battery SOC question

lamplight

OK for all these years i have been using the simple battery voltage as a general method of determining depth of discharge or state of charge, as you will. I have learned on this forum that the battery voltage will likely be a little lower UNDER LOAD , so that might not be always accurate but IF under load, take that into account.

With my recent upgrades and significant investments im trying to do things right: battery monitor, hydrometer, etc...but i do have a slightly different question on battery voltage vs soc:

I've noticed that when i run my cappuccino machine (WHAT? a man has to live!?!) at night it can draw the battery voltage down quite a bit. no surprise there its about 1400 watts AC for a while there. but anyhow... I noticed that say this morning my battery voltage is 12.51, which if going off a chart like this:


can indicate a state of charge of roughly 90%. My new trimetric battery monitor (LOVE IT!) reports my battery bank is 97% full.

The ALWAYS ON load is about 35 watts DC (web server and inverter).

I will notice that if i don't run any high draw loads the battery voltage will sit at a much higher value the next morning, so that both numbers (battery voltage and percent full from meter) are closer and make more sense (ie: with just the basics running (under load). after a good healthy charging day, i can have 12.7v hours after the sun went down still).

so i am guessing there is some logical reason for this, ive read you guys talking about some strange effects on L.A. batts from certain current levels being withdrawn.. etc... but my question is: which number should I trust the most: the trimetrics value for state of discharge or the voltage, or neither value completely? i know some of you (everyone?) will say the only test is to have no load, let it rest for some time and measure the specific gravity. which yes i intend to do on occasion, but hopefully not too often as its a major pita Just curious what info i CAN get from the meter and the battery voltage.. or more specifically, can i trust one more than the other? clearly battery voltage only does not offer the best gauge of SOC or DOD. hope i phrased the question correctly.

OH, and please also note that im just getting fa,iliar with my trimetric and some parameter values may need to be adjusted to get accurate readings, but so far i think its pretty much right on.

that you for your collective experience as always!

RCinFLA

Re: battery SOC question

In battery operation there is a process called 'kinetics'. This means things are moving around and current is flowing. The chart you show is a battery at rest state, in static equilibrium with no current flow.

When a load is applied there is some of the battery voltage used to move the kinetic processes. This include the movement of molecules in electrolyte, creating and movement of ions, and the molecular state changes of creating lead sulfate and water (from sulphuric acid).

You may have heard of 'Peukerts' effect on battery discharge/capacity. There are a couple of components of this overall effect. Part is rest recoverable due to redistribution of molecules, part is not rest recoverable due to site blockage on the plates.

Basic chemistry of lead acid is positive plate is coated with lead dioxide, negative plate is pure lead, and electrolyte is 32% sulphuric acid/68% water. During discharge, lead is combined with sulphur on positive and negative plates to a brown fuzz of lead sulfate. Sulphuric acid in the electrolyte is converted to water(sulphuric acid providing the sulphur to make lead sulfate on plates). The equilibrium rested voltage is very close to 0.845 + specific gravity for each cell. This means the rested voltage will drop as battery has had more discharge. This lower specific gravity is due to sulphuric acid depletion in electrolyte.

The negative plate is made of porus lead (sponge like) to provide greater effective surface area for sulphuric acid to find and convert to lead sulfate. This allows electrolyte to filter in and find a lead molecule to converter to lead sulphate during discharge. These pores can be plugged up during discharge, particular with high current flow, preventing electrolyte from reaching available lead. This capacity loss will not recover after a rest state (not to be confused with recharging which can clear the blockage).

The other major kinetic factor is dispersal of electrolyte. During discharge the water created must move out of the way to allow sulphuric acid to move in and get access to the plates. High discharge rates will cause a 'backup' of dispersal. This part of the 'Peukerts' process is recoverable when battery is allowed to rest (or go to a lower current draw). The rest allows the electrolyte to remix and become uniform, clearing the higher concentration of water in the plates proximity. Battery terminal voltage will rise as this remix happens.

Another small part of 'Peukerts' loss is the effective series resistance of battery. This is terminals, and plate grid connecting structure. When current flows there is some unrecoverable loss in capacity due to this resistive element. (goes to heat loss)

As to the Trimetric monitor, it will tell you a percentage based on what you put in for battery capacity. It measures average current (as opposed to r.m.s. current) and computes % full based on subtraction from AH capacity you programmed in. It resets itself to 100% (Full) when absorb voltage has been reached for some period of time. This is done to wash out cumulative discharge/recharge summation inaccuracies.

It does not account for Peukerts effect or heavy current draw versus light current draw effect on battery capacity. (or aging capacity loss of battery).

It does have a program entry for recharge efficiency. This is appled to injected recharge current. Again, there is no account for charge efficiency based on state of charge. Below 80% SOC the charging efficiency is quite high, like 95%. Above 80% SOC the charge efficiency drops off to 60% to 70%. An overall combined average recharge efficiency that is typically used is 85%. I believe that is the default number factory set in Trimetric unit.

nsaspook

Re: battery SOC question

The research and methods I have done for my DIY monitor to tell the true SoC/SoH of a battery in a dynamic charge/load condition shows it's a pretty complex process. I currently am logging data every 30 seconds in to a 4Gb SD card (up to 7 years of data) on the controller to get long term power generation vs usage profiles to dynamically update the real capacity, charge efficiency , age and Peukert factors.

What seems to be missing from most monitors is a measure of internal resistance. My plan is to pulse the battery with a series of two load resistors at night and use a method similar to this to generate SoH with the long term data instead of fuzzy logic. One day I might even get it working.:D

http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=/netahtml/PTO/srchnum.htm&r=1&f=G&l=50&s1=7,072,871.PN.&OS=PN/7,072,871&RS=PN/7,072,871

http://www.buchmann.ca/Chap9-page2.asp
http://www.cadex.com/support/analyzers.asp

lamplight

Re: battery SOC question

whoa. thanks guys.. i understand some of that fortunately its in print and i and others can reread for better comprehension in general i get it,

you're saying i should trust neither one

thanks!!

RCinFLA

Re: battery SOC question

You can measure battery internal resistance with a transformer that puts out about 48vac or above and at least 10 amps. You put a series AC motor start or run capacitor sized to get at least 10 amps of AC current. You can assume the battery is zero resistance for the cap size calculation. This makes an approximate 10 amp, 60 hz. current source.

For a 48 vac transformer. the capacitor value would be about 550 uF. An oil filled run cap would be better then electrolytic start cap but 550 uF run cap would be a bit expensive. If using a start cap you run it just long enough to measure the AC voltage ripple on the battery as the start caps are made for short active period.

A 120v 1kw isolation transformer would work even better as the higher the voltage the lower the series cap value. For 120vac at 10 amps, the cap value drops to 220 uF, a range that is reasonable for an AC run cap.

If you want to live dangerously, you could use AC grid directly. Make sure neutral line is used for your measurement ground reference, cap in hot line.

You should run the battery down a little, to about 90% SOC because a fully charged lead acid battery just taken off charger will have a higher reading series resistance .

You need a voltmeter capable of better then 1 mV AC resolution which would correspond to 0.1 milliohms of battery Rs for the 10 amp AC current source. I use an old HP400G AC voltmeter across battery terminals.

Positive grid corrosion due to overcharging is a common cause of increased battery internal resistance on a not so old battery. Other extreme is sulfated battery that has a lot of plate area under a lead sulfate crystals insulation blanket.


Rs of battery is not too much a factor in capacity measurement unless you have a damaged/very old battery or drawing greater then 20% AH rating current from battery. Rs does increase as battery discharges but it not a reliable prediction of SOC. It is a good indication of state of health of battery.

RCinFLA

Re: battery SOC question

lamplight wrote: »

whoa. thanks guys.. i understand some of that fortunately its in print and i and others can reread for better comprehension in general i get it,

you're saying i should trust neither one

thanks!!

The Trimetric is a good little unit. I use one. I programmed battery capacity to half my actual battery capacity to keep me under consumption control.

lamplight

Re: battery SOC question

RCinFLA wrote: »

You can measure..

You lost me right about there... LOL :roll:

no matter.. i have the basics down and im blindly hopeful i will not ruin these batteries before their time really great description of the lead acid battery in your other post before that. thank you !

nsaspook

Re: battery SOC question

lamplight wrote: »

whoa. thanks guys.. i understand some of that fortunately its in print and i and others can reread for better comprehension in general i get it,

you're saying i should trust neither one

thanks!!

You can trust your meter. It's just that what it can do is limited to mainly static conditions of power in/out. That works well for systems with lots of margin (50% derated) but I'm working to find the max energy that can be supplied with limited solar inputs and still keep the storage system healthy. I've hurt a few batteries already with trial runs. (Thank goodness for 1 year warranties)