that termination point and see if that voltage is good. You may well have a better and more robust idea. Worth trying out.

Ok so to a large extent you are masking the issue by having a very small charge current compared to battery capacity. But you can slightly see the issue - the current start to fall before the regulated voltage is reached ie between B and C. And actually, the current was decreasing somewhat before B. This is the “soft” area of regulation . At C your charger is shutting off when the current is still 24A, so you aren’t fully charged, but you are close. If you did this with a much bigger charger, say a 100A Combi charger like we have, the duration between B and C would be much longer and you would be wasting time charging at unnecessarily low current. Equally if the charger cut off as soon as the voltage reaches 13.9 with the current say still at 70A, you would be much less than fully charged. This is why, with a big alternator or charger, you need the CV phase ie holding the voltage at (imo) highest cell voltage 3.6v, whilst the current asymptotically decreases to say 5% of capacity. And of course it is during this phase that automatic balancing can take place. With your graph, there is no significant scope for balancing.

 

ps totally agree with MP.

 

 

Edited November 14, 2019 by nicknorman

[MoominPapa]

30 minutes ago, Mike the Boilerman said:

I'm beginning to think I need a hefty cable connected to each interlink, eight ammeters and volt meters permanently wired in, then I can easily charge or discharge each cell individually every once in a while to keep them balanced, whenever I deems it necessary.

You will, of course, suitably fuse all this spaghetti close to the battery terminals?

 

I'm finding that my 200mA balance system is keeping the cells automatically balanced without any messing around. If I need to manually balance, I tell the computer to do it. I can only change the balance by 5 Ah per day, but I just tell the computer how much I want, and it does it over the next few  days without me having to worry.

 

MP.

 

Edited November 14, 2019 by MoominPapa

[nicknorman]

21 minutes ago, Mike the Boilerman said:

I'm beginning to think I need a hefty cable connected to each interlink, eight ammeters and volt meters permanently wired in, then I can easily charge or discharge each cell individually every once in a while to keep them balanced, whenever I deems it necessary.

 

And I thought desulphating a set of LAs was a load of mucking about!

 

 

Not really! If you want to do it manually, you need a switch and resistor across each cell. You turn on the switch on the higher cells to dump charge through the resistor, so that the higher cells are brought down to the lower cells. Not really feasible to do unless the battery is fully charged. Better to have an automatic system for it though - fit and forget!

[WotEver]

On 12/11/2019 at 11:36, Dr Bob said:

What thickness wire and how long to get a current of 20-30A at 3.3V (or even 10A)?

The resistance of a copper cored cable is approximately ohms/km (or milliohms/metre) =18/size in mm²  
 

You can do your own maths

 

[MtB]

2 hours ago, MoominPapa said:

You will, of course, suitably fuse all this spaghetti close to the battery terminals?

 

Of course. That is something that disturbs me with that young lad's videos. He never puts fuses in his wires on the interconnects. 

 

But nor does anyone else it seems. Maybe their wires are made of fuse wire...

 

 

 

[Dr Bob]

2 hours ago, MoominPapa said:

So, what is the effect of eliding the CV phase? The answer answer is that it depends on the magnitude of the CC current. If the current is small, then the SoC will be nearly 100% when the CV voltage is reached, and CV phase would be very short anyway, so there is very little difference. If the CC current is large, then the SoC when the CV voltage is reached will be far from 100%, and missing out the CV  phase will ignore a large portion of the battery capacity. This is why your oft-repeated statement that 13.8 corresponds with 80% SoC is nonsense. It may be true for your charging system, but it's not universal: it depends on the maximum available charge current.

 

OK, thanks for jumping in and putting me straight on the high Amps charging. I understand fully what you are saying but not sure what the actual effect of high Amps will be. At 150A charging what do you estimate 'far from 100%' means or  put another way how long in minutes will the CV phase be? Does that mean that the true measure of 100% is when the tail current drops to 4% (or wotever) after the highest voltage is reached?

Of course charging at high amps the 13.8V value is wrong for 80% but it is approx Ok in the 30-50A range for my system.

[Dr Bob]

2 hours ago, nicknorman said:

 At C your charger is shutting off when the current is still 24A, so you aren’t fully charged, but you are close.

 

Yes, not quite 100% ...need to let the tail current drop more ....but it was terminated by the voltage triggering the relay in the BMV and the auto disconnect working. The charger would have carried on putting the current in. I only use the charger on the few occasions I go to 100% (or near) so it doesnt automatically shut down before full despite it is on the victron LiFePO4 setting -hence it is isolated by the 'emergency' disconnect.

[MoominPapa]

18 minutes ago, Dr Bob said:

OK, thanks for jumping in and putting me straight on the high Amps charging. I understand fully what you are saying but not sure what the actual effect of high Amps will be. At 150A charging what do you estimate 'far from 100%' means or  put another way how long in minutes will the CV phase be? Does that mean that the true measure of 100% is when the tail current drops to 4% (or wotever) after the highest voltage is reached?

Of course charging at high amps the 13.8V value is wrong for 80% but it is approx Ok in the 30-50A range for my system.

 

Difficult for me to estimate, especially as I don't (can't) charge at 150A either.  When I was doing initial test with a lab power supply and 3.5Ah cells, it was possible to get even deeply discharged cells over 3.6V with a charge current of 3C, so in theory the whole of the charging cycle could be in CV mode with a large enough charge source.

 

It's certainly true that the tail current is the marker for 100% charge. The voltage part of the test is just there to discriminate against the situation where the current drops because a large load is using most of the alternator output, or the alternator is being turned very slowly by an idling engine, or similar. 

 

One thing I've found is that using charge termination to re-sync an Ah counter works much better if the charge termination point is consistent. It doesn't matter too much is there are always  10% if the Lithium ions in the wrong place when you declare charge termination, but if there are sometimes 5% and sometimes 15% then that really degrades the accuracy of ongoing SoC determination. Getting consistent charge termination is made more difficult by varying loads on the system, and changes in alternator regulation with engine speed, for instance.

 

MP.

 

 

 

 

 

[Dr Bob]

20 minutes ago, MoominPapa said:

 

Difficult for me to estimate, especially as I don't (can't) charge at 150A either.  When I was doing initial test with a lab power supply and 3.5Ah cells, it was possible to get even deeply discharged cells over 3.6V with a charge current of 3C, so in theory the whole of the charging cycle could be in CV mode with a large enough charge source.

 

It's certainly true that the tail current is the marker for 100% charge. The voltage part of the test is just there to discriminate against the situation where the current drops because a large load is using most of the alternator output, or the alternator is being turned very slowly by an idling engine, or similar. 

 

One thing I've found is that using charge termination to re-sync an Ah counter works much better if the charge termination point is consistent. It doesn't matter too much is there are always  10% if the Lithium ions in the wrong place when you declare charge termination, but if there are sometimes 5% and sometimes 15% then that really degrades the accuracy of ongoing SoC determination. Getting consistent charge termination is made more difficult by varying loads on the system, and changes in alternator regulation with engine speed, for instance.

 

MP.

 

 

 

 

 

Ok thanks for that. I had forgotten the stuff you did on the 3.5Ahr cells. That makes sense now and I can see why you would have a long CV mode at that charge rate. 

I don't have the issue of re syncing as I rarely get above 80% ...although I guess I could get it to sync at that. I am just relying on voltage to tell me where I am.

It would be good if you could re sync the BMV at 80% and it display 80% rather than 100%. 

 

 

Eta.... when you commercialize your system, can I buy one!

Edited November 14, 2019 by Dr Bob

[rusty69]

10 hours ago, Dr Bob said:

Eta.... when you commercialize your system, can I buy one!

Yep, me too please MP.

 

All this top balancing, bottom balancing LVD,load dissconnect.LA dump load. Low temperature disconnect, High temperature disconnect, high voltage disconnect, cell monitoring,accuracy, repeatability,resolution, disollution 2P4S, 3P4S, C3P0, R2D2,7of9 drop-in's, drop-outs, shake it all abouts, 1C,2C,3C,ICU,BMS,BCG,BFG,ABYC.....Top knee, bottom knee, water on the knee..........................I JUST WANT TO CHARGE MY PHONE.

Edited November 15, 2019 by rusty69

[nicknorman]

Discussing these things and having various ideas thrown around is really helpful in coming up with a design. So my overall concept is now:

 

Two separate modules:

 

Module 1 is a measuring/monitoring/protection module attached to the batteries and comprises the AD7280A battery monitoring chip controlled by a PIC micro. This would be responsible for measuring cell voltages and battery temperature, operating an audio alarm and a mechanical isolation switch in the event of emergency, and operating the balancing MOSFETS/resistors. Rather than reinvent the wheel, it could optionally incorporate the BG-8S for a monitoring display. Information about cell voltages etc would be sent down a serial data link. This would be permanently powered, but spending a lot of time sleeping when the engine isn't running, so the overall current drain would be perhaps be a couple of mA plus whatever the BG-8S uses (can the BG-8S be turned off without disconnecting all 5 (or 9) wires?)

 

Module 2 comprises the AR6000 alternator regulator chip controlled over LIN by a PIC micro. It would receive the serial data from module 1 so it knows individual cell voltages. It would only be powered when the ignition was on (well the AR6000 would be permanently powered but it only uses 60 microamps when "sleeping"). The regulator chip has a stand alone mode that would be set to a low safe regulation voltage, say 13.6v - it would go to this mode a few seconds after no longer receiving LIN communications from the PIC. ie fail safe. Otherwise, the PIC would send target regulation voltage commands and maybe field current limit commands to the regulator. It would base this on the cell voltage information received via serial comms, alternator temperature, battery current etc. If it lost comms with Module 1 it would go into a fallback mode ie regulation at 13.6v.

 

That's it really, what could possibly go wrong?

 

Anyway, I need to leave the land of dreams for a while, this project will be for next year at the earliest!

Edited November 15, 2019 by nicknorman

[Dr Bob]

28 minutes ago, nicknorman said:

 so the overall current drain would be perhaps be a couple of mA plus whatever the BG-8S uses (can the BG-8S be turned off without disconnecting all 5 (or 9) wires?)

 

 

I am pretty sure you just disconnect the top wire which is the -ve from the battery. The display is dead when you do that.