battery monitoring

[Graham.m]

With the batteries were at their low point I did measure the sg of a couple of cells. It was 1.212, significantly higher than I expected. I think it was due to poor mixing and I couldn't find my large "turkey baster" to give it a good mix, the pipette thingy you get with the refractometers is very small.

 

Well taking that as termp corrected that would be of the order of Soc of 68%. So maybe the SoC was not as low as you believed, Surely the top layer of electrolyte would be a low SG compared with the lower layers as the sulphuric is heavier than the water.

[nicknorman]

Well taking that as termp corrected that would be of the order of Soc of 68%. So maybe the SoC was not as low as you believed, Surely the top layer of electrolyte would be a low SG compared with the lower layers as the sulphuric is heavier than the water.

 

The mastershunt was on 57% so the actual SoC couldn't be 68% unless the batterie had a lot more capacity than badged, which seems highly unlikely.

 

The stratifying effect you mention is a slow process. More likely I think is that the acid around the plates was weaker and had not mixed with the stronger spare electrolyte above the plates. It's a test to do when I find the "turkey baster" but not today as I'm flying home.

[Graham.m]

The mastershunt was on 57% so the actual SoC couldn't be 68% unless the batterie had a lot more capacity than badged, which seems highly unlikely.

 

The stratifying effect you mention is a slow process. More likely I think is that the acid around the plates was weaker and had not mixed with the stronger spare electrolyte above the plates. It's a test to do when I find the "turkey baster" but not today as I'm flying home.

 

From using a hydrometer without stirring I have found readings tend to be lower than stirred. Hence my normal process is to fill and empty the hydrometer twice and read the third.

 

As to capacity I have no idea but strange things happen with batteries. Maybe next time run hydrometer readings in parallel

 

Have a good flight home

Edited January 26, 2016 by Graham.m

[nicknorman]

The other issue is that I haven't topped up the electrolyte for a long time. It's about half way between where it should be, and the top of the plates. Perhaps that is why the sg was high.

[Top cat]

What a minefield this determining state of charge is, it seems every method open to us has errors. It seems to me that the best we can do is be conservative in how deep we discharge them and rely of the charging current tailing off to tell us when they are charged and noting how many Ah we put back in compared to the norm.

I'm just hoping that the technology being developed for electric cars can be applied to boats soon, but I ain't holding my breath.

 

Top Cat

[Graham.m]

The other issue is that I haven't topped up the electrolyte for a long time. It's about half way between where it should be, and the top of the plates. Perhaps that is why the sg was high.

 

So many variables to this thing call SoC, I think the only good way is Specific Gravity on a battery with full electrolyte. Maybe a full charge top up and see where it goes

[nicknorman]

So many variables to this thing call SoC, I think the only good way is Specific Gravity on a battery with full electrolyte. Maybe a full charge top up and see where it goes

Possibly, but most people can't be bothered with that. Boating is about having fun, not messing with dangerously corrosive chemicals! Hence the compromise of the Smartgauge!

 

Heading back to the marina (in a gale!) and thus on charge. MS reading 81%, SG 79%, which is a proximity of readings I'm more used to seeing.

Edited January 26, 2016 by nicknorman

[Graham.m]

Possibly, but most people can't be bothered with that. Boating is about having fun, not messing with dangerously corrosive chemicals! Hence the compromise of the Smartgauge!

 

Heading back to the marina (in a gale!) and thus on charge. MS reading 81%, SG 79%, which is a proximity of readings I'm more used to seeing.

 

Those reading suggest that whatever they are tracking they give a similar result

 

But that is the problem those instruments, whatever type, have to give reliable readings otherwise the users who use their batteries hard and discharge down to 50% according to one of those gauges could be shortening battery life. That is one of the reason why I recommend discharging to 25% rather than 50% it covers the area of possible errors and gives some leeway in emergencies.

 

Hope the wind is/was not too much of a problem and things are going OK

[WotEver]

Possibly, but most people can't be bothered with that. Boating is about having fun, not messing with dangerously corrosive chemicals! Hence the compromise of the Smartgauge!

Not only that but as the battery ages and becomes more sulphated so the specific gravity readings become correspondingly less accurate.

 

As has been said several times in recent posts it just goes to show what an imprecise science it is.

 

Tony

[Graham.m]

Not only that but as the battery ages and becomes more sulphated so the specific gravity readings become correspondingly less accurate.

 

As has been said several times in recent posts it just goes to show what an imprecise science it is.

 

Tony

 

No SG readings do not become less accurate, they are still accurate, they become lower as some of the sulphuric acid is not released and highlight the problem of a dying battery. That highlights the need to take remedial action such as a high voltage charge to reduce the sulphation. Something none of the gauges highlights.

[Tony Brooks]

Well I would say that a correctly read the rested voltage is indicating a different state of charge to a hydrometer one of the readings will be inaccurate as far as state of charge is concerned. This is really all an ordinary boater needs to know.

 

I know it is still an accurate reading of the acid's relative density but it an inaccurate reading of a battery's state of charge.

 

I think Tony's statement is a fair one to be made on this forum. It would be different if the forum only catered for highly technical people who would discuss the meaning of accuracy without being confused.

[nicknorman]

Those reading suggest that whatever they are tracking they give a similar result

 

 

Hope the wind is/was not too much of a problem and things are going OK

I normally reverse along the pontoons (long pontoon with 7 or 8 boats, we are at the far end, just a foot or so of clearance each side) but after 2 attempts to get positioned to reverse into the slot it obviously wasn't going to work out without clonking the boats at the mouth of the pontoons (both occupied!) so I took an executive decision to wimp out and go in forwards. It just means I have to use the long power lead and traipse it the length of the boat.

 

Anyway, now on shore charge and both MS and SG showing 86% with 34A. I have to leave at 1pm so unfortunately I won't be able to see fully charged being achieved but I'll log it until I go.

[peterboat]

The other issue is that I haven't topped up the electrolyte for a long time. It's about half way between where it should be, and the top of the plates. Perhaps that is why the sg was high.

cant you fit watering systems to trojans, thats one of the great things i like with full tractions just check the bottle and top up if needed

[WotEver]

No SG readings do not become less accurate, they are still accurate, they become lower...

And a less accurate measure of SoC.

 

Tiny

 

Tiny?

[WotEver]

cant you fit watering systems to trojans, thats one of the great things i like with full tractions just check the bottle and top up if needed

Yes

http://www.trojanbattery.com/products/hydrolink-watering-system/

[cuthound]

As has been said several times in recent posts it just goes to show what an imprecise science it is.

Tony

But any fule knows that batteries are based on black art and not science

[nicknorman]

cant you fit watering systems to trojans, thats one of the great things i like with full tractions just check the bottle and top up if needed

Yes you can. But I've had them 2 years and only filled once so far - although they need doing again - so it's not much of a chore to do it manually. I have good access to the tops too.

Edited January 26, 2016 by nicknorman

[nicknorman]

Bit rushed today, had to get back to the marina in time to put the boat to bed, visit my mother, and then get back up the M42 in rush hour to the airport. And it was REALLY WINDY! However I remembered to log some data during this morning's cruise back / recharge, but missed the first bit. Here it is. So had the MS been a BMV out of the box it would have gone to 100% at 18A (4%). In the event the SG reached 100% at 12A (2.6%) which is perhaps a little high for some (including me) but not too bad. Some folk including Victron recommend 4% as a compromise between genny run time and battery life.

 

You'll notice I've put a second order regression line on the MS's curve (yellow dashed line). I haven't done the maths yet but I am pretty sure it would never reach 100%. Or more simplistically if you look at the AH needed to be put back to satisfy the MS - 6% or 27 AH - and the rate of decrease of current, its hard to see how 27AH is going to get back into the batteries. The MS would take a jump from some mid 90s value, straight to 100%, once the current got down to say 1%. This is a reflection of the accumulating error in the AH gauge despite it having the correct capacity and doing just 2 cycles between 100%. Of course it is in "dead reckoning mode" all the time it is not at 100%.

 

 

Graph is a bit lumpy due to the resolution of 1% over a fairly small range. Also I moored up and changed from alternator power to shore charger power, hence the slightly odd current curve.

Edited January 26, 2016 by nicknorman

[nicknorman]

Ok so doing the math on the regression formula (using c-b^2/4a) gives me a peak of 94.42%. Is that right? (Seems a little low). Of course there is nothing to say that the shape of the curve will remain constant (extrapolating is always dangerous!) but it certainly seems unlikely to get close to 100%

[Graham.m]

Bit rushed today, had to get back to the marina in time to put the boat to bed, visit my mother, and then get back up the M42 in rush hour to the airport. And it was REALLY WINDY! However I remembered to log some data during this morning's cruise back / recharge, but missed the first bit. Here it is. So had the MS been a BMV out of the box it would have gone to 100% at 18A (4%). In the event the SG reached 100% at 12A (2.6%) which is perhaps a little high for some (including me) but not too bad. Some folk including Victron recommend 4% as a compromise between genny run time and battery life.

 

You'll notice I've put a second order regression line on the MS's curve (yellow dashed line). I haven't done the maths yet but I am pretty sure it would never reach 100%. Or more simplistically if you look at the AH needed to be put back to satisfy the MS - 6% or 27 AH - and the rate of decrease of current, its hard to see how 27AH is going to get back into the batteries. The MS would take a jump from some mid 90s value, straight to 100%, once the current got down to say 1%. This is a reflection of the accumulating error in the AH gauge despite it having the correct capacity and doing just 2 cycles between 100%. Of course it is in "dead reckoning mode" all the time it is not at 100%.

 

Recharge2.JPG

 

Graph is a bit lumpy due to the resolution of 1% over a fairly small range. Also I moored up and changed from alternator power to shore charger power, hence the slightly odd current curve.

 

Ok so doing the math on the regression formula (using c-b^2/4a) gives me a peak of 94.42%. Is that right? (Seems a little low). Of course there is nothing to say that the shape of the curve will remain constant (extrapolating is always dangerous!) but it certainly seems unlikely to get close to 100%

 

Nick many thanks for the graphs, well done. I wish I were better with words, my apologies that I am not.

 

Maybe there is a basic problem here, we have no known reference to compare against the gauges. You have said in the past that your batteries are fully (100%) charged when the tail current was 1 amp and I seem to remember 1/2 amp also being quoted. Yet here there seems to be an acceptance that 100% has been reached because the Sgauge has reach 100%. The Sgauge has a specified error of +/-10%. Also with a tail current of 15 amps (3.3% of capacity) you appear to have accepted the batteries are at 100% SoC. I think maybe you were tired and rushing the tests.

 

The current slope from 15A onwards will get flatter, as the slope has all the way through the charging process, as it approaches zero current, as the rate of assimilation of the charge gets slower and slower as is normal for a charging battery, as will the slope for the Msgauge. Now we are back to what is 100% SoC.

 

I think it is fair to say that the Sgauge has rushed to 100% yet the batteries are not at 100% as demonstrated by the tail current of 15 amps approx, being of the order of 3.3% of the bank capacity. You recommend the aim should be for 1% of capacity (4.5A), you have also quoted 0.5A and seem to regularly quote 1 amp tail current as being the 100% point for your battery bank.

 

The graphs are very interesting and have shown that the slopes of the two gauges are different. I do not think they show nor prove any level of accuracy nor inaccuracy on the part of either box and surely cannot without a calibrated reference. They show only that the algorithms of the two gauges are different.

 

We have no calibrated reference of a higher accuracy than the gauges being tested to compare against. No hydrometer readings to actually check the state of charge within the battery or to use as a comparison.

 

Looking at the graphs, at the point the readings stopped they show a difference of 6% between the two gauges. What caused that difference; I would suggest we do not know. It is within the error allowed in the Sgauge specification and probably the same for the MSgauge expectations. You have mentioned that you normally expect a closer correlation between the two gauges nearer to 1 or 2%. So what happened this time.

 

I am not sure what all your hard work has told us. Sorry

 

ETA As for the remark, I assume about the MSgauge "...but it certainly seems unlikely to get close to 100%" As the spec on the Sgauge is +/-10%; the smartgauge manufacturer considers 100% +/-10% is acceptable i.e. between 90% and 110% when the reading should be 100%. So I think the remark is rot being we have two gauges with the batteries below 100% capacity reading 94% and 100%.being in mind that 6% of the batteries capacity is 27Ahs.

Edited January 27, 2016 by Graham.m

[nicknorman]

I said in the text that the tail current was 12A / 2.6% when SG got to 100% and yet you talk about 15A / 3.3%. Why is that?

 

Anyway as I've said before it boils down to what you want to consider is 100%. Yes I can get the tail current right down to 0.5% or less, after several days. But no-one wants to charge for several days each time. It thus becomes a compromise which is a personal choice. Victron says 4%, Mastervolt says 2% (that's the default value on the MS), you and I might say 1%. But none of these is the "right answer", they are personal choices.

 

Another thing to consider is the relationship between SoC and tail current. At these very small tail currents, hardly any AH remains to be added. Due to the asymptotic nature of the current curve I suggest the difference in SoC % between 2% and 1% tail current is minimal, perhaps 1% SoC or less. The point is that even if we had a super-precise way of measuring SoC it still isn't a good way to know what the tail current is nor, if tail current is the important parameter, when to stop charging, especially of this fictitious gauge only reads to 1% SoC resolution.

 

Then there is the question about whether it matters and if so how much. No-one has any data that says batteries charged to 1% tail current do any better than those charged to 2%. Or even 4%. Or if there is a difference, quantifying it in terms of battery life vs extra genny running time.

 

It all boils down to personal prejudice.

 

In the event when the SG read 100% it was perhaps 1 or 2% out. The MS was about 5% out the other way. Of course both the SG and the MS are in dead reckoning mode during charge, with the SG using parameters to guess the SoC that are inherently less accurate than those available to the MS. But during discharge, whilst the MS is still in dead reckoning mode, the SG has good references to work from, as can be seen on the earlier graph in the way it adjusted itself to the actual SoC (based on rested voltage) once the intermediate charge was stopped.

 

You asked why I appeared to be seeing a different relationship between the gauges on charge. I think it is simply because normally I fully recharge after each discharge, whereas in these tests there was a partial recharge. This seemed to throw the MS somewhat so it started from an erroneously low value, thus the SG was ahead of it more than usual. What I normally see is that, starting from similar values at say 70% SoC, the MS races ahead with the >100A charge current, but once the high 90s are approached the MS has slowed right down and the SG catches up, so that they both get near 100% at much the same time.

 

Finally I'd say that a digital readout of SoC is not a particularly good way to decide when to stop charging since the curve is so flat at that point. I'd say the same thing about specific gravity since this varies between cells, is temperature dependent, electrolyte level dependant and messy and error prone to measure. Not to mention inconvenient.

 

Tail current is easy to measure precisely and thus even if it's curve is fairly flat too, it is easy to determine a time when it falls to say 1% and stop charging. So it is a precise way of measuring something, but what is the significance of that something? Who knows, especially given the variance of people's ideas of what it should be.

 

Bottom line for me is that the SG gives a reasonable guide to when to stop charging that is probably as good in practice as any. And very easy easy for a numpty. Loads of people use it thus. Is there any evidence that the diligent chap who charges to 1% tail current has a better battery experience, or does he simply have a bigger genny fuel bill?

Edited January 27, 2016 by nicknorman

[Top cat]

All very interesting stuff and thanks for doing the graphs they are most informative. What im taking from all this is that the best way to determine when to stop charging is using the tail current. Whether to stop at 4, 0.5% or anywhere in between comes down to personal preference and circumstances.

In our case the final few Ah get put in by the solar panel for free and quietly so we typically get to 0.5% or less but on days when we dont get that I'm not overly bothered because we are likley to tomorrow.

 

So if I was advising a technically challenged boater who just wants to know when to stop charging. I would suggest a good ammeter and voltmeter plus a solar panel or two.

 

If I knew then what I know now I wouldn't have bought the BVM but got the separate meters and saved money.

 

Top Cat

[Graham.m]

I said in the text that the tail current was 12A / 2.6% when SG got to 100% and yet you talk about 15A / 3.3%. Why is that?

 

Anyway as I've said before it boils down to what you want to consider is 100%. Yes I can get the tail current right down to 0.5% or less, after several days. But no-one wants to charge for several days each time. It thus becomes a compromise which is a personal choice. Victron says 4%, Mastervolt says 2% (that's the default value on the MS), you and I might say 1%. But none of these is the "right answer", they are personal choices.

 

Another thing to consider is the relationship between SoC and tail current. At these very small tail currents, hardly any AH remains to be added. Due to the asymptotic nature of the current curve I suggest the difference in SoC % between 2% and 1% tail current is minimal, perhaps 1% SoC or less. The point is that even if we had a super-precise way of measuring SoC it still isn't a good way to know what the tail current is nor, if tail current is the important parameter, when to stop charging, especially of this fictitious gauge only reads to 1% SoC resolution.

 

Then there is the question about whether it matters and if so how much. No-one has any data that says batteries charged to 1% tail current do any better than those charged to 2%. Or even 4%. Or if there is a difference, quantifying it in terms of battery life vs extra genny running time.

 

It all boils down to personal prejudice.

 

In the event when the SG read 100% it was perhaps 1 or 2% out. The MS was about 5% out the other way. Of course both the SG and the MS are in dead reckoning mode during charge, with the SG using parameters to guess the SoC that are inherently less accurate than those available to the MS. But during discharge, whilst the MS is still in dead reckoning mode, the SG has good references to work from, as can be seen on the earlier graph in the way it adjusted itself to the actual SoC (based on rested voltage) once the intermediate charge was stopped.

 

You asked why I appeared to be seeing a different relationship between the gauges on charge. I think it is simply because normally I fully recharge after each discharge, whereas in these tests there was a partial recharge. This seemed to throw the MS somewhat so it started from an erroneously low value, thus the SG was ahead of it more than usual. What I normally see is that, starting from similar values at say 70% SoC, the MS races ahead with the >100A charge current, but once the high 90s are approached the MS has slowed right down and the SG catches up, so that they both get near 100% at much the same time.

 

Finally I'd say that a digital readout of SoC is not a particularly good way to decide when to stop charging since the curve is so flat at that point. I'd say the same thing about specific gravity since this varies between cells, is temperature dependent, electrolyte level dependant and messy and error prone to measure. Not to mention inconvenient.

 

Tail current is easy to measure precisely and thus even if it's curve is fairly flat too, it is easy to determine a time when it falls to say 1% and stop charging. So it is a precise way of measuring something, but what is the significance of that something? Who knows, especially given the variance of people's ideas of what it should be.

 

Bottom line for me is that the SG gives a reasonable guide to when to stop charging that is probably as good in practice as any. And very easy easy for a numpty. Loads of people use it thus. Is there any evidence that the diligent chap who charges to 1% tail current has a better battery experience, or does he simply have a bigger genny fuel bill?

 

I am sorry Nick, that is a pile of words supporting the sgauge. It does not deal with the failures of the experiment to compare two gauges without any reference. There is nothing to say that the sgauge was not reading high because it was not referenced to something else. If the MSgauge was the reference then the Sgauge read high, if Specific Gravity oh we don't have that. If tail current the sgauge read high.

 

Both boxes were within the ball park when the experiment was stopped, I suspect because the sgauge will not read any higher than 100%; whilst the batteries were under 100% charged measured by the parameter you seem to favour tail current at 1%.

 

As TC says an ammeter and a good voltmeter will do the job far more economically than the sgauge

Edited January 27, 2016 by Graham.m

[nicknorman]

I am sorry Nick, that is a pile of words supporting the sgauge. It does not deal with the failures of the experiment to compare two gauges without any reference. There is nothing to say that the sgauge was not reading high because it was not referenced to something else. If the MSgauge was the reference then the Sgauge read high, if Specific Gravity oh we don't have that. If tail current the sgauge read high.

 

Both boxes were within the ball park when the experiment was stopped, I suspect because the sgauge will not read any higher than 100%; whilst the batteries were under 100% charged measured by the parameter you seem to favour tail current at 1%.

 

As TC says an ammeter and a good voltmeter will do the job far more economically than the sgauge

 

Since the experiment didn't have any stated aim, I don't see how it can be a failure. I would say it was a success in that it shows the vagaries of the whole issue. It does however show the strengths and weaknesses of both types of gauge, neither is the panacea, so it depends on which types of faults you are prepared to tolerate. I do wonder if you spend enough thine reading what I have written. You say "if tail current, the SG read high" which is exactly what I already said, when the SG read 100% the actual SoC was probably 98 or 99% and at that moment it was a more accurate reflection of the actual SoC than the MS.

 

As I think I said, I didn't intentionally terminate the experiment, I had to leave the boat at that point, it was coincidence that the SG got to 100% just in time. I would have preferred to be able to continue to log the tail current and MS readings. Of course I left the batteries on charge and the telemetry now tells me the charger has gone to float.

 

Yes perhaps I do tend to focus on supporting the SG although I have highlighted the flaws in both gauges. Perhaps that is to compensate for your habit of highlighting any flaws in the SG whilst completely ignoring the flaws of an AH-counting gauge.

 

Anyway to sum up I'd say the "experiment" (it wasn't really an experiment, just a report) was a success as it shows there is no clear answer. If there were a clear answer we would have nothing left to talk about!

 

Regarding your final sentence yes there is no doubt in my mind that the best and coincidentally cheapest way to determine when to stop charging is an ammeter. However that only addresses one aspect of the need to monitor batteries, just as important is to know the actual SoC during discharge so as to avoid over-discharging (ie when to start charging) and also to have an idea of battery health. For that, the SG is the only player in the game for a numpty or anyone who wants an easy to live with solution.

Edited January 27, 2016 by nicknorman

[Graham.m]

Anyway to sum up I'd say the "experiment" (it wasn't really an experiment, just a report) was a success as it shows there is no clear answer. If there were a clear answer we would have nothing left to talk about!

 

 

So let us leave it there. As you say there is no perfect easy solution to the problem of battery management. Each has to read and try to make their own choice. It is such a pity that batteries are such an expensive disposable item in a boat's equipment. If they were the same order of cost as mooring lines I doubt this discussion would even have started.