Battery charge levels up and down like a yoyo

[Cardo]

We've been sat in a harbour with no services for the last 4 days. Sun's been shining pretty much all day.

We have 500Ah of sealed LAs in our domestic bank (though I have a sneaking suspicion the capacity is nowhere near that much). We have 300 watts of solar panels (though not in the most efficient places, so aside from a couple of hours a day when the sun is right up at the top, there's usually enough shade on one of the side panels to make its output negligible). We also have an Aquair hoisted in the rigging, though that doesn't do too much.

We have a SmartGauge keeping an eye on the batteries, in hopes that we'd have an idea of the charge levels.

We have a 50l front opening fridge, and we use laptops, iPad, phones and TV a fair bit, so these are often sat on charge. We also run the watermaker for an hour or two a day, when the sun is strongest and we're getting 10+ amps from the panels. Not really any other drains at the moment, as we're not going anywhere.

Up until yesterday, the Smartgauge was telling us we were sitting in the 70 - 80% range. Batteries would be low 70's in the morning, and we'd be up to around 80 or low 80s by the evening. Voltages in the evening seemed decent, around 12.7 - 12.8 during rest, and down to 12.5 - 12.6 with a load.

However, yesterday was a bit of an odd'un. We ran the engine for a half hour to warm up some showering water, which usually puts at least a couple of % points on the batteries. The sun was also pretty strong. By the evening, the batteries had shot up to 91%. Voltages looked good, high 12.somethings once the sun had gone down. By the time we went to bed, having watched a film on my laptop (plugged in, did seem to be using a fair bit of power), they'd dropped to 81%.
This morning we're down to 68%. Voltages (though the panels are starting to work) are around 12.5 and dropping down to 12.3ish with the fridge running.

Anyone have any ideas what's happening here?

 

[GHA]

Could it be the smartgauge getting a bit confused while the batteries were being charged?

http://www.pbase.com/mainecruising/smart_gauge

The Smart Gauge seems to work as advertised on GEL, AGM and FLA batteries in discharge mode.


What does that mean?


It means that I did see the Smart Gauge get a bit confused when the bank was being charged. It can't really track the capacity of a battery charger now can it...? However we are only talking about 10-12% variation

​

 

[RAI]

Could it mean that the Smart gauge now has a better handle on the true capacity / state of charge of your batteries after having them raised to a higher voltage and watched them descend.

 

[halcyon]

This morning we're down to 68%. Voltages (though the panels are starting to work) are around 12.5 and dropping down to 12.3ish with the fridge running.

Anyone have any ideas what's happening here?

Not a lot, 12.5 is around 70% , though battery temp effects it, 12.3 on load is reasonable, but 12.7 / 12.8 is nearer 90 / 100% try letting the batteries rest and give the smartgauge chance to reset charge level.

Brian

 

[William_H]

Another near tragedy averted

News tonight has story of a 62 yo lone sailor who was knocked of his boat in big seas. He had a lifejacket on and swam for 3 hrs to shore. He crawled out on a stone break water and got some fishermen to help him. He is now in hospital with hyperthermia. The boat was a popular local design S80 26ft more of a racing boat than cruising but very stiff. Powered by outboard motor and pictures looks like he did not have life lines. Seas have been rough due to a passage of a strong front over the weekend.
He was doing a voyage to Bunbury about 100NM south of from Fremantle. He apparently did not set up search and rescue time with local rescue group. They have installed repeaters down the coast so all in VHF range. He was lucky to survive. No one knew he was there. His boat was pictured being towed in apparently none the worse for wear. I recognised the boat as one that used to race at our club. The moral of this story is use a safety harness. Do not fall off your boat. be careful; olewill

 

[sailinglegend420]

I could make some smart comment that SmartGauge would have saved him..........

The OPs situation clearly shows SmartGauge doing its job. If you don't think it should be this low then check the batteries by isolating each one in turn from all charging and loads for at least 4 hours and check their open circuit voltage. You might find one is duff and dragging the others down, in which case you won't have to wait 4 hours to see that!!!

You thought the capacity was less than it used to be and this is showing as a bigger loss of SoC overnight than you have been used to. An old battery monitor might not have shown you this so dramatically.

You don't say who long you have had Smartgauge, but it does take many charge and discharge cycles to fully learn your batteries. It keeps on learning and will only become more accurate, unlike a conventional shunt based Battery Monitor that becomes more and more inaccurate.

SmartGauge is less accurate when charging, but mine tends to say you are 100% full too early. You say you work in the 70 to 80% range - how often to you get back to 100%?

 

[tawhiri]

Is the battery monitor showing the real battery voltage. My BEP battery monitor is about 0.2 volts higher than reality.
If I don't see 13 volts or more on the monitor then the bateries are under charged.
500Ah should be plenty to run for several days before getting to half charged, but if one of the bank is on its last legs, then it will drag the rest down.
by the way what are LAs

 

[Cardo]

I could make some smart comment that SmartGauge would have saved him..........

The OPs situation clearly shows SmartGauge doing its job. If you don't think it should be this low then check the batteries by isolating each one in turn from all charging and loads for at least 4 hours and check their open circuit voltage. You might find one is duff and dragging the others down, in which case you won't have to wait 4 hours to see that!!!

You thought the capacity was less than it used to be and this is showing as a bigger loss of SoC overnight than you have been used to. An old battery monitor might not have shown you this so dramatically.

You don't say who long you have had Smartgauge, but it does take many charge and discharge cycles to fully learn your batteries. It keeps on learning and will only become more accurate, unlike a conventional shunt based Battery Monitor that becomes more and more inaccurate.

SmartGauge is less accurate when charging, but mine tends to say you are 100% full too early. You say you work in the 70 to 80% range - how often to you get back to 100%?

The 70 to 80% area where we were the previous few days seemed like the right ballpark. And down to 68% this morning was probably about right, I'm just curious as to why it shot up to 91% yesterday, only to tumble to 68% this morning. That's quite a big drop, far more than we normally get overnight. I can only assume the SG got confused as to how much charge had gone into the batteries and was over reading.

The SG has been fitted for the past 2.5 years, together with the batteries, so it should have learned them, by now!

We rarely get into the 90's when parked up with only solar/wind, but if we motor for 2 to 3 hours it normally pushes the batteries into the mid-high 90s. If we motor for longer, or there's plenty of sun once we arrive, we sometimes end up on 100% before the sun goes down.

 

[Cardo]

Is the battery monitor showing the real battery voltage. My BEP battery monitor is about 0.2 volts higher than reality.
If I don't see 13 volts or more on the monitor then the bateries are under charged.
500Ah should be plenty to run for several days before getting to half charged, but if one of the bank is on its last legs, then it will drag the rest down.
by the way what are LAs

I've checked each battery by leaving it disconnected for an hour or so, but not seen a noticeable drop in voltage. I don't think any particular battery is duff. And we can normally go for a few days before we reach 50%. Though it does depend on how much we're taking out of them. The fridge uses a lot, as does the watermaker.

LA = Lead Acid

 

[VicMallows]

You don't say who long you have had Smartgauge, but it does take many charge and discharge cycles to fully learn your batteries. It keeps on learning and will only become more accurate, unlike a conventional shunt based Battery Monitor that becomes more and more inaccurate.

Sorry, but I much prefer to charge my batteries fully, monitor how much power has been taken out ..... and how much put back in, and then draw my own conclusion as to remaining charge. Trying to predict available/remaining charge purely on (even a 'learned') voltage-based algorithm is a total anathema to me.

 

[pvb]

Sorry, but I much prefer to charge my batteries fully, monitor how much power has been taken out ..... and how much put back in, and then draw my own conclusion as to remaining charge. Trying to predict available/remaining charge purely on (even a 'learned') voltage-based algorithm is a total anathema to me.

Have to agree with you. My main experience with battery monitors has been with the Xantrex Link 10, which resets itself once the batteries are fully charged. I certainly didn't notice it becoming "more and more inaccurate", but then I only had it for about 15 years, so maybe sailinglegend420 is talking about longer periods.

 

[Mistroma]

Have to agree with you. My main experience with battery monitors has been with the Xantrex Link 10, which resets itself once the batteries are fully charged. I certainly didn't notice it becoming "more and more inaccurate", but then I only had it for about 15 years, so maybe sailinglegend420 is talking about longer periods.

I assume your Xantrex Link 10 resets itself to read 100% capacity when it reaches full charge and that you need to program battery capacity (i.e. Ah expected) and this bit is where the guess work comes in. A new 100 Ah battery might only deliver 90Ah, get 100Ah after a while and then lose capacity again. So telling you that you have 100% of nnn Ah left isn't much help if your actual capacity turns out to be 70% of nnn. The user needs to adjust the system to reflect actual capacity in order to get an accurate % figure. I believe that Smartguage doesn't say anything about actual battery capacity, just that it is currently at a certain percentage of whatever capacity it is actually capable of holding.

Perhaps the Xantrex unit is more complex and does use some method of continually estimating and updating actual battery capacity. I don't think that many "amp counting" type meters do this and sailinglegend420 is probably referring to the dumber variety.

I think that this might be what sailinglegend420 is talking about, perhaps he will confirm this point or explain my error.

 

[pvb]

I assume your Xantrex Link 10 resets itself to read 100% capacity when it reaches full charge and that you need to program battery capacity (i.e. Ah expected) and this bit is where the guess work comes in. A new 100 Ah battery might only deliver 90Ah, get 100Ah after a while and then lose capacity again. So telling you that you have 100% of nnn Ah left isn't much help if your actual capacity turns out to be 70% of nnn. The user needs to adjust the system to reflect actual capacity in order to get an accurate % figure. I believe that Smartguage doesn't say anything about actual battery capacity, just that it is currently at a certain percentage of whatever capacity it is actually capable of holding.

Perhaps the Xantrex unit is more complex and does use some method of continually estimating and updating actual battery capacity. I don't think that many "amp counting" type meters do this and sailinglegend420 is probably referring to the dumber variety.

I think that this might be what sailinglegend420 is talking about, perhaps he will confirm this point or explain my error.

The Link 10 only has a crude LED bar graph to indicate percentage charge, but it does have an Ah counter which is a good indicator (with the voltage reading) of the charge state of the batteries.

Naturally, for best results, it's important to program the monitor with a realistic battery capacity. Every time the batteries reach the fully charged state, the Link 10 resets the Ah counter to zero, so it can't become "more and more inaccurate".

 

[Mistroma]

Yes, my point exactly, it resets the Ah counter and you say it needs a realistic entry for capacity. Even if you run an initial battery capacity check to find that your 100Ah battery is currently only 95Ah then it might be 90Ah in a month or so. If the initial test was at C20 rate and you actually draw at C15 or C30 the capacity will be different again.

So the Ah available is a moving target. I think that's why the Link 10 might become less accurate over time (vs. a different system which doesn't need to be told about battery capacity). Perhaps you meant that the level of the error didn't bother you with the Link 10 and the readings were good enough for your needs.

 

[[2574]]

So telling you that you have 100% of nnn Ah left isn't much help if your actual capacity turns out to be 70% of nnn. The user needs to adjust the system to reflect actual capacity in order to get an accurate % figure.

Succinctly and accurately put; a statistic of "x% of an unknown" is not a very useful statistic. As assessing AH capacity seems to be a black art I can't see the value in reprogramming an AH counter as then the formula is x% of a guess; equally unhelpful. Maybe if the Smartgauge really is as accurate in assessing % SoC as is discussed hereabouts then we should work it backwards from there - so if a battery is 70% charged (per Smartgauge) after 48hrs at anchor and we assess our 24hr average consumption to be 3A/hr (from watching the ammeter) then capacity must be 480AH; call it 500AH, near enough. I mentioned on another post that both Smartgauge and Ammeter/AH counter is the best monitoring solution; I am increasingly convinced of the need for both.

 

[GHA]

I mentioned on another post that both Smartgauge and Ammeter/AH counter is the best monitoring solution; I am increasingly convinced of the need for both.

Pity smartgauge don't offer a model incorporating an ammeter and ampH counter, that really would be a must have.

 

[sailinglegend420]

..... I certainly didn't notice it becoming "more and more inaccurate", but then I only had it for about 15 years, so maybe sailinglegend420 is talking about longer periods.

OK let me give a long and detailed explanation as to why SmartGauge is much more accurate than a shunt based Battery Monitor. A leading US marine engineer has just posted his results of lengthy tests with the SmartGauge and is very surprised by the results. He had put SmartGauge and their claims under a file - "Snake Oil"!

For a battery monitor to display the Ah capacity properly the battery parameters have to be correctly input to the software on installation, but many people leave the default settings as they are. Nasa do not allow two very important parameters to be programmed for different batteries – Peukert’s Constant and the Charge Efficiency, but they seem to use the battery voltage to help "integrate" the Ah capacity.

The one value that MUST be input is the battery capacity of the service bank. This changes over time as the batteries sulfaate and there is no simple way to determine their true value. An estimate of the capacity reduction can be guessed and input every 12 months.

The other parameters that are needed are Peukert’s constant for the batteries being used – sometimes this is not easy to find from the battery manufacturers. Peukert’s constant adjusts the actual discharged Ah values depending on the size of the discharge current – because Peukert’s law is logarithmic a very small changes can make a big difference to the Ah count, so change this value carefully! If a battery monitor does not allow for Peukert’s to be entered then it will be inaccurate, especially with batteries like AGM and Lithium. Peukerts value also changes as the batteries age. There is an effect similar to Peukert's effect during charging where the Ah returned depend on the actual charging current – so charging at 40 amps does *not* return twice as much energy as charging at 20 amps - it returns less). Unfortunately, unlike Peukert's effect, there is no way to calculate this effect.

An accurate charging Ah count is also affected by the “efficiency” of the battery at accepting charge. Wet cells may be only 75% efficient whereas AGMs may be 98%. These figures are given as an overall efficiency, but they change with the State of Charge. At 50% SoC the battery efficiency maybe nearly 95% and at 85% SoC the battery efficiency maybe 50%. If a bank is always only discharged to 80% before re-charging then the Ah count will be wrong.

If all these values are not input correctly and updated as the batteries age the Ah count will become less and less accurate.

Batteries also have a self-discharge up to about 10% per month which the battery monitor cannot detect. This will also introduce large errors.

Auto Re-Sync Feature
To increase their accuracy the Ah count must be reset to 100% when the bank is 100% charged.

A manufacturer’s Auto Re-sync feature will have default values so that when the charging current is typically below 2% of the battery Ah Capacity (C) for 5 minutes at a voltage above 13.5 volts then the Ah capacity is reset to 100%. These values are usually much too conservative, but they do want the battery monitor to reset earlier rather than never.

A battery is fully charged when the current is 0.5% C, but at a voltage of about 14.4v. This voltage/current cannot be used as it is impossible to achieve because all regulators will have reset to a float voltage much lower than this long before the current has reduced this far.

“The Reset Gochya”
There is also a problem with solar panels that under cloudy conditions may not deliver a high enough current but may meet the Auto Re-sync voltage values. This will prematurely re-set the Battery Monitor to 100% too early.

This auto re-sync feature can be disabled completely by raising the reset voltage to something like 16 volts – which it should never reach. Then the Battery Monitor can be reset manually when YOU know the batteries are fully charged.

The Battery Monitor state of charge readings should always be crossed checked with the service battery voltage as a rough guide to the accuracy of the displayed Ah count. At 50% DOD the voltage should be about 12.2 volts - when no current is being drawn and there are no charging sources.

With all these inaccuracy problems the Military finally demanded more accurate Battery Monitors - so enter the SmartGauge which only measures battery voltage. The man who invented it used to work for Link and got fed up spending all his time on the phone explaining why their Battery Monitors were not accurate. I have a BEP battery Monitor and bought the SmartGauge because the BEP was becoming useless, no matter how much I tried to change the parameters.

 

[sailinglegend420]

Pity smartgauge don't offer a model incorporating an ammeter and ampH counter, that really would be a must have.

Having spoken to the MD I know they will never do that because they don't need to, but more importantly they know that the shunt may be installed incorrectly. Maine Sail, who has just tested the SmartGauge, has said that a very high number of the installations he sees are wrong - I think over 75%, and 90% of the Battery Monitors are programmed incorrectly. Always buy a cheap Digital Amp meter with the SmartGauge.

 

[[2574]]

Matt,

Erudite and enlightening.

Thanks,

Rob

 

[pvb]

If the initial test was at C20 rate and you actually draw at C15 or C30 the capacity will be different again.

That's not really relevant, as the Link 10 calculates using Peukert's exponent, so compensates for differing discharge rates. The key reason for setting the capacity fairly accurately is that the zero reset when fully charged requires, as one of its parameters, the charge current to drop below 2% of the battery capacity figure (ie charge current below 4A for a 200Ah capacity).

I reckon knowing how many Ah you've used is, together with the voltage reading, a pretty useful indication.