Why are my batteries absorbing and floating at a temperature much lower than what is set?

mevenable

how do I measure the voltage at the "device terminals"?

BB.

Placing the meter tips directly on the metal post or terminal of the device. [corrected typos from my phone. -BB]

If you have a loose binding screw or dirty connection, the cable can be at higher voltage than the device terminal.

Bill

mevenable

Estragon - an observation on that question about falling below 50 V - last night I was running a load test on my new battery bank, letting it handle a modest load for the night by itself and see how it goes.  Anyhow, I put the system on grid from 4-6 pm, to float the batteries with the last bit of sun and to not have any charging during the period I was measuring.  Right before I put the load on the batteries, they were in 51.15 V, and the Mate3 had both the temperature compensated and uncompensated voltages at 51.2 V, even though the battery T was 33C.  In other words, like you said, if the batteries are at rest the two are the same, and Mate3 recognized that.  Then when I switched the load to the batteries (0.1 kW) they dropped to 50 V within 30 seconds, but the temperature compensated voltage went to 50.8 V.  I am still trying to wrap my head around this, but it does seem like the green battery status light (50V and above) does work with uncompensated voltage and will turn yellow after less accumulated load if the batteries are warm, which is pretty well what Outback guy said.  Back in December-February the ambient temperature was lower, and the temperature sensor was not properly located, so the green light stayed on longer for the same accumulated load.  What I am not really sure of, is whether this means anything for the batteries further down the SOC curve - what is the best voltage to work with when figuring SOC from voltage.  Probably T-compensated?  If the batteries are at 33C and the uncompensated voltage is down around 49.2, they really might be around 80% SOC. 

 Interestingly, on my load test for the new battery string I used about 1.8 kWH accumulated load during the night, which is fairly approximate at such low levels, and the batteries were just dropping from 47.6 V to 47.2 V at the end, and when load was removed they recovered to 48.0 V, which being at rest is the same with and without T comp.  The kWh and the log files inverter current amp column data both indicate that I went to about 60% SOC for a 100 amp string.   Yet the Energy Cell manual says a rest voltage of 48.8 is 50% and the Outback guy said a rest voltage of 48.8 is 65%.  And the Mate3 manual says that 60% SOC is 46 V with load.  Granted I only let the batteries rest a few hours, but from my experience in full tests, they do not go up a whole lot more after the first hour or so.  Is there any truth I can cling to here?  I am thinking that if I set my HBX grid-switch to 47.6 V for 30 minutes when running the three banks, I would probably never get that low overnight, but if I did I would certainly be somewhere in the 65-70% SOC range.  Sound reasonable?

 

mcgivor

The whole voltage representation of state of charge is open to interpretation, even with  a known healthy battery is merely a guide at best. A battery which has lost 99% of its original capacity can have a resting voltage which represents 100% state of charge, put a load on it and the voltage collapses to near zero. The information in the following link is extremely informative which answers many questions https://www.solar-electric.com/learning-center/batteries-and-charging/deep-cycle-battery-faq.html#Temperature Effects on Batteries

Estragon

Determining SOC using voltage while discharging (or charging) is difficult. Variables include the size of the load, battery construction, temperature, string wiring, and initial SOC. To assume a 60% SOC at 46v "under load" is assuming away the variables involved. Even fairly complex SOC meters get out of sync because they make assumptions which sometimes don't hold.

Your experience that batteries recover quickly at first, and more slowly as time passes matches my own. I haven't plotted it, but I suspect it would be a logarithmic curve. Voltage after resting for an hour would be reasonably indicative of resting SOC, with the caveat that voltage is likely to rise (or fall) further with more time after discharge (or charge). Depending on the size and duration of the load, even a few minutes resting can give a useful indication.

Resting voltage for a nominal 48v lead acid battery at 50% SOC is about 48.2v.

BB.

Fixed typos from my phone:

Placing the meter tips directly on the metal post or terminal of the device.

If you have a loose binding screw or dirty connection, the cable can be at higher voltage than the device terminal.

Bill

-Bill