Patrick OReilly

You're doing an awesome job on this forum Jason (and whoever might be assisting you). Keep up the excellent work!

(I've struggled with the user manual). Hmmm - yes, I've also found the manuals to be very verbose, but not very technically enlightening. But I'm an amateur, so perhaps I'm missing stuff that's obvious to those "in the know". This Forum has been my best source for helping to understand stuff a bit.

Caution is the better part of valor!!!! Do be careful. 😀 It's all been said above, but I'll try to summarize briefly. You could do this using KWh or Ah as the base for calculations. I prefer to use the KWh because that takes care of the fact that your DC and AC are running at different voltages. Batteries have a Capacity - usually expressed in KWh. e.g. A 5KWh battery can provide 5KW for one hour. or 1KW for 5 hours. In theory. That answers the question: HOW MUCH power is stored? Batteries also have a "C" rating - usually expressed as a simple number with "C". e.g. 1C, or 0.5C. This basically tells you how much of the battery's total Capacity you may use in one hour. You could convert it to a percentage - That makes more sense in my head. 0.5C = 50% of Capacity per hour. That answers the question: HOW FAST can the power be delivered? My battery is 4.8 KWh, 0.5C. This means that I should only pull 50% x 4.8 = 2.4KW max load at any time. If I do that I could (theoretically) pull that load for 2 hours. (Except - don't do the full 2 hours because you don't want to go below about 20% SoC.) So - check your battery spec - if you have 5KWh 1C (mentioned by one of the responders above) - you should be able to pull 5KW from it, but it will only last an hour (less 20%). So please check your battery's specs. And, if you have two batteries, then you can ADD them together. Effectively you have a 10KWh Battery, still with 1C rating... And then set your Inverter to limit the current accordingly - that will protect your batteries from accidental abuse and damage. (PS, though may battery is rated 0.5C, I set my Inverter to max at 45%, or 0.45 C.)

Hi. Thanks @P1000 and @Bobster. I appreciate your time in responding. Yes, I am aware that my measurement is not in laboratory conditions. One would have to start and end with equal SoC, and implement a whole bunch of other controls. Hence why I entitled this topic "How efficient are LiFePO4 batteries in Real Life". What I want to check is whether the "real world" efficiency I'm getting is OK, because there's a big gap between 80% and "high 90s". So, my numbers were simply over a period of a month, as reported by the inverter. I think a month is long enough to establish a reasonable average, or trend. There may have been a difference in SoC at midnight 31 October vs 30 November. But at the very most that would affect the results by about 2-3%, since I keep my battery >= 40% at almost all times. Inverter losses, heat losses, etc. all play into this. But it is something to think about when 20% of the power I try to store disappears. And once one knows what that is going to be, it will have some impact on your strategy for charging your batteries. @I84RiS said that he "did this calculation" (by which I understand he used my "real world" method), and he found 92% efficiency. I don't know what his setup is so I can't draw any lessons yet. But I think if I'd achieved a number like that I would not have been concerned. Anyway - so this is really what I was looking for - a number of people to look at their system in "real world" conditions and see whether my result is good, bad, or ugly... Do you guys, or anyone else, have your November stats at hand to do this quick calculation? (See the OP for the very simple method I used) Let's see what the results of the majority look like.

Thanks @iiznh From the behavior of the inverter, I don't think there is any step-down as it nears full. Assuming good sunshine, I usually see the SoC get to 100%, and PV production and battery charge (negative power to the battery) continues apace for a while. It can be anything from 5 to 30 minutes. Then Battery voltage quite suddenly increased from about 51V to 55V, and at that time the Inverter stops pumping power into the battery. So, it looks like the inverter does not really use the "SoC = 100%" as the trigger to stop charging - there's something else that triggers that, and it's more closely related to the battery voltage than to the BMS reported SoC. Regarding claiming the battery BMS is faulty - I have no idea whether what I have is enough evidence to demonstrate that as a fact... I'll see whether I can get any response from that avenue.

Warning : I'm just a novice amateur owner, so take my opinions from when they come... So, I'm using SunSynk. which means some of the config (software) looks different, but I understand that the hardware and capabilities are much the same as the Deye. Note: I'm not familiar with the bottom-right graph you have - I've not seen one like that on the SunSynk - so it's not clear to me what it means exactly. But it does seem odd that it shows negative values. It might be telling you that you've already done what I will suggest below... For starters, I will assume you have the CT installed where it should be... And thus, you'll have one DB (non-essential load) on the Grid side, between the CT and the Inverter, with things like geyser and stove/oven, and another DB (essential load) on the Load side with all your household load (lights, plugs, etc.) If you set "Limit to Load", then you'll only use your inverter power (battery and/or PV) for the "Load", which is also described as essential load. This will NOT provide any power to your non-essential load items. If that's what you want - then you're all good. However, since you clearly have more production than you are using (on sunny days), you might want to allow the Inverter to send some power to the non-essential load. To do this you must set "Limit to load" OFF, but you must set "Zero Export" ON. (I don't know what "Zero Export" might be called on the Deye Inverter.) When I run the SunSynk this way, it then reports the non-essential load as part of the load on the graphs. It also means that I get to use some of the free energy (PV) to supply my geyser. Suddenly a timer switch on the geyser starts to make sense... Let the geyser come on during the day, once your batteries are full and you still have good sunlight - heat the geyser and still recharge the batteries afterward.

Interesting. I'd feel quite happy with that efficiency I think. Could you share which batteries/Inverter you're running? And perhaps any particular setting you use that you think may be pertinent...?

Hi all. How Efficient Are LiFePO4 Batteries in actual practice? A quick google gave me answers of "over 90%", "90-100%", and then a very bold "100%". In my (lay-man's) opinion, the round-trip efficiency of a battery is simply: What percentage of the power I put IN when charging the battery will I get back OUT when I use it? However, I'm looking at the following figures from my Inverter after my first full month of operations: This is a single 4.8 KWh Dyness battery. (Sorry - I don't know why the colors are so faded in the screengrab.) So. over a 30-day period, I see 137.3 KWh of charge, and only 110.9 KWh of discharge. By my math, that is just over 80% efficiency. Now, If I'm getting free power from the sun, then the 20% inefficiency is not too serious because 20% of zero cost is still zero. But this makes me think twice about using the Grid to charge a battery. Instead of paying R3.00 per KWh (round numbers), I need to buy 1.25 KWh to charge the battery, to be able to consume only 1.0KWh. So now my grid power is costing me R3.75 per KWh. So, my strategy is leaning toward ONLY using solar to charge the battery, and then using it as much as possible. But using grid to charge the battery makes no (financial) sense at all. Obviously, we also contend with load shedding, and that may be a good reason to charge a battery from the grid. Grid+25% is still cheaper than running a generator. But for the purposes of this thread, I just want to get clarity on the practical experience of others, and possibly the theory (if there are experts here who know the science) as to what is a realistic expectation of the round-trip efficiency of a LiFePO4 battery.

So, I've had this problem happen to me again. The battery SOC from BMS is sitting at 40%, and then suddenly it's 0%. At the time, the lowest battery voltage reported was 45.8V. On the first occasion it was 45.1V. I'm thinking that the problem may be related to the BMS becoming poorly calibrated over time. The symptoms prior to the problem are that for a few days the batteries had not ever been charged to 100% and pushed till voltage peaks about 55V. Based on the comment from IIZNH above, this may cause the BMS calibration to become inaccurate, and thus to be feeding the inverter bad SOC info, and so SOC may be worse than reported. Does this make sense? And this leads to a question: Is there a recommended proper procedure for ensuring you keep your battery healthy and the BMS working reliably? Should I be using grid to ensure my battery is properly topped up at least every 42 hours? (or 48, 72?) Anything else? My approach to my solar installation has been to maximize the capture of Solar to reduce my bills, while also maintaining a minimum SOC of 40% so that I can survive a Loadshedding interruption of up to 3 hours. So, I run the system with ZERO export, battery priority, but I leave the SOC minimum level at 40% all day and night. On good sunny days I easily top up the battery before noon, sometimes before 10. But with recent weather I sometimes only get to somewhere between 80-100%. I figured that's fine, I'll use whatever I have and just save the last 40% for loadshedding. But perhaps this strategy is not battery-friendly... ? I also have another question regarding battery efficiency (power IN vs. power OUT) - but I'll put that on a different thread.

Yes, it is correct - that was a setting I noticed; it was correct all along. Thanks.

Thanks again @mzezman and @wolfandy. Yes - I have only one battery, so max ~2.2 KW feed there. New system - started conservatively. Anyway, I will make work of re-stringing my panels and see what impact that has on the picture.

Another Question - which I am now thinking might be related to the above.... I have noticed quite often that in clear sunny conditions the Solar power seems to be very slow to respond to a load spike, which results in the inverter using grid power to meet the demand, instead of the solar that should be freely available. But this seems to only be an issue when the panels have been "throttled back" by the inverter, after battery charging is 100%. If Solar is busy feeding, and there's a spike, the solar continues happily. Consider this graph from two days ago: So, there was a load spike about 9:40. At that time Solar was still in full flight finishing off the battery charge, so it continued to feed, and battery joined the party, so there was very little power taken off the grid. This looks OK to me. Then there was another load spike about 12:15. At that time Solar had been "throttled" by the inverter for a couple of hours (forgive my non-technical language). During the spike the battery responded immediately, but the rest of the demand was met by GRID power, while Solar did nothing. Then only about 10 minutes later the solar eventually responded to recharge the battery. I see this sort of behavior quite regularly. So, my Questions: Is this an Inverter issue - being slow to respond to demand with Pv? Is this a Panel issue - being slow to "wake up" when the Invertr calls for power. (New option after discussion above) Is this perhaps an installation issue causing start-up delays because my Strings are only on 130V per MPPT, whereas the Inverter says the minimum Pv voltage should be 150V.

Thanks @freierMensch Very interesting, and also a good cost saving. Maybe one day I'll try something similar...

Thanks for sharing your experience. I am getting a bit of a gut feel that it may have been a result of cell imbalance. For several days prior to this the battery had not been fully charged. (Why? because we had so much rainy weather, and I was not doing any charge-from-grid). I do see that the battery voltage was lower in the last day or two than it has been at any other time. So, over a period of maybe 48 hours of mixed charge and discharge cycles, without a solid "fill it up to the brim", I guess there was a lot of opportunity for the cells to become imbalanced. I'll keep an eye on things. If it should recur, I'll keep all the feedback in this topic in mind.

Awesome! Thanks to all of you for your feedback. I'll have a chat with my installer to arrange some changes...

Thanks for your speedy responses, guys! A couple of questions, if I may: Would I then see 260V on the one string, instead of 130v per string? Does this put me anywhere near the 13A limit per MPPT that I've also seen mentioned? I have seen peaks of between 10-12 A concurrently on each string in the current configuration! I'm hoping (assuming) a string is in series, so voltage will be higher, but current will be the same. 3270W (nominal) / +-260V is pretty close to 13A. Is there any potential disadvantage of doing it this way?

Hi all. I did not want to hijack the other similar thread - my setup is a bit different... When mine was installed, I was oblivious of - well - a LOT. So, panels per string etc., was not even a thought. My installer just did stuff and that was that. I have 6 x 545W JA solar panels. 2 strings of three each. from other discussions I'm learning that may be a problem due to difficulty "starting up". My inverter stats report that each string is delivering about 130V whenever there's enough sunshine. On another thread it seemed that 150V was only just enough... ? I'm lost. It seems to work OK, but I have no clue if it would be better as one string... or... ? Any advice (and explanations) would be most welcome. Edit: all 6 panels are mounted on the same piece of roof - facing the same way.

Thank you @Nexuss ! So, as I said - I'm not a sparky, so I was never aware that there was a bond between Neutral and Earth. But I think I see the logic of why there should be. and therefore, why you'd need to take care of that when the normal neutral is out of action. Thanks for taking the time to explain that to one of us civilians. 🙂 I will be having a look at my installation to see whether this has been done by my installer.

Hi @Nexuss. I think I have a good feel now for what the role of the CT is, and as a result I'm seeing some nice support for the non-essential load from my system. Question: you mentioned the "Earth Neutral bond Relay" a couple of times. I have no clue what this means, but 100v floating around uncontrolled does worry me. Could you please explain that a little? I'm not a sparky though - so words of one syllable please... 😊 I'm learning so much on this forum!!!

Thansk @Scubadude . Since it seems to be behaving ever since, would that rule out the hardware diode issue? Or should I be checking this?

Thank you @mzezman !

Thanks for this @iiznh. I have noticed that the inverter does keep charging for some time after the SOC reports 100%, so it seems to be doing what you have said. Since this thing happened two days ago, I've had a full cycle of charge and discharge (to 40%), and charge again today with good sunlight. It seems to be well behaved now. I'm developing a suspicion that the several days of poor/incomplete charging cycles due to the horrible weather (from a PV point of view) may have allowed the BMs and actual cells to get out of sync, resulting in the Cells hitting min voltage even though the BMS reported 40%. This morning when at 40% (my configured minimum) the voltage was still reporting 48.8V, which seems much healthier. FYI: Out of an abundance of caution I have limited my battery to 45A for max rates of charge & discharge, rather than 50A. Just because I like to go easy on machinery. I don't know if this really matters, but it makes me feel a little better about life. I am a software guy myself, and all the talk about running your own HomeAssistant and writing your own scripts, etc., is sounding more and more appealing. But I have to make sure I understand the nature of the beast first, and I really appreciate what I am learning from all of you!

I believe I am yes - I did not install this myself, so I'll double check the installation when I am at the site again.

Hi @GreenFields. the inverter is SunSynk - so I'm using their App to monitor. Yes, there is info under the Equipment, "Battery" tab. There are graphs for SOC, Voltage, Current, Temperature and Capacity. They all show history, except the Temperature which seems to be non-functioning - reports -100 C all the way. From this is see the voltage was at a low point of 45.1V at the time this happened. This is also the lowest voltage I've seen in the 2 weeks. The Battery Setting tab has a field for "Battery empty V" which has a value of 45. this was set by my installer. I'm not sure what the Inverter is supposed to do if it hits 45V, but this behavior does not seem to make any sense. If anything, I'd expect the inverter to "switch off" the battery till it gets some charge. OR something like that. Strangely - I had let the battery run down to 30% SOC the day before, and at that time it reported 46.3V, which is 1.2v higher. This is not making sense to me. But I am a noob to all this.

Thanks @iiznh , you clearly understand this a lot better than I do !!! Is there a way to "calibrate" the battery? I'm so new to all of this. The battery is Dyness 4.8KWh. this is the best technical info I can find at the moment: Looks like 15S? It has a BMS type connection to the Inverter, so I would have assumed battery state should be well managed. But maybe not?