 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Electronics : a new way to balance lithium ion battery
  Page 5 of 11   |
|||||
| Author | Message | ||||
| Warpspeed Guru  Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
It may already exist, but someone somewhere would have been the first guy to do it. Not sure I have ever seen a battery monitoring system that can both monitor and balance very large numbers of cells with a large screen display that requires only about a watt to run, but I could be wrong. Cheers, Tony. |
||||
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Tony, perhaps you have not noticed me mentioning that I *no* longer use ultracapacitors. The charge transfer from 50 Farad was simply too slow on my system which transferred charge from the high cell to the low cell. At least 3 seconds connect time plus 0.5 second disconnect (to settle all relay contacts) adds up over 16 cells. I inject a fixed (adjustable) voltage now. Each cell takes whatever current it requires to equalise. The timing is still as above but equalising is faster. Its done in sequence rather than having software trying to figure out which cell requires extra charge and which cell is donating it. If you tinker with relays, there *must* be a period when all relays are off. Its way too risky to count on the changeover contacts alone to prevent 2 cells being connected at once. Klaus |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
With the way I plan to connect up all the relays you could operate all the relays together or in any combination and nothing bad would happen. Its pretty fail safe. I did not realise you no longer used capacitors to balance. From what I have read it does not take very much. Those cell balancers using a shunt voltage regulator could only safely dissipate perhaps a couple of hundred milliwatts. Just reading up on the Nissan Leaf EV battery, it uses switched 420 ohm surface mount resistors across 65 Ah cells to balance. That is only 10mA and 40mW. Anyhow, the flying capacitor method of voltage measurement will inherently aid balance even if the effects are very small. As I plan to run it continuously its not going to do any harm. How are the Winston Thundersky batteries going ? I may get some of those myself. Cheers, Tony. |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Tony, I thought mine was fail safe too, that is, before I realised just how little contact clearance there is in those relays. And the power of a 10KW battery bank is amazing when let loose on relay contacts & PCB tracks. I still would advise to run the relays with a distinctive 'all off' period between each step. And add fuses, they are easier to replace than blasted PCB tracks. I'd been there and done that when it comes to bust things. And I have a nearly full shoebox of discarded PCB's before I came to my present set up. I do have a shunt regulator across each cell too. Mine can shunt up to 4A. But, these days the shunts don't do anything, they only operate if a cell voltage gets too high. This is not happening once the bank was balanced. So the shunt regulators are just insurance now. The Winston batteries have exceeded my expectations of them. By a long shot. Yes, they are expensive but with LA I would need 2 or 3 times that capacity to get a similar performance day to day. I would have 3 days max battery storage with my 10KW bank if the sun goes truly into hiding. I would like I had another 200Ah bank but I don't really need it here in sunny Perth. Klaus |
||||
yahoo2 Guru Joined: 05/04/2011 Location: AustraliaPosts: 1166 |
I have one system I look after is now out to 460 days without a balance and it is still within a couple of millivolts at float voltage the other battery banks are checked and cycled every three months to check the shunts and contactors are working. the voltage spreads a bit when it gets above 3.6 volts but we are still only talking 20-30 mV at the most. They might flow 250 mV for 30-50 seconds on one or two cells. For new matched good quality cells in a system that is not run to extremes it is going to be years until the balance circuit does much. I dont know about used EV cells, factory seconds or cells that have been hot or overcharged. perhaps they would need some more management. The most important thing is an individual cell low volt disconnect, if the BMS cant do that then sooner or later a cell is going to get driven in reverse under load and it will be toast! busman, I am a little perplexed by your draw from 6 cells. I have had trouble with the laminated terminal connectors working the bolts loose on a portable pack, that give me some funny voltages. I guess heat expansion causes it. Anyway i replaced them with braided connectors and it has been fine since. I'm confused, no wait... maybe I'm not... |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
From what I have read, if you do not mercilessly thrash these lithium cells they stay pretty much in tune with each other. Agree absolutely that individual cell voltages should be monitored, and with an under voltage and over voltage battery disconnect that can be tripped by any one cell. Klaus, glad to hear you are very happy with your Winstons. I have replaced more than a few relays that had open coils over the years, that seems to be the main failure mode. Burned contacts are more a design stuff up rather than a reliability issue, and small fuses at each cell for the monitoring wiring are absolutely essential. Finding suitably robust relays for this should not be a major problem, although they will probably not be cheap. I will throw it all together initially to see some results, then after probably about a year get some really nice circuit boards made once all the changes and inevitable modifications are finalised. I think we may need more than one layer of protection that can operate quite independently. These batteries are too valuable to leave much to chance. Cheers, Tony. |
||||
| busman Newbie  Joined: 30/10/2016 Location: AustraliaPosts: 21 |
In reply to Yahoo2: My banks in the motorhome (200 ah @ 12v winstons and 800 ah @ 24 v sinopolys are now over 3 years old. At that time about all we could find for monitoring/protection were Junsi cell loggers, so the banks were set up using those as protection for out of balance cells. One was set so if a Junsi alarm call for an over volt cell happened, a voltage stepper came into ply, sending a higher than programmed voltage (voltage sensing wires) to the Rich Supercombi power management centre. This made the SC think battery was over, so it told the 3 solar regs (all data cabled together) to stop charging. At the same time a water heating element was switched on to "pull back" any errant cell. After a set time, usually 5 mins or so (set by timer) it all reverted to what was happening before the vent. A counter told me how many times this had happened, or how much hot water I had ! Another junsi was set to set off a piezo alarm on high or low voltage. The junsi only drew it's operating power from the first 6 cells, no problem for the 12 v bank, but 2 of them on 24/7 created an imbalance meaning you had to stop charging at a lower voltage and charge at a slower rate or 7 and 8 would run away. Remember this was years ago and it was a real problem. Now both banks are fitted with balancers, probably not necessary for the 12 v bank as they seem to stay pretty balanced anyway, but essential for the 24 using the earlier regime. There has been call since then. In fact the Junsi is disconnected, just plug it in occasionally to check, never more than 20 mv.Plug it in when travelling though. However I have added a different level of protection. On the battery positive lead I have installed 2 levels of protection, one is a high/low voltage cut out from VDO. The levels are adjustable and this drives a Gigavac latching relay. So if voltage goes past the high or low parameter the Gigavac "unlatches" and stays that way until I switch a current to it to "latch" it again. This way, if I am away the battery disconnects and stays disconnected. I also have another high voltage only device driving a NO Gigavac contactor. If the voltage goes over 30 this one drops out allowing the contactor to go to its normally open state. So fail safe is taken care of. Why the 2 units ? It is electronic, therefore at some time in the future it will fail, maybe tommorow, maybe 100 years, I don't know when.  Also I use solid copper bars 4 mm thick to join my cells (64 in big bank) with no problem, been through a wide range of temps as well. I know the logic says you must buy expensive links but that is not my experience, the copper bar is many, many times cheaper than links. Oh, and from my experience, with a balancer in place you do not really need much else. If I was to set up another system it would have a balancer and similar high/low voltage protection. These cells can take a lot of abuse, even if you had a single runaway it would not get to any dangerous area, if you have your parameters tight. So, there you are, my current BMS, all off the shelf, anyone want to buy version 1 ? Cheers William |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
William, that is pretty much exactly what I plan to do eventually. Instead of a Junsi, I am home brewing something that does pretty much the same thing but will monitor more cells and have a much larger display that records peak max and min cell voltages on screen. As its my own software I can do pretty much anything regarding voltage cut out thresholds or alarms. Instead of a VDO, I will be using a programmable voltmeter that has two separate fully adjustable alarm outputs. http://www.lightobject.com/Programmable-4-Digit-Red-LED-ACDC-Volt-Meter-with-Dual-Control-Good-for-HHO-System-P408.aspx That will just monitor total battery voltage as a backup. Have not decided what type of bistable cutouts to use yet, but will definitely look at those Gigavac contactors. Also the charger and inverter have their own shut downs. It seems like the most logical solution, glad to see someone else has very similar thinking. Cheers, Tony. |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Hi Busman, Your gigavac latching relays aroused my interest as I was not aware of that brand. However, on their website there seem to be only high voltage latching relays I could find. What I would require; to replace the 12V 80A latching relays I had used with my 24V battery bank but are too low voltage for safe usage on the 48v bank; is a latching relay that could disconnect 150A @ about 56V DC. This would be the worst case situation should something short out the bank while I'm away. What rating is the relay shown in your picture? And where did you get it from? With my 24v battery bank I used a back to back 120A schottky diode in series of the battery cable. These diodes were normally shorted out by a latching relay across each. Over voltage would trigger one relay, preventing further charge flowing into the battery but it still powering the load. Under voltage would disconnect the load but leave the charging side intact. Thanks. Klaus |
||||
| busman Newbie Joined: 30/10/2016 Location: AustraliaPosts: 21 |
Hello Klaus, Gigavac is well known, they do make a great product. My bistate (latching) relay is http://www.gigavac.com/sites/default/files/catalog/spec_sheet/gxl14.pdf 350 amps. Only listed as 12/24 v coil but I have found them to be very easy to work with and would not be surprised if they will make one with your required coil voltage I brought direct, the guy you want to send an email to is Marcus, send it to info@gigavac.com and mark it for his attention and ask if they can make one of the above with your size coil, I would not be surprised if they say yes. Not cheap but then what piece of quality equipment is ? About $200 by memory. I have about 4 in my setup, I use 2 NC to divert 2Kw of solar to my hot water heating, through 3 pairs of elements, each a 12 and 24 volt in series. 1/2 an hour in the middle of the day gives me 50 ltrs of very hot water. Hope this hepls, Cheers William |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Thanks William. These relays have impressive specs but, as you said, it comes at a price. I'm interested although it will have to wait until I finally get the inverter I have been tinkering with running reliably. There is plenty of extra solar power here from my 3KW of panels but it goes to waste as my hot water is instant gas. Heating water by solar is a bit down the list of things to do at this stage  . Good on yer for doing it. Klaus |
||||
| busman Newbie Joined: 30/10/2016 Location: AustraliaPosts: 21 |
Tony, thanks for the kind words. When we started with Liths information was difficult, got a lot from a guy called T1 who started a forum where many different systems were developed. But now, after all this time, other equipment has become available, knowledge has been gained and a rethink is required from what we started with. IF, you have active cell balancing in some form, and I am not convinced this needs to be 24/7, a lot of drama can be avoided. Once this is in place it is really only a matter of deciding how you want to back that system up, I no longer have the Junsi connected all the time with it's disconnect/load/timer regime because I think it unnecessary, although I do plug it in when actually travelling because we are at the front of the coach and you do not know a lot of what is happening down the back. So, it stands to reason if you have something keeping balance in check, you can set pretty tight parameters that in our case is a disconnect, I am not worried about keeping some circuits alive like a fridge, because this will be such a rare event I feel. I wish I had your hardware/software abilities, I can do a lot of stuff but came along just a bit before all this stuff became common. And I don't have the head for it anyway, wish I did. Next will be a system for the house. I am happy with the coach setup now, have added a second SuperCombi giving us 6000/12000 watts so all cooking is now induction, apart from the gas barbie. Although I must admit while giving the trailer roof a wash today I was thinking "I could squeeze another 1Kw on here if I was careful" How many RV's have 4 Kw of solar ? Not too many I reckon, I have not found anyone with our current 3 yet but they are nudging closer. Cheers William |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
You guys are way ahead of me in all of this, I still have to even get a lithium battery. But reading about the ideas and practical real world experience of others is an absolute gold mine of information. I suppose the first thing is a really good monitoring system to understand what is really going on, and to try to understand system behaviour and tweak things to stay within safe limits. Once the system is balanced and settles down, hopefully it should require minimal attention after that. Cheers, Tony. |
||||
| busman Newbie Joined: 30/10/2016 Location: AustraliaPosts: 21 |
You are overthinking it just a little Tony, if you want to know what is going on you only need to know individual cell voltages. If you have a decent way of looking at that (a Junsi is good enough) you can configure it to act as a policeman if you wish, or simply add some other equipment like a latching disconnect under tight parameters (with a balancer fitted) and that can be the end of the story. That is where our big 24 v bank is at, never look at individual cell voltages now. junsi not even connected, just the high/low cut in place as final line of protection. William |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
What I plan to do is connect my home made cell voltage logger, which will also have disconnect, and a full time (ultra low power) balancer. All this will draw very low continuous power of around one watt from across the entire battery. My system will be fairly unusual in that it is a high voltage dc system with a lot of fairly small cells to monitor. If I want to, I can switch on an LCD monitor fed from the VGA output signal from the logger and see at a glance what every cell is doing, or has been doing. Its has been a fair bit of work to develop and build all of this, but it has not been all that expensive, as its built entirely from integrated circuits I already had. As I am retired, its a slow long term project. Something to fill in the time whenever I am in the mood. And at the end I should have something fairly unique and pretty good. If I was not doing this I would be doing something else very similar anyway. When its complete I will definitely start on some other project. Many people probably see this as a lot of effort for something I could just go out and buy. But retirement changes your perspective on a great many things, including how to spend a great deal of spare time enjoyably. Cheers, Tony. |
||||
| yahoo2 Guru Joined: 05/04/2011 Location: AustraliaPosts: 1166 |
What worries me slightly Tony is what happens with a cell that is slightly higher resistance or slightly smaller in capacity in a system that constantly balances. Im thinking that that cell could be at a different voltage that the rest of the pack through a complete cycle of charge and draw and high and low charge (and temp) except for the one point/ charge rate that it was at when it was last balanced. My preference is to balance at a point that suits the batteries use. solar systems rarely get to the bottom and the current draw is fairly low so it makes sense to keep the cell voltages even at the top knee then if the pack is emptied the BMS will detect an individual cell voltage that is lower than the others at the bottom knee and stop the discharge. Electric Vehicles are pretty useless unless we can get to a charge station, mash your foot on the accelerator at low voltage and the car will see a low voltage cell even if there is still plenty of juice left and trigger the protection circuit and we are stranded. the options are to balance at the bottom knee and leave the voltages looking ragged at the top OR still balance near the top at a trickle charge rate and have a voltage margin at the bottom that is under current restriction (limp home or turtle mode) one of the first BMS units I tested was designed for a car, I didn't realise there was a delay programed in for the total pack low voltage cutout, that scared the crap out of me waiting for it to trigger. the pack was genuinely empty and it wasn't switching off  I damn near had a heart attack!I'm confused, no wait... maybe I'm not... |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
Yes, quite right, you cannot really balance a 50Ah cell with a bunch of 100 Ah cells all connected in series. A weak cell should show both a higher peak charging voltage, and a lower peak discharge voltage. If its about the same as all the others when at around the 3.2v mid point voltage, I would think that would be about optimum balance for that cell. Continuous cell balancing at very low power over a very long period, should converge to that condition, or something close to it. That is why I feel recording the peak max and min voltages reached is just as important as displaying continuous real time cell voltages. As soon as max and min start to spread significantly further apart for one particular cell, even under very light useage, that should be a pretty good early indicator of some capacity loss or internal resistance rise. Cheers, Tony. |
||||
| yahoo2 Guru Joined: 05/04/2011 Location: AustraliaPosts: 1166 |
I find when I lean on the cells with a fast charger everything is fine, it will throttle back with a 5 or 10% margin then slowly top up the pack pushing against some resistance. If we come in slow with a trickle charge there is almost no resistance and i can get to 99.9% capacity at just a few amps. @ 3.59 volts they are all even @ 3.61 volts they may not be, the full cells are starting to spike up. the manufacturers say there is some tolerance and it is dissipated as heat but I tend to read that as "acceptable cell degradation" to reach a cycle target number. a standard constant voltage charger only looks at the average pack voltage so at 3 or 4 amps the first cell to hit 100% will shoot up in voltage in a few seconds, max out the 2 amp shunt and keep climbing till it triggers the shut off. the more cells in series the more likely this is going to happen and the smaller the parallel cells the quicker the spike will be. The cells are very well behaved when they are new, I have always expected to have to change the charging strategy as they age and develop some quirks but the oldest ones I look at are still below 2500 cycles and seem to work fine, no leaky ones, no hot ones so I have kinda given up, we check the cells with a multimeter once every three months, check the shunts for faults and cycle the contactors every month and that is about it. I'm confused, no wait... maybe I'm not... |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
A few seconds ? That is terrifying. If this is true, it needs a shunt regulator on each cell hooked up through an opto isolator for instantaneous charger shut down. Not difficult to do, just difficult to believe because nobody else seems to need that. I would have thought that voltage sampling of each cell every couple of minutes would be plenty fast enough to detect a charging voltage spike on one cell. If the voltage rise is as explosive as you say, its not really a problem, if its real and expected. Cheers, Tony. |
||||
| busman Newbie Joined: 30/10/2016 Location: AustraliaPosts: 21 |
I am sure I have the most expensive pack of the people having this discussion, and a few years experience with my 2 packs as well. These things are pretty tough, some of the mistakes made in the early days (like overcharging, flattening to 0 volts, dead shorting etc (in one case for 6 weeks)) resulted in bulged cells, but many of those have been rescued and are still in work today. Everyone seems to think they are such babies, but that has not proven to be the case. So very easy to want to protect your expensive investment, in a couple of years you might just be saying "why did I waste my time on that?" when I could have been fishing or whoring, whatever gets your goat. In my experience it just ain't necessary and I have a pretty big investment in these things. If I thought it was the Junsis would still be connected etc etc, but they ain't, that must tell you something. You just do not need the level of information you guys are trying for, it can all be done pretty simply now with off the shelf stuff. William BTW we routinely see 100+ amps coming in from the solar and we can charge at 140 via the supercombis, without our latest steps, that was not possible |
||||
| Page 5 of 11 |
|||||
 Print this page |
