Is my battery efficiency set too high

[off-grid-geeks]

From my reading, the 'battery efficiency' setting for the WBjr is a compensation for how many AH need to be measured going into the battery opposed to how many actually count towards the charge.

If the battery were 100% efficient then every AH measured going into the battery would actually be stored for later usage. But batteries are not perfect, so the efficiency setting allows us to compensate for power that is lost in the process.

I have also read in many of the 'Absorb Time' discussions that when you see the falling curve of amps going into the battery start to level off, the absorb is done. When you drop out of bulk and into absorb, a constant absorb point voltage is held, and it is held by dropping off the current flowing to the battery, thereby holding a constant voltage. Once the current levels off, then absorb is done. Those that use 'Ending Amps' to terminate absorb phase rely on this fact.

So, I'm reasoning that if my SOC reaches 100% and my measured current going into the battery is still declining, then I have my 'battery efficiency' set too high. The WBjr believes that enough AH have gone into the battery to fill it, yet those actual amps are showing that it is not full. If I lowered the 'battery efficiency' setting, it should slew the SOC towards the leveling off of the charge amps.

Note that every time I have referred to AH going into the battery, I am using the WBjr amps, not the classic amps, so I am only talking about amps going into the batteries, I am not including amps going to loads.

Please look at this clip of a sample day's charging. This clip shows the WBjr amps going into the batteries curve (orange line) and the SOC (blue line). I have also placed a vertical  line at the point where I feel the charging amps have leveled off, indicating when the absorb phase should really be completed. As you can see, the SOC has been at 100% for a considerable amount of time prior to that point. This leads me to believe that I need to lower my 'Battery Efficiency' setting. It is currently at 94%, and my batteries are 4 years old (FLA GC2).

Thanks for all comments and shared opinions. For me, this is more about my education than having a perfect SOC.

[Vic]

Hi o-g-g,

Being no expert,  let me say the following;

FLAs are not particularly efficient.   IMO,  94% is way too high.   The common wisdom about FLA efficiency is that they are about 85% efficient.   AND,  as they age,  they are probably not THAT efficient.

Also,  IMO,   believe that one needs to wait until there has been no change in the battery charge current for a period of time,  like about 20 of so minutes.

Seems to me,  that your Orange line still has some slope to it.

Guess that the lumps in the Orange line are from varying loads on the system,  with some reasonable voltage drop between the CC and battery (?? ?).

More Later,  good question,   Vic

[off-grid-geeks]

Thanks Vic,

I've seen 2 conflicting opinions for the starting point for the efficiency setting; 85% and 94%.

If there is agreement that my graph shows I'm too high, then my next attempt was going to be 85% and see what that looked like.

Those bumps in the orange WBjr amps line are my fridge kicking on and off. The ONLY item I have running on an inverter is the fridge, which is actually a new, efficient, 5.5 cubic foot chest freezer repurposed to a fridge by using a brewer's thermostat. Everything else in the house runs at 12 volts.

So, I'm going to config for 85% and see the difference, and anyone else that has a comment or suggestion to make, please chime in.
Thanks!

[Vic]

o-g-g,

AGM batteries,  which are considerably more efficient than are FLAs  run around 94% efficiency,  as I read it.

BUT,  I am no expert.   With FLAs,  the great thing,  is that we can easily actually measure the SOC with a Hydrometer.   So,  on the systems here  I really pay no attention to anything other than Net Ah,  on the first morning after a full-charge (as we Skip three days,  and Charge on the fourth).

Good Luck,    Vic

[dgd]

Ogg,

Where you have drawn the vertical line, where the amps tail off, what is the actual amps reading and what is the Ahr rating of your battery bank?

There is no time axis on the graph so what is the time elapsed from reaching 100% soc to your vertical line.

As an aside I never found fiddling with the battery efficiency figure by +- 5% to make any difference and even if the absorb timing is too short the battery still charges towards a real 100% soc when in float
4 year old FLAs I would tend to keep about 90 to 93% efficiency
Just MHO

dgd

[zoneblue]

From my reading, the 'battery efficiency' setting for the WBjr is a compensation for how many AH need to be measured going into the battery opposed to how many actually count towards the charge.

If the battery were 100% efficient then every AH measured going into the battery would actually be stored for later usage. But batteries are not perfect, so the efficiency setting allows us to compensate for power that is lost in the process.

Thats about right. The "battery efficiency" setting is the coulombic efficiency, that is to say its not the Wh in / Wh out, thats much lower. Wh for Wh you account for both current and voltage, and voltage is lost from heating/internal resistance, and that isnt accounted for in the Ah efficiency.

Unless bob is going to explain his algorithim, its assumed to be a two factor thing, the battery effiency setting scales the charge Ah, and the temperature setting scales the SOC. Theres no peukart, or at least he never mentioned it.

The thing about the AH efficiency is it varys by SOC. During bulk it is 100% efficient. (Each Ah you put in stays in. though you still have a Wh loss). Once absorb starts the (FLA) battery starts to bubble and gradually the AH efficiency falls, tapering to zero once float is reached. The coulombic energy is lost to hydrogen gas production.

The number one hit on google for lead acid battery efficiency gives you the first grpah below, which shows this clearly. (http://electronics.stackexchange.com/questions/153517/what-is-the-charge-discharge-watt-hour-wh-efficiency-for-lead-acid-batteries-a)

That article also references a sandia study (http://www.otherpower.com/images/scimages/7427/Lead_Acid_Battery_Efficiency.pdf), (which is a shockingly poor study when you examine it closely), but it does confirm that battery efficiency depends on where you run your system. For 95% of the year our battery achieves float every day, and for 80% of the year it gets several float hours each day. And our daily drawdown is around 10-15%, thus the battery spends its entire life in the less "efficient" part of the charge cycle. Thus you should factor this into your efficiency setting. However who cares whether your SOC is 90 or 95... its when the bank is low that you need reliable SOC figures. Thus it was our recent experience here to have a month of usually sh*tty weather and we went 21 days without float. This showed me how the simple "average efficiency" falls down. During this kind of lower PSOC cylcing, the efficiency is much higher.

And... AGM changes everything. According to one Rolls document, AGM batterys are 99% AH efficient. This is because there is no gassing. The hydrogen is reabsorbed using catylysts.

My quick analysis of this last month indicates that i have to increase the efficiency to 100% to get the SOC back to 100% after those 20 odd low SOC days. (This analysis used a 15min sample interval hence its rough). See graph below (blue curve was with classic eff set to 94%, and SOC drifted a full 25%, the other two curves are my crude modelling). I have 1sec data, and i will probably redo the analysis using it, but its a bucket load of data, and excell wont handle it, so will need some code. But it somes seem to indicate that 99% is in the right ball park for AGM.

I have also read in many of the 'Absorb Time' discussions that when you see the falling curve of amps going into the battery start to level off, the absorb is done. When you drop out of bulk and into absorb, a constant absorb point voltage is held, and it is held by dropping off the current flowing to the battery, thereby holding a constant voltage. Once the current levels off, then absorb is done. Those that use 'Ending Amps' to terminate absorb phase rely on this fact.

This is a separate issue. Its not easy to see the levelling off point on a curve thats (somewhat) logarithmic. The exact level that the charge acceptance current/power reaches depends on the battery type ( again lower for AGM), and the batterys age. I use 0.005C + 0.5A for our EA setting, and that works well. For FLA this might be more like 0.02C.  Even so during the summer i sometimes see it taper down to as low as 0.0015C. You just have to use a pragmatic compromise. But that also shows you how a decent post absorb float can really put the topping on your absorbs. (Sadly the classic cant be configured to do a post absorb EQ/finish, another subject). See todays curve, went to float at 15:50. (This was with WBJr efficiency set to 99%, and it started the day at 77%. Also notice that there is no discernable jump in the SOC at float.

Note that if you set you EA too low it may not trigger, or, trigger less often thus youll get less frequent SOC reset.

So, I'm reasoning that if my SOC reaches 100% and my measured current going into the battery is still declining, then I have my 'battery efficiency' set too high. The WBjr believes that enough AH have gone into the battery to fill it, yet those actual amps are showing that it is not full. If I lowered the 'battery efficiency' setting, it should slew the SOC towards the leveling off of the charge amps.

On an average efficiency basis: If the SOC jumps up at float, then the efficiency is set too low. If it gets to 100% before float its set too high.

Note that every time I have referred to AH going into the battery, I am using the WBjr amps, not the classic amps, so I am only talking about amps going into the batteries, I am not including amps going to loads.

Please look at this clip of a sample day's charging. This clip shows the WBjr amps going into the batteries curve (orange line) and the SOC (blue line). I have also placed a vertical  line at the point where I feel the charging amps have leveled off, indicating when the absorb phase should really be completed. As you can see, the SOC has been at 100% for a considerable amount of time prior to that point. This leads me to believe that I need to lower my 'Battery Efficiency' setting. It is currently at 94%, and my batteries are 4 years old (FLA GC2).

Thats right.

Thanks for all comments and shared opinions. For me, this is more about my education than having a perfect SOC.

The above is probably too much information, but it is a kinda interesting subject.

[zoneblue]

Is was Trojan not Rolls:
http://www.trojanbattery.com/Tech-Support/FAQ/BatterySelection.aspx

[Vic]

This is an interesting Topic.

Am an agnostic when it comes to Battery Monitoring devices,  as most of them do not have enough parameters to really model the battery being used.   And,  who would know enough to enter all of the parameter data,  anyway ...

It is correct that even FLAs approach 100% efficiency,  at least for about 80% of the Bulk (or so ...   depends) stage,  and during Absorb the efficiency decreases to essentially zero at the last part of Absorb.

Personally,  I would guess that AGM batteries do gas,  and this gassing is much of the inefficiency of AGMs,  but the Catalyst recombines the Hydrogen and Oxygen back into water.  But,  it is the breaking down of water even in AGMs that is the dominant loss of efficiency,  even if,  the gasses are later recombined.    Believe that AGM and Gel LA batteries are Lead-Calcium (plus some other trace metals)  and gas considerably less than do Lead-Antimony batteries.   IMO,  Gel batteries have the maximum charge voltage a bit below the Gassing Voltage.

Have never used AGMs,  and in every other way,  I am NO expert on any of  this ...   and there are SO many variables.   "It all depends  ..."

FWIW,    Vic

[off-grid-geeks]

Thanks for all the good input and great discussion!

DgD, the amps at the vertical line are at 12.2 amps. The bank capacity is 440 AH. The elapsed time from 100% SOC to the vertical line is 1:15

I'm now running with a setting of 85% to see what the difference is.

[dgd]

Ogg,

From that data the charge is levelling out between 2 to 3% of battery capacity which is a good indication of a healthy FLA battery.
I would set up the Classic for an EA setting of 13amps and run an efficiency of 93 for a while and see if the SOC aligns closer to absorb termination.
But do remember this is not an exact science and as others have said there are just way too many variables at play. It doesn't really matter too much if SOC reaches 100 before absorb ends
I have seen my FLA bank after not completing absorb for several days then jump from 95 to 100 when an absorb cycle completes, other times 100% is reached well before absorb completes

Dgd

[zoneblue]

I have seen my FLA bank after not completing absorb for several days then jump from 95 to 100 when an absorb cycle completes, other times 100% is reached well before absorb completes

Exactly. Depending on how low the SOC got the efficiency setting will either over compensate or under compensate.

Vic, i also am wary of battery monitors. However with AGM, you dont have many choices, and its only now that the Mrs can see the SOC on the kitchen wall in 36pt font, that she really has ANY sense of whats going on, whether she can run the little electric oven we have, whether she can go mad with the vacuum, or just do a quick lick.

Heres how i see coulombic efficiency. An amp and and amphour are comprised of coulombs. A coulomb is a finite number of actual electrons. In a closed electric circuit those electrons can therefore go near else but around the circuit. Hence coulombically its 100% 'efficient' right? In a FLA when hydrogen is produced and vented to the atmosphere, the battery in the circuit becomes a hydrolysis generator, and thus electrons become lost to hydrogen (and oxygen) atoms. Therefore the way i see it is if the gases are recombined then the electrons are not lost, and the efficiency is more or less 100% as with the closed circuit. Ill be able to see after running 99% for the next couple of months how this works out.

Also agreed about the difficulty of defining and programming SOC parameters, but all i know is that our classic was reading 25% too low, when it was reporting a SOC of, at its lowpoint, 17%. And that messed with her mind, and i had to translate the figure to what i thought it to be.

What i thought it to to be, was derived from, noticing the progress of partial absorb tapers during the month, and, Blackbox each day reporting the highest battery voltage between 5am and 5.30am. This is quasi 'rest voltage'. It obviously isnt true rest voltage, but, it lets the fridge cycle off. If i correlate these figures during this event, this sort of tallys with the pink 100% curve on the graph. 

If i was to try to write something for the manual, it would be (provisionally):

"For FLA start with 89%, and for AGM start at 99%. Then adjust as needed by observing your charge cycles. The SOC resets each time the battery floats, and if the SOC jumps up at float, then the efficiency is set too low. If it gets to 100% before float its set too high. However be aware that, the efficiency varies by SOC, so use a setting that gives the best result in the band that you tend to cycle your battery. It also varies by battery age, and ultimately, any battery monitor is just a guide".

Maybe this is a good time to vent my little rant about battery monitor shunts. It seems to me that all battery monitors have a basic flaw, by using only one shunt. For 25 extra bucks adding a second shunt would produce a much better picture of whats happening to the battery. One shunt would be in the load path, and the other in the production path. From the first time i ever saw a battery monitor, the question was always in my mind, so that battery is currently charging, but how much of that is production and how much consumption?

Below is the current blackbox home screen. Of the three panels, the energy corner shows charge watts, load watts (i have to fudge this by using use both classic amps and Wbjr amps), SOC and Vbat (for reality check). The scrolling graph shows the last 10mins of production and load watts. The weather helps you make load decisions, and the time, well its the only accurate darn clock in the house.

Those four figures show that the kettle has just boiled, but that otherwise theres just base loads, and that the battery is near end of absorb.

[off-grid-geeks]

Dgd,
Yep, I read awhile back that chasing an accurate SOC is nothing but frustration. I'm not really trying to do that, I just wanted to understand the interaction of the battery efficiency setting and SOC. And thanks to all the responses, I see that I am understanding it.
I'll play around a bit to just to learn more, but will settle for 'close enough'. when I'm done.

[dgd]

Maybe this is a good time to vent my little rant about battery monitor shunts. It seems to me that all battery monitors have a basic flaw, by using only one shunt. For 25 extra bucks adding a second shunt would produce a much better picture of whats happening to the battery. One shunt would be in the load path, and the other in the production path. From the first time i ever saw a battery monitor, the question was always in my mind, so that battery is currently charging, but how much of that is production and how much consumption?

... moving a bit off topic here 

You seem to be referring to a battery monitor plus a load monitor system and somehow integrating the data from both to provide a more complete current usage picture.
With the current Classic's shunts it takes some simple calculations to figure out load and tare usage.
With two or more charge sources its getting more complicated.
I did think that a 2 Classic system would allow a second shunt/WBjr combo to measure other DC current, eg an inverter load,  but it seem this has its own complications.
I found the easy way was to just add other DC and AC current measuring devices placed where needed.
One DC device, the ACS758, proved simple to use with Arduino/Cubie/rPi/BBB, its a high side +ve line device. These are available up to bi direction 200amps, I use a 150A uni type to measure DC usage of my 24v 3.3Kw inverter.
With an Uno or Mega ethernet connected there could be several AC and DC current and maybe an AC voltage/Hz sensor, plus additional temperature probes or other sensors making data available to your Black Box system monitor (in a similar way your Arduino based weather monitor does)

dgd

[off-grid-geeks]

Latest update.
I've been using an ending amps of 13 amps, which is 3% of the C20 rating, as was suggested earlier in this thread.
The WBjr amps absorb curve does level off at that point.

I've been chasing the elusive 100% SOC to match this EA point. I know, I know, its not important. I'm just doing it for the education.
Today I got the two to coincide. The battery efficiency setting that got me to that point was 70%.

Seems low. 4+ year old batteries, Full charge is showing an average SG of 1260.
But it works.

Learning a lot!

[zoneblue]

.. moving a bit off topic here 

A bit!

You seem to be referring to a battery monitor plus a load monitor system and somehow integrating the data from both to provide a more complete current usage picture.
With the current Classic's shunts it takes some simple calculations to figure out load and tare usage.

The problem is that the classic amps and the wbjr amps are not averaged in the same timeframe (a product of the modular nature of the WBJr), and so you get these odd spikes and negative load data. The classic amps is also not as accurate as the WBJr, and theres related issues with teh classic tare.

With two or more charge sources its getting more complicated.
I did think that a 2 Classic system would allow a second shunt/WBjr combo to measure other DC current, eg an inverter load,  but it seem this has its own complications.

Yeah ive never tried followme, so not sure how that would work. Never mind adding other charge sources, as you say.

I found the easy way was to just add other DC and AC current measuring devices placed where needed.
One DC device, the ACS758, proved simple to use with Arduino/Cubie/rPi/BBB, its a high side +ve line device. These are available up to bi direction 200amps, I use a 150A uni type to measure DC usage of my 24v 3.3Kw inverter.
With an Uno or Mega ethernet connected there could be several AC and DC current and maybe an AC voltage/Hz sensor, plus additional temperature probes or other sensors making data available to your Black Box system monitor (in a similar way your Arduino based weather monitor does)

Ok so im interested, but.... In order to keep EA, the existing low side battery line shunt needs to stay. Thus one other shunt either in the charge or the discharge line, ought to do it. However to account for both shunts in the same time frame, i guess that takes us back to a matched pair of shunts and instrument amps, one for each of charge and discharge paths. That means 3 shunts, and 3 lots of 10x mV voltage drops. Joy.

Ok so use the WBJr shunt, with another amp. Thats down to two. Both in the low side, or one low and the other high? We kinda want to avoid negative bus potential differences, so maybe high side as you said.