I've been using this device for several months of continuous cruising now.
I am still convinced that amp-counting
battery monitors are practically useless (I removed, then finally
sold my
Victron one). That's not to say that they are not capable of giving any useful information at all, but by the time you keep them in proper calibration and interpret them properly you will have expended a lot more effort to get a lot less precise information, than you can get by other means.
The SmartGauge is widely praised as a superior approach to
battery monitoring, and is used by military and
commercial operations, has been thoroughly tested by different people including our own MaineSail, all with the results that this device gives a much better picture of the state of your
batteries.
This all may be true, HOWEVER -- my experience has been that I am getting practically zero information from the SmartGauge, which I was not getting from simply monitoring system voltage and referring to a simple table taken from the Trojan
Batteries site.
I used system voltage plus the table for several years, and checked the accuracy several times by taking specific gravity measurements.
My method for using system voltage is very simple -- you don't try to do Open Circuit Voltage, after resting the batteries for 12 hours etc., which is the technically correct way to do it. That's because this is completely impractical on a cruising
boat which is in actual use. What I do is to read system voltage when there are no big loads on, and no
charging has taken place for some time. The idea is that small loads will not pull down the voltage that much, AND any error will understate the battery state of charge, which is a harmless error. I decided that this would
work ok because I don't actually need more precision than 10% or so.
Specific gravity tests showed that the method is actually far more accurate than that -- more like 2% or 3%.
I was delighted with this, as it was far simpler than interpreting an amp-counting meter, and required nothing more than a
cheap voltmeter (but carefully wired directly to the batts with large cross section cables).
The only problem was that this method cannot tell you how much
charging you have accomplished, during a charging run. If you charge for a while on
generator and stop before the
charger goes into float mode, you just don't know how far you got until some hours later when the surface charge is off the batts.
So I thought the Smart Gauge might be the answer to this. I bought one last year and finally got around to installing it over the
winter.
Well, the Smart Gauge is not an answer to this. It does attempt to estimate how far you have gotten during a partial charging run, but it is wildly inaccurate. It settles down and gives you a true reading only -- hah, after several hours, after the surface charge is off the batts.
So in summary -- it's a good device, undoubtedly more useful and far more accurate than amp-counting gauges, but I have not discovered any way in which it is significantly better than simply monitoring system voltage.
Now I am sure that this can be overcome with further development. Voltage rises throughout the
absorption bulk phase. This must correspond to state of charge achieved. A really accurate picture could surely be achieved by mapping charging
current vs voltage -- no? Then the device could time the absorption phase and estimate the state of charge achieved per minute of absorption charging by "learning". So maybe SmartGauge v2.0 could have a Hall-effect ammeter to measure charging
current, and could use this data to calculate state of charge achieved during charging.
Of course it may be that we don't actually need that much information. The main thing we need to know is when to charge. We probably don't need to know that with very great accuracy. But we would not want an error on the wrong side -- where we have an exaggerated idea of the state of charge of the batts, so we charge too late, because we were not informed about how deeply discharged the batts had become.
So for this, actually, simple voltage monitoring fits the bill exactly. The state of charge can be understated if you read the voltage with too much load on, but CANNOT be exaggerated. And it's actually remarkably accurate, about 2% to 3% according to my tests.
I still have my old table from Trojan, on my instrument panel. It's amusing to see that the battery state of charge read from this table, always corresponds EXACTLY to what the SmartGauge is telling me.