| General > General Technical Chat |
| Lithium Iron Phosphate - "Low current overcharge" |
| (1/2) > >> |
| paulca:
I have come across this phenomena recently on the internet. YouTube, Forums, etc. etc. What I have not found is a coherent explanation suitable enough to even determine if it's "old wives tales" or an actual effect. Proponents claim, if a LFP cell with a charge termination voltage of 3.650V will be damaged if it is left with 3.650V across it after it's termination taper current is reached. The trouble with most of these claims is they do not provide any evidence for this "damage" or only make "hand wavey" conjecture as to what is occuring. As an example, some of the proponents will claim the cell is fully charged and has no "free lithium", so any current at all goes into breaking down the chemistry. This sounds plausible, until you look back at their charging profile and compare it to the datasheet, or even conduct some rudimentary tests. For example, while they are typing up their "Low current overcharge" post, quickly charge the cell up to 3.8V, the capacity test it back to 3.65V and just before they hit "Post", tap them on the shoulder an point to the other few AmpHour of "free lithuim" the cell had. In my limited and humble understanding, if you hold a lithium cell at any voltage under it's Do Not Exceed voltage the current it will ultimately consume, when it finally reachs a steady state will not only be tiny, but will do nothing more than generate a tiny amount of heat in the cell (compared to the cells mass). I can conceive how that tiny amount of heat could, possibly, wear the cell. However I cannot see this being a large concern. A 100Ah cell at 3.65V will settle to a quiscent current of something less than 100mA. That's less than 1 watt of heat in a cell that weighs 2Kg. It's going to get FAR hotter dumping at 1C discharge or charging at 0.5C! So, what are your thoughts? 1. Absolutely a grave concern, often overlooked. 2. Something to consider, but it might cost you less than 1% cycle life. 3. True but blown way out of proportion. 4. Ole wives tale. 5. Actual CAUTIONs and WARNINGs taken from other cell chemistries and their unique charge curves, which DO require holding them at 4.20V for only just long enough and taking them back out of 4.20V as soon as possible, but do NOT apply to LFP at 3.650V. ? |
| Siwastaja:
Who knows? Exact chemistry details are important, you can't generalize the problem to all LFP cells from all manufacturers. Follow the datasheet conditions (usually: terminating the charge and not keep at 3.65V) unless you know very well what you are doing. As often discussed here, you can "float charge" or keep a cell connected to a current-limited voltage source that matches the open-circuit voltage of the cell, and no current would flow except that compensating the tiny self-discharge current, but 3.65V is not the open-circuit voltage for a 100% SoC LFP cell. Therefore I don't see the point of doing that. Just look up the exact OCV value of 100% (for example, by charging to 3.65V C/20 as per datasheet and let the cell sit for a day) and use that as your float voltage instead of 3.65V. From memory, it's something around 3.45V. |
| paulca:
Yes, you see I agree with that. I think it is the "generalisation" occurring, where people are parroting some small contextualised bit of information and then applying it in far broader terms than it deserves. For instance, you mentioned "open-circuit voltage". That is a term most of these people don't understand. They will give you the voltage they see on the harbour freight meter to 2 decimals, but with a load applied and then confidently state something like "Look, as I said the rest voltage for the 4S pack is exactly 13.20V. " So it's been frustrating me. I stopped and asked, what exactly am I trying to answer that is leading me through this frustrating content. I am working on and testing the top (and bottom end) responses in my own solar/lithium system. So I am dealing in some of the finer details such as balancing, absorption (or tapering depending on what you call it). Trying to get the battery with it's BMS in harmony with the MPPT such that (1) I am getting the full capacity, (2) I am not invoking the BMS disconnect and (3) I am not over stressing things unnecessarily. Obviously any voltage measured while load is applied is either invalid, or incomplete. I am not making things easy on myself, because I am conducting these tests in-situ, not "lab conditions". I stubbornly refuse to cheat and waste power or use the mains to charge the pack. So I wait on the sun to work with me. What is making that easier is persistent high resolution data across multiple measurements. So if the sun plays ball while I am not on the ball, I still get to review what happened afterwards. Not a lot of what people are claiming is adding up. I come here and people usually make sense. Which is why I came back here. On the datasheets. They are very, very hard to come by. What you usually get is a single page which has the bog standard stuff on it. Do not exceed, 4.20V. Max charge voltage: 3.65V, Max discharge 2.5V. etc. The best you can hope for is the capacity test conditions and results. That is about as close as you'll get for "charge parameters". They vary between tests, manufacturers, batteries and the various studies I have read. It just seems they are "parameters" and without defining them all the rest of your results are invalid. What I'm saying is, a datasheet might state it's capacity test as 3.65V until a current of 0.05C. I feel that is a statement of necessity and having to draw a line somewhere rather than a limit. I am rapidly approaching the point where I'm realising the slim benefits that exists anywhere in that top few % of capacity are just not worth the diligence or the chance of mistake. Coming back around on this for about the dozen'th time now I am starting to realise, some of those completely "made up" figures people throw around, the over zealous babying and artificially lowered limits, are not just safety margins of "lieing to the children", and no they are not valid and correct to the letter of any documented process... what they are however is what people have decided to target which lies somewhere between stupidly babying and probably doing more damage versus running them to within an "inch" of their limits. It was when I pondered buying a much bigger battery and looking at £2500 I started to consider if I would be so willing to do these experiements and argue/haggle and debate over the "peaks" or would I rather not set the operational limits well inside the datasheet maximums and not have as much concern, monitoring, risk of damage and just "chill" and let it be. If I did that, how close would I end up to these "made up figures" which annoy me. |
| bdunham7:
--- Quote from: paulca on April 17, 2023, 04:44:19 pm ---Not a lot of what people are claiming is adding up. I come here and people usually make sense. Which is why I came back here... ...If I did that, how close would I end up to these "made up figures" which annoy me. --- End quote --- I can give you rules of thumb off the top of my head with no attribution and you can choose to believe them or you can research them further. There's a tension between charging stress and cell capacity--since battery purchasers obsess over capacity per dollar, battery sellers want to maximise the number that they can give you. For most Li-ion cells, assuming reasonable temperatures, you can float them at 3.92V without stressing them. Some chemistries might be able to go a little higher and you can choose a higher voltage, like 4.00V, with a small tradeoff of lifetime for some increased capacity. If youfloat cells higher than that, they will start to have greatly diminished lifetimes. Floating a typical Li-ion cell at 3.92V gives you about 75% of the capacity that a charge to a 4.2V/0.1C cutoff. For LiFePO4, the magic numbers are 3.30, 3.45 and 3.65. If you float at 3.3V, you'll get ~70-80% capacity and low stress. You can bump the float to 3.45V float and eventually get almost the same capacity as a 3.65/0.1C charge, but greater cell stress. I know there's this idea that LiFePO4 can take abuse, but I'm not so sure that abusing them is a good recipe for durability. Now that's all just folklore and myths as best I remember, so take it for what it is worth. Research may give you some varying answers, specific cell chemistries and designs may yield differing results, but the general principle regarding tradeoffs is sound. |
| paulca:
In part the trade off for me is more how long, how slowly, to hold them at 3.6xV to allow for the potential of imbalance before cutting back to float. Do I cut back to 13.6V or 13.4? Do I let the MPPT trickle the pack for 3 hours? 2 hours? I suppose I'll need to just review the data I am getting and decide, how long does it normally take, how often is there a "slow cell" and tune and refine over time. At the moment, 14.40V pack limit for the MPPT seems to provide enough headroom for the BMS to carry on balancing without tripping the cell over volt at 3.65V. It hasn't been tested in full direct sun yet though, I'm need to test if the MPPT tapers fast enough when it's hits 14.40V to allow the balancer to 'tidy up'. If when things settle in, it turns out the cells are all in order within 10 minutes of hitting top end paramters, then I can set the boost duration accordingly. |
| Navigation |
| Message Index |
| Next page |