Matthew, It sounds like your pack has cells that are severely out of balance and now the charger does not push the pack voltage high enough to where the Curtis re-calibrates its SoC estimation, so it continues to go off a bad estimation. If you have a way to push a smaller (1 or 2 Amp) current into the pack with all the balancers powered (if they need 12V then also charge your Aux battery!!!) so that they can shunt the current away from the cells that are already full and you only charge the cells that are way lower, or if you have a hobby charger that can charge 1 or more LiFePO4 cells at a time, then you can apply that charger to one or more cells at a time, the hobby charger will charge and balance all connected cells and then you move it to the next set. Say if it can do 4 cells at a time and it takes overnight to charge and balance, then after 6 or 7 nights you have a balanced pack... Then the Curtis should recalibrate and the pack should work normally.
BTW, what you are seeing is also the reason that some Chinese products go up in smoke: the BMS is designed to take care of minimal differences of SoC and balance them away over time, combined with a charger to a fixed voltage. So have one or more bad cell(s) that discharge too low and the BMS can't balance it away before the charger is pushing the good cells way over their max cell voltage, resulting in smoke and fire. Or a cell is so bad that it self-discharges to zero and the charger tries to bring it back up at full bore, generating enough gassing and pressure to burst the pouch with the same effects: smoke and fire. So, stay safe, don't let your cells go over max voltage, balance the cells to the point that all of them reach full around the same time and the BMS can balance the remaining differences away at the end of the charging cycle. Success! Cor. -----Original Message----- From: EV [mailto:ev-boun...@lists.evdl.org] On Behalf Of Matthew Quitter via EV Sent: Tuesday, August 08, 2017 5:57 AM To: Electric Vehicle Discussion List Subject: Re: [EVDL] Curtis 1238-6501 State of Charge Hi Guys, Thanks for all the feedback, really interesting. I have a 10 year old battery pack, I got it second hand out of an old conversion. The balancers do a great job of keeping the pack balanced, I think they can move around 10A. Most of the time they can keep the charge in control, it's only in the last 20% where I guess the charger is doing constant voltage and I guess is pushing more current than the balancers can move. Or the batteries are reaching their limit and no amount of balancing is going to add more charge to certain cells. Basically my pack has a couple of outlier cells that charge really slowly (as in their voltage doesn't change) and a couple of cells that charge really quickly (the voltage rises up and over 4V quite quickly). I've found that looking at the cell voltages gives a fairly accurate representation of the SOC. I've logged each cell every few hours, particularly after a charge or a drive, over a week or two and everything was fine until recently. I measure down to the millivolt. Is that accurate enough? And our temperature here is so stable, particularly in the underground garage where I keep the car that I can't imagine the internal temp of the battery is really that different to ambient but I don't know much about all this. Will go and do some research. It's like the Curtis SOC has missed a chunk of Ah going into the pack. I guess I'll have to draw the pack down using a load to get the charger and the Curtis on the same wavelength again. I looked at single cell charger/discharger but the ones I saw do max 2A and the maths of 25 160Ah cells meant I was facing two months of slowly discharging cells! I've gotten a lot conflicting advice about the max charge voltage of a Thundersky cell. Bill said to not let them go above 3.8V while the Thundersky page - http://www.thunderstruck-ev.com/Manuals/Thundersky%20Product%20Manual.pd f - says 4.25V. I tend to stop the charging when cells get over 4V. Is that causing damage to the battery? I was planning on building my own charge monitoring system using an arduino to monitor individual cell voltages and if any one cell goes above say 3.8V then shut off charging for a period of time to let the balancers work. And then shut off the entire charge once the pack reached a fixed voltage - say 88V. But you guys are saying that shutting of charge at 88V would not give me an accurate reflection of a fully charged pack. I'll look at the Ah meters you linked to. Many thanks, Matthew 07966 806 727 On 8 August 2017 at 08:00, Bill Dube via EV <email@example.com> wrote: > In LFP, temperature influences voltage more than SOC in the region of > 20% to 80% SOC. The voltage changes just a tiny amount between 20% and 80% SOC. > > Basically, if the voltage changes it is more likely to be due to a > change in temperature, than a change in SOC. > > Indeed, you could measure the temperature of each cell, along with the > voltage, to tease out the SOC, but you need the average internal > temperature of the cell, rather than the external temperature of the > cell, and you need a very accurate measure of voltage. It is way > simpler and cheaper to count amps. > > > > On 8/7/2017 2:31 PM, paul dove via EV wrote: > >> Sure you can. Voltage is the only way to discern SOC. >> You just can't measure accurately under load. You have to measure >> open circuit voltage. >> www.mdpi.com/1996-1073/9/11/900/pdf >> >> >> >> From: Willie via EV <firstname.lastname@example.org> >> To: Matthew Quitter via EV <email@example.com> >> Sent: Sunday, August 6, 2017 9:13 PM >> Subject: Re: [EVDL] Curtis 1238-6501 State of Charge >> >> >> On 08/06/2017 07:38 PM, Matthew Quitter via EV wrote: >> >>> Hi Guys, >>> >>> My 1238-6501 Curtis controller gives very different SOC % for >>> similar voltages. >>> >>> For example I've recorded a pack voltage of 82.2 and had the Curtis >>> show 22%, 31%, 43%. >>> >>> The pack is made up of 25 Thundersky 160Ah batteries. >>> >> With LFP cells, you can't accurately infer SOC from voltages in the >> middle of the range. The voltage/SOC curve is too flat. >> _______________________________________________ >> UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub >> http://lists.evdl.org/listinfo.cgi/ev-evdl.org >> Read EVAngel's EV News at http://evdl.org/evln/ Please discuss EV >> drag racing at NEDRA (http://groups.yahoo.com/group >> /NEDRA) >> >> >> >> -------------- next part -------------- An HTML attachment was >> scrubbed... >> URL: <http://lists.evdl.org/private.cgi/ev-evdl.org/attachments/ >> 20170807/6b534a2b/attachment.htm> >> _______________________________________________ >> UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub >> http://lists.evdl.org/listinfo.cgi/ev-evdl.org >> Read EVAngel's EV News at http://evdl.org/evln/ Please discuss EV >> drag racing at NEDRA (http://groups.yahoo.com/group >> /NEDRA) >> >> > _______________________________________________ > UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub > http://lists.evdl.org/listinfo.cgi/ev-evdl.org > Read EVAngel's EV News at http://evdl.org/evln/ Please discuss EV drag > racing at NEDRA (http://groups.yahoo.com/group > /NEDRA) > > -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.evdl.org/private.cgi/ev-evdl.org/attachments/20170808/182d caa4/attachment.htm> _______________________________________________ UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub http://lists.evdl.org/listinfo.cgi/ev-evdl.org Read EVAngel's EV News at http://evdl.org/evln/ Please discuss EV drag racing at NEDRA (http://groups.yahoo.com/group/NEDRA) _______________________________________________ UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub http://lists.evdl.org/listinfo.cgi/ev-evdl.org Read EVAngel's EV News at http://evdl.org/evln/ Please discuss EV drag racing at NEDRA (http://groups.yahoo.com/group/NEDRA)