A couple of folks have said now that the drop in capacity is expected given the age of the cells, so I'm starting to feel a bit better about the results.
Hi, Dave and Magda from Mallorca here. We are part of a team advising and training installers and project managers/owners on marine projects of all sizes and styles. Cell 12 is showing signs of operating out of sync and differently from the others in the pack and this will worsen over time. Strong recommendation is to use one of your 4 spare cells you mentioned to swap out number 12 for one with the same calendar ageing as the others in the pack. Although the balancer will mask the issue it will eventually mean the equalization of this pack is compromised long term. If the nice supplier will rma the cell, just keep to one side for experimentation/emergencies as it's lower aging will again cause some (very minor) imbalance. Since you have the perfect spare it would be best to use that. Normally we stay quiet on RU-vid and comments, but I enjoy your channel too much not to try and help. Professionally we work with 5 main LFP marine battery suppliers and Magda had a big hand in developing/testing a US battery for the marine market. We are LFP tech heads who also happen to be life long sailors.
Thanks for the advice. I probably will swap it out before the batteries get installed on the boat. Thankfully, with a bank of 6 (or 7, if I can fit it) packs in parallel, the early discharge/loss of one pack won't be too catastrophic. Cheers
@@TheDigitalMermaid Hi there, i would probably also swap it with the spare - maybe you want to test if the cell drops quick in voltage (with deactivated balancer). If there are dentrides peaking the seperator, the cell will selfdischarge over time and the balancer will have trouble to keep it under control over time. Greetings from germany and a happy new year!
I just wanted to suggest an alternative to your resistor that would give a visual indication of your capacitor charging. If you could string some incandescent light bulbs together to match the voltage being applied to charge the capacitor and place them in series with the line instead of the resistor they would both limit the current draw and dim as the capacitor became charged. Car headlight bulbs could work and some are dual filament so you could easily connect them in series so one bulb would be good for 24 vdc. Of course you would have to be careful not to handle the bulb by the glass as it can get hot very fast. I'm old and I've used this type of tool for years. Because of the current draw it is great at finding poor connections as well where some of the young guys I was working with saw the voltage on their meter then were totally baffled when they tried to power up the circuit and the cot breaker was bad so the voltage just went away. I haven't worked much with Lithium chargers or battery packs so I sure do appreciate your show.
I said to myself “did she reset the smart shunt?” 🙂 Next scene: I forgot to reset the shunt. Laughed my ass off. Don’t worry, I can’t remember how many times I’ve heard Andy say that when he does a capacity test.
@@sebastienl2140 For that, I would need a shunt per battery... That's something I considered back in the first battery build, but gave up on for cost and space reasons.
I would like to point out that at 17:52 in the video you nailed the balance of your battery - you were done! It was balanced at 55.6V or so (your absorption voltage?) giving about 3.475V/cell all matching and the current drops! Nice. At 16:02 you were out of balance on several cells, at 17:52 you were in balance. To me that would be happy time: I am perfectly happy with a balance of +-10mV for a balance voltage is 3.45V. Once balanced the current will drop, and all cells are near your balance voltage. That’s a good thing. You are done! I use the current drop and the matching voltage as my balanced indicator. If I balance my battery on the bench (using CV /CC suppy) I use the current drop as a balance indicator. If you use an accurate current meter you will see the current drops to something in the 10mA-30mA range after a long time- (several hours 'absorbtion'). This current depends on your BMS or other things you connected. If you say: but I need to balance at 3.6V that's fine and you could increase the voltage and repeat process. But you will not really gain much in practice (vs 3.475V)- you get to choose your balance voltage and balance the battery at that point. The important thing is that you have a decent balancer- and the JK BMS you use seems to have a nice one. The important thing is you choose a balance voltage higher than 3.4V or so. 3.45V, 3.5, 3.6 all work. Some balancers work better at higher voltages, but I found the my JK is perfectly happy at 3.45V. I noticed you seem to increase the voltage to 3.65V/cell a the end? That is a pretty hot charging voltage. You could try to balance there but you would not gain much if anything.
I set the 3.65 as the cut-off voltage, so it's only a safety thing. I know there's basically not energy of note up that high. I was more keen to just make sure the cells are all as absolutely balanced as they could be before the test started. I won't do that when they're installed .
its always a possibility that temperature and some kind of wake up facts might work on such big packs.... in a permenant usage it will be so much better... HAPPY NEW year btw
Good-part-of-the-day to you 🙂 One thing that I think I might have been missing in your measurements is the voltage drop over the cables between the batteries and the inverter. As you are pulling around 22 amps on discharge (at the 7 minute mark) -it would be interesting to measure the voltage drop over those cables. As each cable has some resistance, (al be it quite low), those 600W most likely is partly dissipated in heat within those cables. This is loss that can't be measured with the shunt, but it is still power as stored in the batteries. Maybe you can check that. As a radio-amateur, I have seen voltage drops over the (original!) power cables of the IC-7300 that were not insignificant. Those cables have in line fuses that do corrode a litle and guess where the cable gets a little warmer than it would be without 🙂 Good luck and thanks for sharing your video's!
I suspect you're right. The resistance of the cables / voltage drop is pretty low over the relatively short runs and 4awg wires, but it's not zero, at all. Given how long it takes to discharge, the loss to heat is certainly a measurable component. Combine that with the calendar ageing of the cells, and I think the capacity I saw is understandable and also just fine. For me, one of the top goals has been trying to understand how much usable capacity I'll have when I'm on the boat, and these tests are a decent analogue for low-speed steaming.
@@TheDigitalMermaid I personally think you did an excellent job there. Just from an engineering point of view I couldn't think of anything else that might have been missed out. I am not an engineer by the way, just a hobbyist / amateur :) 4% indeed isn't too much but like me you would like to see the 100% usability which can happen if batteries have a little more than the given capacity which usually only happens to brand new ones. It all depends on the specifications though if sonething is within the acceptable parameters. Keep up the good work and... The sparks got me too in your previous video. All the best from the Netherlands!
And now on the elephant in the room: cell 12 voltage drop. What you have here is what I call a leaker cell. Why? You balance your battery. A week later you check (with full recharge) and you find once more that cell is low. It feels like someone drained it a bit… In your case, that cell seems to have similar capacity as other cells (it is not the lowest). It probably has similar internal resistance too. But like clockwork, a couple of weeks later (or even the next day) it will be low again when you look in the balance range. Off course to see that the cell drained a bit you need to charge into the balance range - above 3.4V and wait for current to drop. Below 3.35V you will not see this effect. You can also get the opposite - a cell that is always higher after a time - and anything in between. If you have the luxury of getting it replaced, do so, or live with it and simply rebalance occasionally doing exactly what you were doing - charge to 55.6V and hold it there until the current drops and the cells are balanced. I have dealt with a leaker cell for a couple of years and that cell does not appear any worse than the others in terms of capacity (other than requiring re-balance occasionally). There is a way to spot a leaker cell in action: apply the 55.6V with balancer on until the cells are balanced and the current drops (what I call my standard balance procedure). All cells are the same voltage (or close enough). Now: Keep the 55.6V and **turn balancer off** (keep the 55.6V steady!). Observe the voltages every couple of hours or days. You will see cell 12 drop (in real time!) and some other cells go up, some cells in between. In your case I expect you will see some drop on cell 12 fairly quickly, say a few hours. I have done an experiment with a 4S 100AH battery where I added a small adjustable resistor to each higher voltage ‘less leakage’ cells (in the 1K-4K range 3mA-1mA) that turns non-leaker cells into leaker cells and by some miracle the problem got much better once I tweaked the resistance values to match the leaker cell: it took much longer for the leaker cell to go out of balance. I am not suggesting you add resistors (it is difficult to come up with a good value) I am simply saying that the low cell acts like someone put a resistor across it compared to the other cells. That why I call it a leaker cell.... You could call it self discharge or maybe use a better term. There seems to be an online opinion that LIFEPO4 do not self discharge or leak, so I am waiting for a better explanation of my observation.... All I can say is that my LIFEPO4 seem to self discharge and each cell may have a similar or slightly different self discharge rate. If you have slightly different self discharge rate between cells, you will get your cell 12 effect. (If you are interested, my best estimate for the self discharge (leakage) rate on my LIFEPO4 100AH cells seems to be in the neighborhood of 0.5-3mA. that gives about 1300 days to self discharge an unconnected cell. I expect your ~300AH cells to be in the 1-9mA range.)
I fired up my d.i.y battery build for the first time yesterday, the 2 parallel banks are still ballancing but getting there. Thanks for being an insperation, Andy and yourself are awesome.
@@TheDigitalMermaid Going from a wee battery pack to a big battery shelf is luxurious, I love it, well worth the hard work. It has been great watching you grow and learning off of you. Now I just have to try and generate enough to charge them lol. Stay safe, Love from 🏴 Gaz
BYD, at least the packs certified for victron, are quoted at 80% capacity after 10 years. Assuming that is a cell chemistry thing you seem to be about on track after 2 years. Limiting charge and discharge voltages to keep then conservative may help a little but in an off grid application you are adding cycles by doing that. I am no expert, however from what I have seen your experiance seems pretty typical. BTW I am also running some 12V, nominal, LiFePo4 block type batteries, which have an onboard BMS. The manufacturer, Ultramax, told me to set Bulk and absorbe voltages the same and set an elevated float voltage to give the BMS something to work with. Both my inverters are Multiplus's one 48 and one 24, the former feeding the latter, which still gives me the capacity, kW, as the 24 can assist the 48. Great vid BTW, thanks. You should get VRM set up, way easier to handle configs and a great wat to log stuff. Also, do yiu know yu can run NodeRed on the Cerbo, its great for adding logic, creating a better dashboard and will talk to just about anything it can reach over a network. You can even edit flows directly from VRM, from anywhere. I use NodeRed to coordinate what my two systems are doing and control them from a single dashboard. The 48 is hooked to a CCGX, NodeRed, on the Cerbo talks to that using Modbus TCP.
I think the losses are 100% acceptable based on age alone. That’s as I suspected- I think little harm has been done by the low voltage, even though it’s obviously better to avoid this. It’s also why I use conservative top AND bottom voltages - the capacity gain in going closer to limits is negligible.
You probably know that the standard test conditions says discharging at 0.2C at 25 degree Celcius. You should see a higher capacity at the lower current drain, if all cells are OK.
It's just calendar aging most likely. Don't sweat it. The downside to collecting the materials before use is the aging happens in dry dock as it where. Tech gets faster and cheaper every year in general also but the old stuff we have still works. Just stay focused on getting it all in the boat and back on the water, it'll server just fine for plenty of years yet. Perfect is the enemy of complete a friend of mine is fond of saying. Just get it complete, make it perfect later.
The beauty of battery testing over days and days and days... just for one video. Cell12 seems to be fine and the capacity test needs 0.2C discharge current (56A), unless you have done all tests with this same load and it is therefore comparable. I use an extension cable and the Tesla as load for my battery tests so the energy at least gets recycled as much as possible😉 With 0.2C, you will even pull less energy from the cells though. 4% down is really a bit too much for 2 years, considering you haven't used the battery apart from capacity tests. Can you test another bank to compare? And what were your initial capacity results when you build the batteries?
Hrm, I struggle to find a way to load the batteries above 0.2c... The charger I have is only 13A @ 120v, and it's winter... Maybe I'll have to pickup a heater after all, I've got one on the boat I can bring home. That with the server rack might help me get the C rate higher... Though, interestingly, when the batteries are on the boat, the C-rate draw, even with both motors at full throttle, would be 21kw, or 410A (68.4A/pack, if I have six), which is still only 0.24C, and not a power draw I'd realistically hit. I'm probably going to pull closer to 50A total most of the time, or 0.03C. I did test the other 5 batteries, and (spoiler) they're all in the same 13.7kwh range. I focused on the low C-rate being a benefit for the life of the batteries, but didn't consider it would reduce the usable power. I think I need to run a discharge test with the heater, rack and anything else I can throw at it and see if I get a better result. I mean, you were right about the cell being OK, so I have no reason to think you're not right here, also. I did have a lot of replies suggesting that 4% loss over 2 years is about expected, so I'm super interested in the results. More tests!! ^_^
Usually you get slightly more capacity out of the cells, when the C-Rate is low. But with a high C-Rate, the cells heat up and you get more capacity because of the heat...the efficiency is still higher with a lower C-Rate, since less current means less heat loss due to the internal resistance of the cells. Keep testing. I am keen to see how they perform at 25°C or 30°C. @@TheDigitalMermaid
Most of your problems are due to cable resistance (voltage drop and terminal resistance). Change the cable size to 3/0 gauge (200A). Use contactors for safety.
Andy have been testing the jk BMS with newer firmware were you can adjust the reset to 100% value, with the firmware version you have it will reset when average voltage is 3,6v per cell , typically when the delta is lower than the value you have set to trigger balancing
Aye, but the BMSes I have can't be updated, annoyingly.
7 месяцев назад
You are about 4% down, which is absolutely a reasonable loss n capacity compared to brandnew from what I have read so far. With sime preassembled batteries there is already some reserve builtin so you do not notice this initial loss in capacity, but I would not worry about it too much. rule of thomb for EVs is to size your batteries in a way that two thirds capacity can give you the range you see as your minimum range. this accounts for capacity loss over time and adverse conditions like low temperature and so on, that also diminish your range. From my point of view you should go by the same ballpark numbers and enjoy the wider margin this provides while you have it.
If this is expected capacity loss, then that actually makes me feel better. Spoiler for the next video, but all the packs were down about the same amount, 13.7kwh, +/- 0.1kwh. The ultimate goal has been to determine range, and so if this is the post-new stable, then in a way it's better, as it'll give me a better answer to "how far can I go?" when I get the boat in the water.
7 месяцев назад
@@TheDigitalMermaid Tesla claims 2% on their newer Model 3 with FFP batteries, But I think they are being a bit optimistic - or have a better handle of their batteries. I would RMA Cell 12 as others have suggested for good measure, especially if you have a few spares anyways. One thing to nore for Use n the boat: LFP cells handle cold temperature charging much worse than other battrey types, below a certain cutoff that should be in your datasheets you can be relatively sure you would damage them. To avoid that install a simple silicone-sheet heater in each battery compartement and combine it with a thermostat. Does not cost much and saves you a headache. Also keep an eye on how much they heat up under the load your motor is expected to generate, If they get too warm you might neet some cooling - easiset would probably be metal coolinf fins between each cell to encrease convection cooling. Since you live in a rather cold area I would not care about it too much, but keep it in mind. I would also suggest you put a small generator on board, to avoid going to very low charging states in case you are stuck in bad weather.
@ Tesla has first picks on cells, I am certain, where we normal plebs have to get the EV rejects. Lol. So if EV companies like Tesla expect 2% loss, then I won't feel too bad. On heating; I considered the heated pads, but I am leaning towards putting a radiator under the battery packs. I'm going to have a diesel hydronic heater, which would let me heat the cabin, the water and the batteries with minimal electricity (albeing burning diesel). I do plan to be in cold places, and warm batteries deliver more power, as well as absolutely needing to be warm to charge.
I downloaded the scientific calculator app. Thank you for showing that, always had a hard time trying it through a simple calculator on the phone. My inverter requires the app to be closed as well before the settings come into effect, not sure why. I just treat it as normal procedure. 264/280 = 94.2% battery capacity remaining. I am assuming a linear decrease. Taking a 4,000 cycle charge to reach 80% battery health. .014 ah per cycle drop, that means 1,142 cycles. Assuming one day usage, that means we have used the batteries every day for 3 years. If that is not the case, I would recommend one last test, take a cell out from another battery or a spare cell and test it out again. I suspect, cell 12 does not hold charge as well and has a reduced capacity, possibly. Could be wrong and biased against cell 12 😁
It's a great calculator! In my case, the settings weren't being saved at all, I had to restart and change the values, it's certainly a bug. My batteries have only cycled a few times, they've been sitting idle most of the year on my trip home.
@@TheDigitalMermaid I would definitely recommend changing our cell 12 and testing again. My battery set has been in use for a year and a half. It came down from 300ah to 290ah. It gets used a 80%~100% practically everyday.
Magda, from my experience, cell 12 has a different internal impedance and thus it has a difference discharge/charge profile. It doesn't mean that it is bad, it is just different from the rest of the cells. If it continues to show this, replace it with a cell which has similar characteristics to the cells you are using. This is typically using AC impedance or Capacity to determine the cell is similar to others within a group. Use this anytime you are placing cells in series or in PARALLEL.
I plan to monitor it for at least a few more charge/discharge cycles, ya. If it goes low again, I'll swap it. I'll probably swap it when I move onto the boat anyway.
As it's 16s and will be paralleled at the pack level it's the resistance across the entire pack that will determine the relative discharge rates. Per cell mOhm is only critical when it's cell level parallel (assuming the difference isn't massive lol)
@@robmc3338 If that is what you believe, be my guest, but individual cells that differ will discharge and charge at different rate given internal impedance variation parallel or series.
Tolerances aren't that big if the cells are from the same bin. Using big capacity single cells is a very bad practice if the cells are not top quality. If you use b grade cells or just cheap cells you have to balance them all the time. Charging them with 30 Amps and balance them with 1 amp will take forever to put them to balance.
That'll work only if the cells are of the same basic capacity and internal impedance. The fallacy that most believe is that once a pack is balanced at a specific state of charge, no matter the method to get it there, it will remain balanced. After a few cycles, those cells within the pack that were different will either reach the lid or the lower limit sooner and if they are not monitored will exceed their specs. This can result in pack fires. So it isn't just selecting cells from a common lot or bin, there's a bit of required testing to safely build a good pack. @@bobsnabby2298
Another thing to consider is temperature. If you’re keeping the room cool and the battery is on the floor you can see an easy 15% loss versus a summer day. Don’t sweat it, they are probably fine, just don’t make a habit of it. What was on to draw it down, just the balance circuit or display on?
99% of damaged batteries are caused by parasitic loads, like a BMS being at all connected during storage what you could do it buy one of those ancient LPT switcher devices, use the switched pins to disconnect BMS lines, and then a double breaker on the main current wires if done right you manually isolate the entire pack after a certian point in the year, it should prevent self-discharge for MUCH longer, leading to more capacity available for the next year what could be MORE foolproof is to just have a TINY solar array on the boat roof, basically enough to provide slightly more power than all parasitic loads in the dimmest parts of winter either way leaving a BMS connected AT ALL during a storage period will lead to some charge loss, eventually leading to capacity loss via voltage extremes or cell damage in the case of LiFePO4 cells you want to charge to EXACTLY 3.0v, disconnect all cells from the BMS, then let it sit in storage a funny thing I learned recently is that you CAN store lead-acid, you just have to remove ALL fluids before storage, and storing the acid mixture in sealed jars is actually safe this means if you have the cash to buy a battery you should charge it up, drain out the fluid, rinse the plates with distilled water, invert the batteries, and then store for months/years
Hello. The ambient temperature is 21C , was it not summer the last time you tested capacity? Maybe this reduces the efficiency? More than 0.1 V difference is not balanced for me, it’s ok during normal use, but for testing, it would be fair to balance to 0,000 V. When the battery was left months without charging, it needs a few cycles to "wake up" Your server is nice, but it is not exactly the same as a constant load. In conclusion, the result seems ok, taking in account all the imprecisions. Cheers.
The room was about 18c when I ran these tests... I'm curious now how much temp would actually effect it, but most comments indicate that ~4% loss is expected given their age, so I think it's fine now.
I haven't seen it in the comments, so, just incase it's not there, did you actuality do a separate load test of cell 12? I'm curious to know how that came out, including its resistance compared to the others. Love your work. Thank you from NZ
I didn't test cell 12 on it's own, I tested by watching to see if cell 12 was going to be the cell to trigger low voltage disconnect, and surprisingly it wasn't
@@TheDigitalMermaid I do believe that would be the simplest conclusion to that cell number 12 low V issue. Testing it on its own there is nothing else to "interfere" with the result. It could also be mucking up the results of other cells and the balancing with its mismatched cell resistance. And just so quick and easy to do, even better that they'll still cover it under warranty! So nothing to lose. I've done a bit with 18650 cells and things like this come up a bit, although chemistry different, the similarities are still there.
I bought my lf280k cells may last year, and a recent capacity test with the victron bmv712 yielded a result of only 255ah. I messaged amy at qso and she said that its not enough of a drop to be a warranty claim. I responded that its actually considerable being that theres only 40 cycles on this pack between 3.0 -3.45 per cell in normal operation (2.5 - 3.65 for the test cycle) and qso didnt even honour me with a response. Im being ignored. Whats more is its only one cell that is letting a whole pack down, and they refuse to do anything about it.
Speaking of noisy network gear - if you are going to run Ethernet on the boat, we have had real good experience equipping boats with passively cooled gigabit switches from Mikrotik. They are dirt cheap, very low power, and very compact (far smaller than the rack mount box they are sold in).
@@TheDigitalMermaid You already have a bunch of network stuff to manage batteries. You will likely have more for RADAR and internet access. It's a RU-vid boat after all. Unless you do not want a 5G modem on the mast...
@@alexpyattaev Power consumption / efficiency is going to be a big focus on how I design the electronics on the boat, so I don't expect I'll be using much ethernet when I can avoid it, as it's a bit lossy. I'll have some, for cameras for example, but I'll be keeping it to a minimum and trying to use lower power protocols like NME2k, canbus or serial when I can get away with it.
@@TheDigitalMermaid Compared to the consumption of the motors, water heaters, refrigeration and air conditioning, literally everything else would be a rounding error, or am I wrong? RADAR and lights could consume meaningful amounts. But Ethernet? Would be fun to go through a breakdown of these things for your boat. Also PoE is amazing if you need both data and power at low weight for things like cameras (forward looking cam on the mast does wonders for log detection).
It is not only the drop in capacity but also the efficiency of your inverter. There is a loss of about 10% when inverting and I doubt your inverter is counting this in. Maybe a shunt right after the battery would give you a better data.
I don't think a 5% loss is all that bad. Unless each cell was individually capacity tested out of the box, I wouldn't assume the "280" printed on them is correct. (IME with pallets of these things, it rarely is, but they'll be close. The CALB 180's tested to 180, but they'd been at storage voltage for 3 years.) I also never assume I'll be able to get absolutely 100% of the power out of any battery - mostly because I don't want a 14hr recharge. If my house were run off a string of them, I might be willing to claw every watt out of them, but if solar was my only means of charging, maybe I'd keep that margin. In an EV where I don't want to sit around for 3-5hrs... (I've driven that EV to the point the motor controller cut out. The pack could've gone lower, but the controller can't be programmed to a lower voltage.)
Ya, I'm coming around to the idea that this is expected aging, I'm not going to worry about it much more. I'll do a few cycles and probably make that into a video, but that'll likely be the end of the capacity testing content for a while.
Flickering lights are usually associated with the Quattro (or other hybrid inverter) frequency-shifting the AC output. It will do this when the pack is near full as part of the UL1741 or UL1741SA standard to tell any AC coupled microinverters to reduce their output. Passive balancers can take a very long time to correct an unbalanced pack. Once the pack is balanced, the passive balancer can usually keep it in balance without too much trouble. Also, remember that the voltage deviation between the cells only really matters at the very top, near the charge target voltage, and nowhere else on the charge or discharge curve. Both Bulk and Absorption should be set to the same target voltage (some voltage above 3.45V/cell such that the BMS's balancers are active). 3.55V/cell is a very typical setting because the BMS will disconnect at 3.65V/cell and under normal operation you want the inverter to disconnect the loads and not the BMS. For 16s at 3.55V/cell would be 56.8V. Generally speaking, absorption time is set to 1-2 hours and the current tail is disabled in order to give the BMS time to balance the cells. The FLOAT voltage should always be between 3.35V/cell and 3.375V/cell. For 16s that would be 53.6V to 54.0V. Use 53.6V if expected loads are low, and 54.0V if expected loads are high. Once cells are fully charged, the float will keep the SOC at a high level (above 95%) while simultaneously providing load-support. In anycase,the battery pack appears to be fine. I don't think you damaged that cell. -Matt
The odd thing is that the lights are upstream of the Quattro, so if the noise needs to be going back into the house grid. It's always only those lights, not the ones upstairs, which is why I am thinking (hoping) it's the drivers in the LED lights. Agreed, I think the cell is OK.
@@TheDigitalMermaid There are several possibilities, including the LEDs themselves. Many older LED lights just use the 60hz AC directly without rectifying it and are extremely sensitive to line noise, frequency drops, and dropouts. * Older LED lights don't rectify the AC and are very sensitive to the waveform. * LED lights that are on dimmer circuits often exhibit flickering problems. Specifically LED lights not designed for dimmers. But even some that are designed for dimmers have problems. This is because LEDs are extremely efficient and dimmer circuits don't actually reduce the power sufficiently to "dim" them. So LED lights actually have to have a circuit to detect the dimming operation. * Large changes in loads such as when turning on or off equipment, including when turning on or off an inverter, can temporarily pull-down or push-up the AC voltage, resulting in a reaction from sensitive appliances and lighting. This is sometimes due to an inverter not being beefy enough, and other times may be due to trunk wiring being too thin (too large a voltage drop from the surge due to the wiring). * If there is any AC coupled solar present, the AC voltage can fluctuate when the inverter/charger is unable to absorb changes in the load for a long enough period for the UL1741[SA] servo to react and rebalance generation. These are the most common causes. -Matt
@@junkerzn7312 The LEDs in the basement were installed almost as soon as I moved in, so they're about 7 years old. I should just replace them and call it a day.
A long-shot thought? One thing I've noticed with Pure Sine Wave inverters is that driving an inductive load (things with poor power factors) will push their "inductiveness" back upstream to the DC feed. I have a small solar system that feeds my ham shack, with a Victron 500VA inverter supplying power to the network equipment and a small computer (both with switching supplies). The hum caused by that inverter made the audio parts of the system unusable. It went away if the load was purely resistive (a bank of incandescent lamps). I eventually solved it by putting a big filter (43,000uf cap, big iron choke) on the power leads to the inverter. This issue didn't occur with a "modified sine wave" inverter, though the RF interference it generated was beyond severe so I swapped it out for the Victron. That cured the RF problem, and introduced the hum. In this case, if the Quattro is acting in power assist mode (with power coming from both the inverter / battery and mains), perhaps the same sort of push-back is occurring into the mains? I don't know how the Quattro's inverter is integrated with the mains and battery systems. But anyway, a quick test might be to have the Quattro run a space heater or a couple of toasters instead of the server rack and see if your LED lights react the same way. That won't fix the LED lights, but might at least shed some light (er, sorry) on the cause.
@@tubeuser2350 Yes, that's definitely a possibility (particularly if the LEDs lighting is an older AC-coupled design). Though it should be noted that modern power supplies for computers and servers have PFC circuits in them, so their power factor is typically really good. (This is a consequence of computers demanding ever-higher amounts of power, which in turn means the power supplies have to be far more efficient). Removing hum. Hmmm. hmmmm. hummmmm.... 🙂Well, you are a ham so you probably know this, but you can try adding a capacitive filter on the DC side and some ferrite beads on both sides. The inverter type may make a difference too. High frequency inverters are noisier (~30khz range roughly) but do offer more isolation between the DC and the AC side. Low frequency inverters are quieter but reactive power and noise on the AC side will feed right back through to the DC side with few impedements. -Matt
How sensitive is the capacity test to the C-rate of discharge? The original test (netting the full 280AH) was done with 1/6 as much load per string, or thereabouts, right? If so, I would expect the higher C-rate now should result in a smaller measured capacity. That said, I believe 4% isn't out of line for the normal ageing of cells. There is, as you thought, both a time and cycles aspect to the ageing of the cells. The drop should level off and be minimal for many years, before hitting the EOL stage.
not very sensitive (I think the Peukert curve is 1.05). In any case, my test is drawing a quite low C-rate. I have had many comments say that 4% is not out of line with chemical aging, so I think it's all OK.
Aye, but I missed the actual shut down, it could have been lower and then rebounded. I need to re-run the capacity tests and be more careful with the settings to be certain this ampacity I'm seeing is the real capacity available. I've had a couple folks now say that this amount of capacity loss is in-line with the natural ageing of the cells, so we'll see. Happy New Year!
I don't know why your charger is charging at 28 amps when the battery is at 58.8 volts. imo, the charge current should decrease when the battery reaches 58.8 volts so the battery can absorb every last drop of charge power.
LFP has a crazy low internal resistance, and will pull large amperage until very close to max charge, where it finally falls off. In this case though, the BMS disconnected because of high cell voltage, so the slow down wasn't seen.
@@TheDigitalMermaid Yeah, I get that but check out @19:05 The voltage is reporting 58.8 volts and your charger was set at 58.4 volts. so it seems like a problem with the charger because a charger set to 58.4 volts should never push 20 amps at 58.8 volts.
@@uhjyuff2095Note that the voltage set is the battery voltage, is normal in my experience for chargers to go over that during charging. I've seen the x8 do the same thing also. They watch the pack voltage and stop when it hits the programmed voltage.
@@TheDigitalMermaid I suspect the JKbms has an above average idle energy consumption. I don't use my battery often and it sits idle like yours did. I want to try turning off the bms to save the battery from draining.
Mine experience showed loss of about 6-7% (135Ah to 127-128Ah) in 3.5 years of use. Lots of cycles, and not really paying close attention to what really happens there (actually, first half year was even without BMS). It was pretty hard real life test for users with null experience :) Just disassembled this system to make full cell test and possible to upgrade. System was not huge, 6x300W solar panels, 24v LiFePO 8x135Ah cells and 2kW prety-old-dumb-china-all-in-one-incevret-charger(but all of course we know that 2 is only peak..). Average load was about 150w/h, and max load about 1kW/h and in summer days it was constant for about 10h/day.
I would want to have seen the full discharge test, not just the reading from the shunt after the fact. If you've lost 600wh from a starting point of 14,300wh, that's a 4.1% decline. The fact that you got the same results with dramatically different peak charge voltages, and low-voltage cutoff settings, argues that you got a false reading on the 2nd test. However, even if you are down 4%, from two year old batteries that you haven't been cycling, I wouldn't lose any sleep over that. Presuming a linear decline, that means after 10 years those batteries would only be putting out 80% of their original capacity. Were you planning on more than that?
If you had all 6 banks connected when you got 14.3kWh at 2000w then with only one bank connected you should set the current to 2000/6 to get the same result. At 1000w you are drawing 3 times the current from each cell.
Aye, this is true but given how low the IR is in these cells, that difference is still such a low C-rate that I don't think it would materially change the results. A combination of it being colder now and the calendar ageing of the cells are the most likely reason for the lower results.
@@TheDigitalMermaid I have carried out a great deal of such tests and it is not just the IR directly that you need to consider. Remember that you will have MUCH higher current in the connecting leads and quite a few connections each of which has a small resistance and when you combine that with the additional voltage droop of the batteries because you are drawing higher current then the actual current draw by the inverter is again higher (power = VxI) so even small voltage drops make a significant difference. You are only talking about a few % drop in capacity and that is easily explained by a few 10's of mV droop across the pack (as seen from the input terminals of the inverter). It may be worth measuring the terminal voltage directly at the inverter input terminals at 300W and 1000W to see if the difference can account for the apparent capacity change. Also if you have a current clamp then see what the change in current is and if the drop is linear in proportion to the power out or if the inverter is actually drawing proportionally more power due to droop.
@@TheDigitalMermaid Not sure if you will find this helpful at all but I just did a very quick calculation and estimated your circuit total impedance at 100 milli ohm (including 16 series cell IR although it is probably much higher than this). The compensated power at the inverter input for 300A is 20.2A (this assumes a linear increase to 20A for a 1000W load). However if you include the impedance then the result for 1000W is 21A so the actual draw is 1A higher for the entire 12 hour test or approximately 4%. I suspect that the total impedance with 16 series cells will be more then 100m ohm but this may explain part of the loss.
Hrm, interesting... fwiw, the labelled IR is, I think, 0.18microOhm/cell. The 4awg wires are a bit long but I could measure their resistance if/when I pull them apart, but I suspect the usual voltage drop calculations would be accurate enough, I think the battery to bus bar cable length is about a meter, and probably about the same to the quattro. I suppose if I really wanted accurate numbers, I would need to go directly from the battery -> quattro -> load, or ideally, battery -> test load. In any case, what I am most concerned about is if the cell and the packs are healthy, and I think now I can say they are within nominal for their age.
Its such a long slow proces testing batterys. But finally it gives you some better understanding of yor setup. So many parameters always one little number or setting.... haha. Relax!
It becomes way to easy to get lost in the testing and forget the overall objective. Like the guy who couldn't see the forest for the trees. Back away a bit and use the things. You begin second guessing when you move to close.
I only use 13kw from each of my packs max. in winter I save 45% about in reserve for emergency. Have to make sure furnace can run if grid down. So almost 2 days on my batteries.
I'll be doing something similar, setting the "0%" on the helm to be at such a point that I have some reserve capacity, just in case I need it in an emergency.
you have "start balancing" set to 3.3V there is no point balancing there, and in fact the net effect is probably unbalancing the cells try "start balancing" set to 3.45V
Voltages will down always when disconnect charger/charging. Power will go up after quiting a heavy load. These change is voltages can be significant when loads / charges are high before disconnect. So always count on drop of voltage when charge completed. Always expext higher voltage after quiting a heavy load. Both drop of voltage and rebounce of voltage are natural electrical fenomenon, BUT they are NOT to be played with. Do NOT over charge batteries, Do NOT over discharge batteries either. Batteries like to cycle around there nominal voltages. Full batteries and empty batteries are a weist of money. You should have enough batteries for the task. BIGGER loads make big voltage drop.... but also give a big voltage up jump when disconnecting the load. Small batteries can't have bigger load, because of the voltage drop...
In uses like a boat, adding more batteries hits a limit, space and weight is limited so that's why I do these tests to figure out how to get the most out of them.
12:27 Looks like you have a calculator on your phone. I wonder what the app is, maybe a TI? I run the Droid-48 which is an app as the HP-48. FYI We live on a sailboat.
Try setting the charging absorbtion voltage at just before the bms high voltage disconnect, and give it an hour, and reset higher if possible untill its absorbed at the 3.65 or close too for a good bit. I am fairly new to lifepo4, and have built myself a 600+ah 12 volt bat(firs cycle tested 630ah on the smart shunt) its down to 615 after a few hunred charge cycles. I find that after say 5 days without getting over 70% charge lets say, it takes a bit longer to absorb once good solar conditions arrive vs a daily fill.
Aye, I was doing that with the Phoenix Charger on the boat for the 12v side, it worked a treat. I need a constant voltage PSU to do that for this, properly.
i think your battery pack has a normal behavior, simply is not yet well balanced, the jk bms is a very good bms, activate the balancing over 3.45 and charge the battery at 55.6 /55.7 globally, after 7-8 day your pack will be perfectly balanced, it's more easy and fast to balance the cells on top because on top you can really see the difference between the cells. i solved all my balance problems in this way (sorry for my english)
@@TheDigitalMermaid set inverter-charger to 57.6 bulk, 55.2 float 1-2 hours absorbtion. 58.4 for bulk is way too high as correctly configured BMS (OVP at 3.65, OVP release anywhere at 3.55-3.6) will cut out instead of allowing full charge.
For propulsion you have to really take care of cell deviations and also go to 3,65V to balance as good as possible and also have proper sized active blanacers. Manufacturer of EVs only using metricously matched cells means their internal resistance is as idendical as possible not to get big cell deviations. 56mV is unacceptable and 27mV deviations for propulsion is really bad, for storage thats good. So you need charge to 3,65V and get it as perfectly top balanced as possible. Then run some time like 1 month in the boat and see what deviations you get, if they increase you need a more powerfull cell balancer...neey 4A should do it.
Keep in mind, the loads on the batteries of a sailboat are a LOT lower than the loads in an electric car. I'll have 6 packs, possibly 7 in parallel, but I'll use 6 for this math. The two electric motors, at absolute maximum power, would pull about 21kw, or 410 amps. It almost certain I would never pull this much, but for the sake of the math, that's 68A per pack, or 0.24C. A discharge rate of 0.2C is where you get the highest cycle life. So your point about getting up to 3.65v per cell for better balancing is valid, the idea that this is a high discharge environment isn't. In reality, I suspect I'd be pulling about 5kw on the high side of normal, which is ~100A, or 20A per pack, or 0.07C.
I have found that with good cells, a passive balancer (it isn't actually passive) is very capable of keeping cells in line. But that is just 3.5 years experience with 48 cells in total. I have an active balancer, but only one, and it isn't connected. The biggest consideration is how well you top balance. Do it right, and you don't need to do it again for more than a year.
@@john_in_phoenix I suspect that, before I move the cells onboard, I'll do one more big 3.2v parallel pack and top balance at low current, make sure they're all as balanced as possible.
Not all RVs are Teslas pulling insane C rates from the cells. The ones my buddy build are little city hoppers and commercial EV pulling similar 0.3C from battery pack. But usage pattern is significantly different means constant current draw over long time while offgrid short peak times when you cook or AC running. My buddy is building EVs and powerwalls as NPO so you can believe me that the info i have from him is valid and based on a lot reasearch, doing that business for 20years and he was actually the first one getting EVE and Lishen cells and Yinglong LTO into europe. The cell deviations you have shows they are not matched by factory and most likely grade B cells which the vendor matched in best case...good for storage but suboptimal for propulsion. A very good source is marinehowto by rod collins...sadly he slowly recovers from a stroke ... Best way to top balance is first to 3,4V and let them sit for 6h, then 3,5V, then 3,65V to achieve a proper top balance. A passive balancer is good for offgrid storage but insufficient for propulsion or EV especially if you don’t have metricously matched cells from factory. The balancer have to fix the issue of cells differ and with lifepo4 only have the top knee from 3,4 to 3,65V to do that. Thats the issue all EV manufacturers battle with lifepo4 and their solution is metricously matched cells and sufficient sized active balancer. I have factory matched cells from my buddy and after 3 years still 3-5mV deviation and pulling 1C load they are between 5-8mV deviations, that on my cat as 1088AH hybrid house which also acts as starter for engines. Another good test is when you propulsion is running for 10min start to measure every cell with an IR temp gun especiallythe terminals, if they are all even within 2-degrees celcius thats ok put i bet you will find some 5 degrees above or below, especially the first and last cell of each pack. That tells you some cells do more work and some less and the balancer needs to make this even...
4% loss isnt too bad for 2 years of use, but i would have expected such low cycles to be less lossy - its similar to my loss after 2 years of use (2019 bought cells used from early 2020)
That is calendar aging for you. They will lose capacity just from getting older. One thing that can happen if they have been just aging without being cycled is that you can sometimes regain a bit of capacity back once you start cycling them after they have been sitting a long time.
I don't really have one yet, to be honest... Could be it wasn't as charged as the others, and after the BMS powered off, the internal resistance dragged it down rest of the way? I'm not really sure though...
That is so much trouble, I don't have the nerves for all that. The 8 cell 2p Japanese Cells i have doesn't need a BMS/Balancer. I do have batteries with built in BMS and they work fine. Those Chinese cells seem to have a multitude of problems staying Balanced
@@TheDigitalMermaidAccording to the published aging curves for LiFePO4, this is completely normal. There is a fairly linear drop in capacity for the first few tens of cycles, and it slows down quite a bit after that. So if you get 5% loss over the first 50-100 cycles, you should see more like 2% loss for the next 100 cycles, then 1% for the next 100. After 500 cycles, they seem to level out mostly, and that’s why they manage to keep above 80% for 3000 to 4000 total cycles. This aging can be reduced somewhat by tightening the depth of discharge of course, because most of the aging phenomena occur at the extremes of the voltage curve. Using 3.55 or 3.6V instead of 3.65V charge cutoff seems to have the most effect - with LiFePO4, increasing the lower discharge cutoff voltage is less important. Interestingly, some people have found that even with a 3.55V charging cutoff, the capacity reduction is not a lot - especially if you increase the absorption time to compensate.
The same thing recently happened to me. I left a cell pack on, and two cells in the pack were damaged. I tried to rescue them, and they did charge beck up, but they quickly self discharged. I had to replace two cells. ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-MVBnR2zvuXw.html - Oh, my cells are in a robot.
I'm just bit tired with this persistent changes of parameters setup. Higher/lower current/power. Changes off screen. Mess. No clear, readable conclusion - just flow of thousands words and long tens of minutes. Don't want just complain or touch you - but pointing why me (and maybe other viewers) goes away at some stage. Sorry.
I know my channel and my content is not for everyone. My goal was to show the process of learning, and so that's what I've been doing. They're not instruction videos, there's already many fantastic channels for that.
Stop listening to ANDY! Everything you are doing is because you don't really know what you are doing. Admit that? YES!! B grade cells won't last like Auto Grade. Andy is always making bad assumptions. Jus swap out the cell. Stop listening to a lonely old man in a garage in Australia who is wrong about most of what he says
To suggest that Andy doesn't know what he's doing is laughable. You want me to admit I don't know what I'm doing? Easy, that's the whole premise of this channel. Normally this is where i'd ask you to backup your claims, but to be honest, the opinion of someone who would dismiss someone else because they're "an old lonely man" is not someone whose opinion I'd value in any case.