Conductance test

[a64pilot]

I have not figured my smart gauge out, it seems accurate most of the time, but when I am sure I am at 100% SOC and go off of shore power, very quickly as in an hour or so it will indicate 90% SOC, from there it seems to drop way too slow and the SOC again seems accurate at lower charge levels, it's that real fast drop to 90 and hanging there or high 80's for a long time that I can't figure out.

[btrayfors]

Not sure about your SmartGuage, but it's basically a voltmeter. If you track the voltage at a FLA battery which is 100% charged then has a modest load attached, you'll find a drop pretty quickly to lower-than-expected voltage.

Then, within minutes, the voltage will begin to climb and will pretty soon reach "expected" levels. I suppose this is after the battery "wakes up" and the chemical reactions set in.

Look at the red and blue curves in the attached graph: these are flooded Trojan deep-cycle batteries. The other two are gelled batteries which do not show the same behavior.

[john61ct]

Originally Posted by btrayfors

Haven't had time to puzzle that out, but one thing it wouldn't do is to give you the actual capacity of the battery unless Peukert were accounted for. If the applied load were much less than C/20, then you'd get an over estimate of the total capacity of the battery. Similarly, if the load were much more than C/20 you'd get an underestimate.

And, all that is assuming the 80% SOC is correct.

That assumption was a given, as I said there have been many debates over the device's accuracy. For me wrt Mr Peukert, since Gibbo's in-depth exploration of the topic is the best I've read

http://www.smartgauge.co.uk/peukert_depth.html

I have no doubt he took it into account in the device's algorithms.


I am not trying to claim my quick-and-dirty idea would be as authoritative as a rigorous 20-hour-rate test.

As long as the math / logic of what I outlined holds, and for those that know they can trust the SmartGauge, I bet it would be a decent alternative that could be applied in the field more easily.

[john61ct]

Originally Posted by a64pilot

I have not figured my smart gauge out, it seems accurate most of the time, but when I am sure I am at 100% SOC and go off of shore power, very quickly as in an hour or so it will indicate 90% SOC, from there it seems to drop way too slow and the SOC again seems accurate at lower charge levels, it's that real fast drop to 90 and hanging there or high 80's for a long time that I can't figure out.

It is not supposed to be real-time accurate during changing state conditions like that, can be off by 5+%.

Even under any charge source the accuracy confidence is lowered. Stop charging the bank, after it has been at rest for a few hours, then confidence of that snapshot is within 2-3%.

[john61ct]

Originally Posted by btrayfors

Not sure about your SmartGuage, but it's basically a voltmeter.

No. Don't disparage something you obviously haven't even read much about.

Voltage is not an accurate gauge of SoC.

As the true AH capacity of your bank declines, how are you re-rating your correspondence tables?

[btrayfors]

Originally Posted by john61ct

No. Don't disparage something you obviously haven't even read much about.

Voltage is not an accurate gauge of SoC.

As the true AH capacity of your bank declines, how are you re-rating your correspondence tables?

Perhaps the learned gentleman could explain how, with just a two-wire connection (one to the positive battery post and one to the negative battery post) the SmartGuage could measure anything BUT voltage.

Of course it's a voltmeter. And, additionally, it has built-in smarts to infer/discern/approximate/estimate battery capacity as derived from voltage changes over time, given some pre-programmed stats about the battery (chemistry, amp-hour rating, Peukert, etc.).

And, unlike most other battery monitors, it actually improves its accuracy over time.

How, then, can you state that SoC cannot be derived from voltage, when it's exactly that --- battery voltage -- which drives the calculations in the SmartGuage?

As it turns out, resting voltage is actually a pretty good indication of SoC. But it's not a direct indication of the residual capacity of a battery. That takes some calculation and additional inputs over time.

Far from disparaging the SmartGuage I think it's a brilliant device which with a very simple hookup delivers pretty accurate information and avoids many of the drawbacks of most other battery monitors.

Bill

[boatbod]

If you've got connections to both battery terminals you can pass high frequency signals/pulses between the leads and measure stuff like conductance as well as voltage.

[sailinglegend]

Originally Posted by btrayfors

Perhaps the learned gentleman could explain how, with just a two-wire connection (one to the positive battery post and one to the negative battery post) the SmartGuage could measure anything BUT voltage.

Of course it's a voltmeter....Bill

Chris Gibson's SmartGauge comments

Full post is at,

Smartbank battery management - Page 2
post number #42

"You have to remember that not everyone understands volts, amps, amp hours, Peukert, charge efficiency etc. And for those that don't, an amp hours counter will not do what they require of a battery monitor. i.e. it will not tell them the state of charge of their batteries.
Obviously I am not prepared to disclose how SmartGauge works. Anyone who asks me to do so is simply being totally unreasonable.
But to say that only voltage can be measured via 2 wires is *completely incorrect....
Pull a brief current pulse from the battery and measure the voltage drop, this will give an indication of internal resistance.
Present an AC voltage across the battery and measure the phase angle and amplitude of the resultant current. This will show the AC impedance of the battery.
Do the same thing with a wide variety of frequencies and analyse the results. This is know as AC impedance spectrography."

[john61ct]

Originally Posted by btrayfors

Perhaps the learned gentleman could explain how, with just a two-wire connection (one to the positive battery post and one to the negative battery post) the SmartGuage could measure anything BUT voltage.

How should ** I ** know?

Gibbo's kept it a secret through dozens of threads with very technical people.

Yes it's frustrating, but way things are with the reality of IP enforcement in China, can you blame him?

If you can understand this stuff (it's way beyond me), and really want to try to figure it out, here are my best-guess starting points:

AC impedance spectrography, phase angles

​impedance, internal resistance

pattern recognition engine, neural net learning (AI) against a stored database of known chemistry behaviours (remember no good with LFP)

-----
But the fact few know (could even understand) the "how" is irrelevant to the question of "does it work?"

Google yourself, ​EnerSys' white paper, makers of Odyssey batts

Maine Sail's thorough test report, he was also very skeptical.

​tested against a Vanner Battery Monitor (around £10'000 worth - designed specifically for testing batteries), a Link 10, BEP DCM600 and a Victron BMV. We tested it under various types of load, battery sizes, types etc. After 3-4 cycles (this is needed for the SmartGauge to learn about the battery - and the key to a shuntless design), we found that the SmartGauge was (overall) more accurate than the shunt based meters


It's very widely used in military systems, in fact reading between the lines after Merlin picked up the line that sort of consulting work might have become a lot more important than this little recreational / marine micro market.

​http://www.ybw.com/forums/archive/in.../t-291026.html

This one from early days is pretty funny: http://www.ybw.com/forums/showthread...nagement/page5

OK, you made me do a lot of work here, NOW could you come to a more informed conclusion and post something more appropriate than "it's just a voltmeter"??

Please?

[Paul L]

So if the SmartGauge has all these complex functions to measure charge state, how come it doesn't work while charging? Something that is going on all day on typical cruising boat with solar and maybe at night with wind.

[a64pilot]

Originally Posted by Paul L

So if the SmartGauge has all these complex functions to measure charge state, how come it doesn't work while charging? Something that is going on all day on typical cruising boat with solar and maybe at night with wind.

It does seem to work while charging, at least it will show an increase in SOC as my bank charges, however I only look at acceptance rate to determine full charge, it will prematurely indicate 100% charge, but likely by only a couple of percent. It's the last couple of percent that seem to take more time than you would think.

[Paul L]

Originally Posted by a64pilot

It does seem to work while charging, at least it will show an increase in SOC as my bank charges, however I only look at acceptance rate to determine full charge, it will prematurely indicate 100% charge, but likely by only a couple of percent. It's the last couple of percent that seem to take more time than you would think.

The manual says it doesn't, if I am reading Note 2 correctly.

[john61ct]

You guys talk like this stuff is easy.

The algorithm input factors benchmarked are more in flux while current is flowing than when at rest.

And it's just a bit less accurate, not pretty much useless like readings from most coulomb-counters out there.

I'd love to hear of a better reasonably priced SoC monitor, really, anything proven out there?

Apparently some of the high-end integrated LiFePO4 (LFP) systems auto-reset their calibration when BMS says 100%, but never heard of ANY other monitor tracking diminishing AH capacity as it walks down over time. But Victron and Redarc make nice stuff, just unlikely it works with anyone else's.

But yes, it isn't perfect.

[Paul L]

Originally Posted by john61ct

......

And it's just a bit less accurate, not pretty much useless like readings from most coulomb-counters out there.

I'd love to hear of a better reasonably priced SoC monitor, really, anything proven out there?

......

But yes, it isn't perfect.

I guess it depends on what you are trying to accomplish. As a full-time cruiser with a decent sized solar setup there are really only two significant power decisions that I make on a daily basis. Will I supplement with the genset and will I use excess solar to make water.
Both of my coulomb-counters accurately give me the information I need to make those decisions at anytime of day. I don't see that the Smartgauge would give me more accurate info for those decisions and it could offer less accurate info at the times I want it.

[john61ct]

Sure, I'm not even claiming everyone needs accuracy, and as I said am actively seeking evidence of other tools that are accurate.

How do you verify accuracy in SoC readings?

How often do you reset / recalibrate ?

What is that process like?

Do they take into account or do you need to determine input AH capacity yourself?

I can really see the value of logging / benchmarking Amps in/out, SG doesn't give that.