Smartgauge w/ 3 batterys

[Dockhead]

Originally Posted by Exile

If I understand John's analysis correctly, charging sources misrepresent both VM and SG readings upwards but are quickly corrected (5-10 mins) once removed, whereas battery discharge/cycling misrepresent VM readings downward but take a long time to correct (24-48 hrs). So the SG can fill the gap by accurately showing SOC a short time after charging sources have been removed. In your case, you would either need an easy way to temporarily remove your solar & wind charging, or wait for a calm evening. In any event, it sounds like a SG has limited value if you are using it to determine when your batts. have finished recharging back to 100% as I suspect most probably do.

But again, this also means that an accurate VM reading of 12.2v or above (or whatever V your batt mfg specifies) with charging sources removed probably means a true voltage which is higher, i.e. greater than 50% (unless an aging batt's lower overall capacity influences??). So for those less concerned with precision & want to simply know when it's time to turn on their engine or genset to recharge their batts, then a VM -- along with periodic capacity tests and resetting your amp counter accordingly -- may suffice.

I'm sure John or someone else will correct if I got any of this wrong.

I think this is a very good analysis, and I generally agree with all of it, BUT for one small but significant point --

While it is true that you have to let a battery rest with no loads and no charging for a couple of days to get an IDEAL SOC reading based on voltage, that does NOT mean that what you get 5 minutes after loads are removed, or even when loads are present, is useless. Not at all!

The changes in voltage which happen over two days of resting are VERY small. In fact, some battery makers publish Voltage vs. SOC charts for LOADED condition (MaineSail has warned not to use them, though!).



The difference between system voltage at a low rate of discharge, like C/100, and true open circuit voltage, is small.

On my boat -- and I measured it, doing a series of SG tests over more than a year -- reading system voltage under light loads (and no surface charge, and no charging or heavy discharges in the last hour) against an open circuit voltage chart gave me a reading of SOC which was accurate within a 3 or 4 percent, which is much more accurate than I ever got from an amp-counting meter. What is more, the error is ALWAYS harmless -- that is, slightly understating the true SOC.


So in my experience, SmartGauge doesn't really "fill" any "gap". It's just much more convenient, because you just push a button instead of doing the temperature correction then reading the chart (don't forget to correct all voltage readings for temperature).

Very interestingly -- my SmartGauge seems to have the same error as my voltage method -- it seems to slightly understate the true SOC. But both methods far more accurate than amp counting, in my experience with Link and Victron amp-counting meters.


If we wanted a perfect battery monitor, surely it would use BOTH amps and volts and would learn from the interaction between the two. I wouldn't think this would be so hard to design, actually. Volts at a given rate of charge or discharge and with some analysis of recent history (and proper temperature correction), has got to give the most accurate picture we could possibly have of SOC, and would work equally well when charging or discharging.

If we actually need such an accurate picture -- but that's a different question.

[Dockhead]

Originally Posted by ramblinrod

. . . With an average cruising boat equipped with a solar charging system, having adequate bank size for normal loads, all one needs to know is:

1. If the batteries read 12.75 Vdc at sunset, they were fully charged. (The goal, life is good.)

2. If the batteries read more than 12.2 Vdc at sunrise, they were not depleted below 50%. (The goal, life is good.)

3. With a little experience, based on the morning voltage and the sunshine forecast, one can easily predict how much supplemental charging will be required.. . .

In my opinion, this is absolutely correct.

The proof of it is this -- how could the batteries have LESS than 50% charge, if the system voltage is more than 12.2V after not being charged for 12 hours?

Because of the limitations of this method, you might actually have more -- if the voltage has sagged because of loads (small bank, big loads). But you can't possibly have LESS. So you have achieved your goal of being sure that you aren't hurting your batteries.

Although I don't have solar, the job on my boat is much the same as this -- simply making sure I don't hurt my batteries by drawing them down too far or failing to charge in time.


How can the job be different on other boats? If you are finely managing your loads and trying to get through a period when you can't charge, or really don't want to charge. In this case, you want to see how you're doing so that you can adjust consumption if necessary.

Not my case, because it only takes one press of a button at my nav table to fire up my very quiet, heavy duty low speed generator. But for those trying to live entirely on solar, or who don't want to run their main engines unnecessarily, you might care that your voltage method may be underestimating your SOC. But I don't quite know what to advise such people. I have not observed that my SmartGauge is any more accurate than informed voltage reading (but might be on other boats -- I don't know). Maybe BOTH voltage AND amp counting -- to give two different points of view.

[conachair]

Originally Posted by Dockhead

On my boat -- and I measured it, doing a series of SG tests over more than a year -- reading system voltage under light loads (and no surface charge, and no charging or heavy discharges in the last hour) against an open circuit voltage chart gave me a reading of SOC which was accurate within a 3 or 4 percent,

You need a lot more than an hour with trojans.....

This is 225Ah , 2 x T05s with load varying between about 2A & 5A with the fridge cycling.Accurate voltage straight from the battery terminals.
Sunset 18.30 (utc) Temperature low to mid twenties degC.




How did you get 3/4% accurate SOC reading? Doubt if I could I get that accurate with a hydrometer, with some variance between the cells added.

[Dockhead]

Originally Posted by conachair

You need a lot more than an hour with trojans.....

This is 225Ah , 2 x T05s with load varying between about 2A & 5A with the fridge cycling.Accurate voltage straight from the battery terminals.
Sunset 18.30 (utc) Temperature low to mid twenties degC.

Fair enough --

and worth a big caveat -- you MUST have the surface charge off before voltage means anything. How long it takes to get off varies a lot. I think I mentioned it, but it's worth emphasizing.

The good thing, however, is that you practically never have much of a surface charge, when the batteries are way down when you are really needing to know where you are. If you charge hard when the batteries are way down and interrupt the charging before you get far, I have found that the surface charge goes away fast. If you're 90% or 100% charged, then obviously it's different, and with light loads can take hours.

Originally Posted by conachair

How did you get 3/4% accurate SOC reading? Doubt if I could I get that accurate with a hydrometer, with some variance between the cells added.

I am assuming that SG corrected for temperature, using a good hydrometer and on unstratified electrolyte IS the true SOC. So what I am talking about is correspondence between different SOC readings obtained by other means, and this.

And do we have any more accurate way to know SOC, than SG? A careful SG reading IS the chemical SOC of the battery, I think. Granted, the chemical state of the battery may not perfectly represent amp/hours of power to dead -- but you see, we run up against the limits of definitions and of our knowledge, and I just don't think we care so much, once we get this far. In my opinion, the chemical state of charge is more fundamental to what is actually happening inside the battery, than the exact amount of power we can still get out of it (which varies so much according to other factors anyway). So I think we are in "angels on the head of a pin" territory.

[conachair]

Originally Posted by Dockhead

-- but you see, we run up against the limits of definitions and of our knowledge, and I just don't think we care so much, once we get this far. In my opinion, the chemical state of charge is more fundamental to what is actually happening inside the battery, than the exact amount of power we can still get out of it (which varies so much according to other factors anyway). So I think we are in "angels on the head of a pin" territory.

Personally I really don't care that much , as long as they get to 100% most days .
This i would disagree with, based on logging trojans anyway -
"1. If the batteries read 12.75 Vdc at sunset, they were fully charged. (The goal, life is good.)"
Probably close but almost certainly not all the way to full unless it's been bouncing around u there for a few hours.
This based on watching amps in & occasionally checking against S.G. After a while you get a feel for when the current will finally tail off & they are happily charged.

One more thing, logging is wonderful, and not that expensive. Humans love to see what they want to see, having data logged to display graphed is very powerful for spotting trends - you too could have it for the cost of a solent pint and a bag of chips with your new computer

[ramblinrod]

Originally Posted by conachair

You need a lot more than an hour with trojans.....

This is 225Ah , 2 x T05s with load varying between about 2A & 5A with the fridge cycling.Accurate voltage straight from the battery terminals.
Sunset 18.30 (utc) Temperature low to mid twenties degC.




How did you get 3/4% accurate SOC reading? Doubt if I could I get that accurate with a hydrometer, with some variance between the cells added.

Thanks, this is excellent data to work with.

So to me it looks like:

1. Your charger was in absorption mode (~14.7 Vdc) until 1530

2. At 1530 it started switching in and out of float mode (~13.5 Vdc).

3. Around 1615 an event occurred, most likely cloud or panel shadowing.

4. Around 1700 the sun came back, but intensity was too low to achieve absorption again.

5. At 1800, (about 1/2 hour before sunset) the charge voltage was effectively removed.

6. The first refrigerator cycle, with no charge voltage applied, brought the voltage down to 13.0 Vdc at around.

7. The sunset (per your statement) was at 1820.

8. The first fridge on cycle (around sunset) brought it down to just under 12.9 Vdc.

9. It is around this point where the terminal voltage decline rate decreased substantially.

So this pretty much confirms what I stated:

1. After the removal of the charge voltage (sunset) and application of a light load (fridge, which in your case represents C 0.02), the residual charge voltage is depleted quite quickly.

2. Your batteries were charged full. Else, the terminal voltage would have dropped to a lower level around sunset.

What I also get from this curve is that...

3. You started listening to some music around 2000, and likely had a beverage.

4. You went to the washroom at 2120 and brushed your teeth.

5. You retired for the evening around 2130.

6. You got up briefly around 0030.

7. You have a manual pump head.

8. Might wanna stop drinking tea or beer before bed, or get your blood sugar and prostate checked.

9. At sunrise, I suspect your battery voltage was around 12.38 Vdc. Life is good, well above 50% SOC.

10. You might consider increasing your fridge insulation (or improving the door seal) to reduce the cycling from every 15 minutes to at least 1/2 hour, especially during the night. Then you could reduce the resistor between "C" and "T" of the controller (if currently present) to extend the fridge run time but at a lower current draw.

Thank your for sharing your voltage curve; it pretty much confirms and summarizes what I have been stating; that by simply monitoring battery voltage at sunset and sunrise, you have enough data to manage your battery bank under normal circumstances.

A Smart Gauge would not give a whole lot more useful information than the voltmeter. Your system gives a lot of information. How useful it is, is likely in the eyes of the beholder. Certainly fun for this exercise. Again thanks.

[ramblinrod]

Originally Posted by conachair

Personally I really don't care that much , as long as they get to 100% most days .
This i would disagree with, based on logging trojans anyway -
"1. If the batteries read 12.75 Vdc at sunset, they were fully charged. (The goal, life is good.)"
Probably close but almost certainly not all the way to full unless it's been bouncing around u there for a few hours.
This based on watching amps in & occasionally checking against S.G. After a while you get a feel for when the current will finally tail off & they are happily charged.

One more thing, logging is wonderful, and not that expensive. Humans love to see what they want to see, having data logged to display graphed is very powerful for spotting trends - you too could have it for the cost of a solent pint and a bag of chips with your new computer

Your 100% charge voltage reading (around 13.0 Vdc at 1800) is higher than I would expect and normally see on any system. Your absorption voltage and float voltages are about 0.1 to 0.2 Vdc higher as well.

A friend (actually the one who advised the Smart Gauge was not accurate at all when charging) discussed developing a raspberry pi based battery monitoring system and I actually went to the effort to draw up a tech spec of what it should monitor, analyse, and report / alarm. I very quickly learned that I would spend way more time developing the system and playing with it, than I did sailing (as confirmed by the time I spent looking at and pondering your single, brief data log), so I scrapped the idea.

Oh, one more thing, if your boat is in the water, your stuffing box(es) are likely fine. ;-)

[Dockhead]

Originally Posted by conachair

. .. One more thing, logging is wonderful, and not that expensive. Humans love to see what they want to see, having data logged to display graphed is very powerful for spotting trends - you too could have it for the cost of a solent pint and a bag of chips with your new computer

Yes, I think that even for me -- who does not spend much time worrying about electrical power -- the logs would be useful. I think you can kind of "grok" how your batteries are living, by looking at those curves. I like it.

What's the best way to get voltage data into the computer? I've seen USB data loggers that would do it, but they cost a lot more than a pint at the Jolly Sailor.

[Exile]

Originally Posted by Dockhead

If we actually need such an accurate picture -- but that's a different question.

As a layman maybe I missed something (corrections always appreciated), but I thought this was the entire point of RamblinRod's advice all along. Not that the SG is inaccurate, faulty, or unhelpful, but that it may be unnecessarily redundant for most of his customers. If you want or feel you need more precision then great, but I can't see how a VM reading of 12.2v (or whatever your batt. mfg. specs) with light loads applied can possibly indicate anything other than a true SoC somewhere at or above 50%. For most, this is the most important thing you need to know, right?

One potential disadvantage is I keep reading posters with newly-installed SG's and solar/wind claiming that their batts. have reached "100%". Have they really, or are they looking at an inflated reading due to their solar/wind charging sources? It seems like an amp counter is most useful on this upper end, i.e. to see whether the batteries are only accepting amps at or below .05C (or whatever your battery mfg. specifies). My battery charger has a read-out which shows amps going into the bank, so once it's on float I can simply monitor that (or my Victron BM). Is there anything misleading that could shorten the life of my batts. with this approach?

Having said that, I could very well wind up installing a SG once I go long-distance cruising, if nothing else than to have an add'l and easier way to monitor my batts. when I'm away from shore power for extended periods of time. But I wouldn't be doing so because it's a "necessity," at least in my mind. In the meantime, I'm still at a loss to figure why RR's advice became so "controversial." Compared to VM's & amp counters, the SG is a fairly recent arrival on the scene, right? Has everyone who understands how to properly read VM's been unduly shortening the life of their batts. before the SG came along?

[ramblinrod]

Originally Posted by Exile

As a layman maybe I missed something (corrections always appreciated), but I thought this was the entire point of RamblinRod's advice all along. Not that the SG is inaccurate, faulty, or unhelpful, but that it may be unnecessarily redundant for most of his customers. If you want or feel you need more precision then great, but I can't see how a VM reading of 12.2v (or whatever your batt. mfg. specs) with light loads applied can possibly indicate anything other than a true SoC somewhere at or above 50%. For most, this is the most important thing you need to know, right?

One potential disadvantage is I keep reading posters with newly-installed SG's and solar/wind claiming that their batts. have reached "100%". Have they really, or are they looking at an inflated reading due to their solar/wind charging sources? It seems like an amp counter is most useful on this upper end, i.e. to see whether the batteries are only accepting amps at or below .05C (or whatever your battery mfg. specifies). My battery charger has a read-out which shows amps going into the bank, so once it's on float I can simply monitor that (or my Victron BM). Is there anything misleading that could shorten the life of my batts. with this approach?

Having said that, I could very well wind up installing a SG once I go long-distance cruising, if nothing else than to have an add'l and easier way to monitor my batts. when I'm away from shore power for extended periods of time. But I wouldn't be doing so because it's a "necessity," at least in my mind. In the meantime, I'm still at a loss to figure why RR's advice became so "controversial." Compared to VM's & amp counters, the SG is a fairly recent arrival on the scene, right? Has everyone who understands how to properly read VM's been unduly shortening the life of their batts. before the SG came along?

Excellent point worth re-emphasizing.

If your battery voltage is 12.8 Vdc, and your current going into them from the charger is more than about 1 A per 100 A-hrs of bank size, above the load current being drawn, your batteries are not fully charged, you are likely seeing charge voltage on partial SOC batteries.

How can one easily avoid being fooled by this?

Check your battery voltage at (or shortly after) sunset after solar charge voltage has been removed, and applied loads have depleted residual charge voltage.

Neither a DVM or SG will give accurate SOC while sun is on the panels

[Dockhead]

Originally Posted by Exile

As a layman maybe I missed something (corrections always appreciated), but I thought this was the entire point of RamblinRod's advice all along. Not that the SG is inaccurate, faulty, or unhelpful, but that it may be unnecessarily redundant for most of his customers. If you want or feel you need more precision then great, but I can't see how a VM reading of 12.2v (or whatever your batt. mfg. specs) with light loads applied can possibly indicate anything other than a true SoC somewhere at or above 50%. For most, this is the most important thing you need to know, right?

One potential disadvantage is I keep reading posters with newly-installed SG's and solar/wind claiming that their batts. have reached "100%". Have they really, or are they looking at an inflated reading due to their solar/wind charging sources? It seems like an amp counter is most useful on this upper end, i.e. to see whether the batteries are only accepting amps at or below .05C (or whatever your battery mfg. specifies). My battery charger has a read-out which shows amps going into the bank, so once it's on float I can simply monitor that (or my Victron BM). Is there anything misleading that could shorten the life of my batts. with this approach?

Having said that, I could very well wind up installing a SG once I go long-distance cruising, if nothing else than to have an add'l and easier way to monitor my batts. when I'm away from shore power for extended periods of time. But I wouldn't be doing so because it's a "necessity," at least in my mind. In the meantime, I'm still at a loss to figure why RR's advice became so "controversial." Compared to VM's & amp counters, the SG is a fairly recent arrival on the scene, right? Has everyone who understands how to properly read VM's been unduly shortening the life of their batts. before the SG came along?

I agree completely.

I think most battery abuse comes from (a) people who just don't care until the lights flicker, no matter what kind of technique they use (or don't use) to gauge SOC; and (b) people who overly trust amp-counting meters, particularly those who don't even recalibrate them. I was cruising for many years before the first battery monitor was invented, and we didn't fret about the batteries much, and our batteries lasted a long time, even before we had solar. In a way, battery monitors are a solution to a somewhat non-existent problem, in my opinion.

As to judging a 100% charge -- if you care about this (and why would you, really, other than idle curiosity?), acceptance rate will tell you this, if you have an ammeter. I forgot to mention this when I was telling my view of the pluses and minuses of SmartGauge, but my own SmartGauge seems to get this more or less right. It will flick over to 100% after high voltage for a certain period of time, or at least it seems it does that, but usually it seems to get that about right. A partial charge -- to 80% or so as I usually do on generator power -- it doesn't understand at all.

[daletournier]

Originally Posted by Exile

As a layman maybe I missed something (corrections always appreciated), but I thought this was the entire point of RamblinRod's advice all along. Not that the SG is inaccurate, faulty, or unhelpful, but that it may be unnecessarily redundant for most of his customers. If you want or feel you need more precision then great, but I can't see how a VM reading of 12.2v (or whatever your batt. mfg. specs) with light loads applied can possibly indicate anything other than a true SoC somewhere at or above 50%. For most, this is the most important thing you need to know, right?

One potential disadvantage is I keep reading posters with newly-installed SG's and solar/wind claiming that their batts. have reached "100%". Have they really, or are they looking at an inflated reading due to their solar/wind charging sources? It seems like an amp counter is most useful on this upper end, i.e. to see whether the batteries are only accepting amps at or below .05C (or whatever your battery mfg. specifies). My battery charger has a read-out which shows amps going into the bank, so once it's on float I can simply monitor that (or my Victron BM). Is there anything misleading that could shorten the life of my batts. with this approach?

Having said that, I could very well wind up installing a SG once I go long-distance cruising, if nothing else than to have an add'l and easier way to monitor my batts. when I'm away from shore power for extended periods of time. But I wouldn't be doing so because it's a "necessity," at least in my mind. In the meantime, I'm still at a loss to figure why RR's advice became so "controversial." Compared to VM's & amp counters, the SG is a fairly recent arrival on the scene, right? Has everyone who understands how to properly read VM's been unduly shortening the life of their batts. before the SG came along?

14 pages and youve nailed it, not that hard, not really sure how it can get so heated.

[Exile]

Originally Posted by Dockhead

As to judging a 100% charge -- if you care about this (and why would you, really, other than idle curiosity?), acceptance rate will tell you this, if you have an ammeter.

I thought that, whenever practicable, getting your batts. charged back to 100% (as opposed to 80% or even 97% in the opinion of some) does have an appreciable effect on overall life. Not as consequential as avoiding discharges below 50% perhaps, but enough to make a difference. But this is only based on what I've read . . . .

I forgot to mention this when I was telling my view of the pluses and minuses of SmartGauge, but my own SmartGauge seems to get this more or less right. It will flick over to 100% after high voltage for a certain period of time, or at least it seems it does that, but usually it seems to get that about right. A partial charge -- to 80% or so as I usually do on generator power -- it doesn't understand at all.

If the SG can get the 100% SoC right then this would be a significant plus for those w/o an ammeter, and can provide redundancy & convenience for those who do. Assuming no ammeter, I wonder if there's a way to calculate how long the SG needs to show 100% for that SoC reading to be accurate?

[Dockhead]

Originally Posted by Exile

If the SG can get the 100% SoC right then this would be a significant plus for those w/o an ammeter, and can provide redundancy & convenience for those who do. Assuming no ammeter, I wonder if there's a way to calculate how long the SG needs to show 100% for that SoC reading to be accurate?

The reason why I don't think it matters for most of us is that getting the batts to 100% takes a long, long time. It only happens on my boat if I'm on shorepower overnight (or if I motored for hours). So it's REAL easy for me to know when I got to 100% -- Shore power overnight = I got there! No shorepower overnight = I didn't. No electronics needed.

I guess if you have a large solar installation the question might be more complex, but still -- why do you really care? It's not like you can leave the sun running an extra hour. It is what it is. If after the surface charge is off, you have 25.2V or so (12.6v), then you know you either got to 100%, or you got pretty close. Do you really need to know more than that?

[Exile]

Originally Posted by daletournier

14 pages and youve nailed it, not that hard, not really sure how it can get so heated.

Fwiw, I notice some of these techy threads getting heated due to what I think is a communications gap btwn. those with a high level of technical knowledge who's interest & enthusiasm with the topic far exceeds those of the layman cruiser. By contrast, the latter often views learning how to maintain their boat systems as more of a means to an end, with each system viewed more like "tools" to achieve the overall goal of a more reliable & less costly boat.

In this case, I think RR and Dockhead have done an excellent job bridging that gap, at least for this layman.