I'm checking out the Costco GC2 batteries and Iota 55A converter as well as the new 16-bit a/d for Arduino. Prior to discharging the 208AH GC2 batteries, they were top charged with 16V for about an hour.They are then discharge to about 50% (actually 99.8AH) with a load current of 14A for 7 hours. The load current consists of a 100W light connected to a msw inverter and some lighting inside the MH. Data is acquired using an Arduino Uno board, and a 16-bit a/d converter. Battery voltage and current (measured across a 100A shunt) is recorded every second.The first graph shows battery voltage (in blue) and current (in pink) over the entire 7 hour discharge period. Voltage scale is on the left y-axis and current scale is on the right y-axis. Current is defined as negative when going out of the battery. Top of the graph is 0A and the bottom is -16A. The current oscillates due to inverter operation. It does not pull a constant current.The second graph is same as first, but just looking at the first 30 min of discharge.Like wise the third graph shows the first 5 min of discharge. The inverter is switched on at about the 3/4 minute mark to get -8A, and a little later the add lights were turned on to drop down to -14A. Battery voltage is naturally decaying (from time 0 to 3/4 min) because the 16V charging power supply was just removed. What's surprising is that there are two different voltage discharge slopes. The first slope is fast, going from 14V to 12.4V in about 6 minutes. This may be the top charge. The second slope is very gradual, going from 12.4V to 11.8V in about 7 hours. The next graph shows SOC with respect to battery voltage. The weird looking top I believe is due to the top charge. The last graph is same as above, but the voltage is corrected for a no-load current. The internal resistance of the battery is 12 mohm. The load current multiplied by the resistance is added to the voltage measurement. I also started the plot before the weird looking top comes into play.And here's the test set-up. The 100A shunt is connected directly the the negative battery post. Charge plots still to come.

The 55A Iota IQ4 converter is in boost for 80 minutes. SOC at that point is 87%. At the 65 min mark the battery is at 14.5V and the converter at 14.8V. This triggers a timer within the Iota and it remains in boost mode for an additional 15 min, or to the 80 min mark.Current slopes downward at 65 min because the Iota output voltage reached the limit of 14.8V. In other words, tapering begins.In boost, battery voltage rises fairly linearly, until about 50 min into charging. Voltage begins to increase faster an that point. That's due to the increase in battery resistance starting around 75% SOC.The right side y-axis scale is for current and SOC.Another interesting point is when converter voltage drops from 14.8V to 14.2V. Battery surface charge and IR drop reduces, causing it to accept about 8A more current. In other words, dropping from 14.8V to 14.2V is not that huge.

Hi Salvo,Thanks for taking the time to post this information.

Very interesting graphs. (Not at all ugly either! :) )I do wonder about the charging graph's SOC plot though. It seems to be showing about 5% too high an SOC at that 65 minute point, compared with what I have seen with my recharges. Perhaps there is a different allowance for heat loss?I also can't reconcile your battery acceptance rate of 20 amps at about 95 minutes at "90%" SOC at 14.2 volts with my "marker" of 5 amps per 110AH at 14.5v, which would be 9.5 amps for 208AH, and less than that at a lower voltage (14.2 vs 14.5)That would also make your 50-90 take about an hour and a half, which is about a half hour too soon, based on what I have seen.So either the SOC plot there has something wrong with it, or there is more to be learned here about this. (My "marker" works out right cross-checking with the Trimetric AH counter, so it is not "wrong"---the thing to figure out here is what is making the difference)"The AH returned" plot on this graph is linear, which would keep the SOC percentage lower than the SOC plot in the above graph.http://www.engineersedge.com/battery/specific_gravity_battery.htmFor sure the drop from 14.6 battery to 14.2 has less of a drop in amps than dropping from 14.6 to 13.6 does. 13.6 means twice the amps drop.

It's hard to say there's something wrong with my methodology. I measure current every second. During discharge, I have almost 26,000 measurements. All 26,000 current measurements are added together and divided by 60 sec/min * 60 min/hr. The total is 100AH removed during discharge. During charge, my SOC starting point is 108 AH / 208 AH = 51.9%When charging, each one-second measurement adds to the initial 108 AH capacity. There are no losses involved. In other words, when in bulk, if 10A are charging the battery for one hour then battery capacity increases by 10 AH.

Nope.The data acquisition is in my humble opinion, flawless. Two ears and a tail.My inability to produce graphing here left me frustrated in showing the relevance of my Vmax charging regimen and value of 2nd stage 14.0 vdc charging with generator. Your graph illustrates this point beautifully.Observation and reaction to trends & tendencies data are worth their weight in gold. All my A to D data disappeared when I was hospitalized 26-years ago. It was fun. Configuring a reactive charging regimen to the data was an obsession. Brings back fond memories.What can be done for many hundreds of dollars these days, took many tens of thousands of dollars and hundreds of hours of untallied labor back then. But it was worth it.

Some cool battery graphs | Good Sam Community

grizzzman
Explorer
Jan 02, 2016
Salvo wrote:
I don't know how much electronic knowledge you have. Just a cautionary note. The a/d has a 4V supply voltage. That means you can not measure voltages greater than 4V. In order to measure battery voltage you first use a resistor divider to reduce voltage to a usable range for the a/d and then multiply the a/d output by a specific factor to get the actual voltage.

You don't need any resistor divider to measure shunt voltage. This voltage is down in the millivolts. In order to measure both charge and discharge currents you need two a/d channels, operating in differential mode. The a/d you're getting has 4 channels. One channel is used to measure battery voltage and 2 channels are used to measure current.

I use a separate 9.6V Ni-CD battery (and 5.3V dc/dc converter) to power Arduino. This particular battery was used in a remote control car for grandson. I think you will get a measurement error if you connect Arduino directly to house battery. You could get an isolated dc/dc converter to get power directly from house battery.

The first step with Arduino is to write code to make a led blink. It takes a few baby steps before getting to the stage of measuring and storing data.

grizzzman wrote:
Ordered 2 temp sensors and the screen. I want live and stored data.

Thanks Salvo for the information. The hardware side is no big deal for me. I ordered a couple of voltage dividers that had screw terminals when i ordered the temp sensors. That prompted me to order a screw proto shield.
The C+ code is another matter. I remember looooonnnng ago spending an hr programming (I guess you would call it an early pc?) A 6"ish green round screen with a keyboard and a 5 1/2 floppy drive. Just to get a tiny stick figure to move its arms and legs......lost ALL intrust in C at that time.

That being said, I would like to play with home automation (Trailer) solenoid triggered with a temp sensor hot water tank return system , a hydronic water heater system with a 300 watt solar assist heater coil and then lastly, A BMS system for a future lifepo4 battery upgrade.( After alot of research i dont agree with what amounts to high voltage charging for the stupid balancers. This free's me up to set a more conservative voltage setting and the ability to shut down the step up voltage system coming from the tow rigs alt.

I agree baby steps. I have read several books on the arduino programming and have a better grasp. That being said at times my eyes would glaze over and all i could see was "bla bla bla"

Thanks again
Salvo
Explorer
Dec 22, 2015
I don't know how much electronic knowledge you have. Just a cautionary note. The a/d has a 4V supply voltage. That means you can not measure voltages greater than 4V. In order to measure battery voltage you first use a resistor divider to reduce voltage to a usable range for the a/d and then multiply the a/d output by a specific factor to get the actual voltage.

You don't need any resistor divider to measure shunt voltage. This voltage is down in the millivolts. In order to measure both charge and discharge currents you need two a/d channels, operating in differential mode. The a/d you're getting has 4 channels. One channel is used to measure battery voltage and 2 channels are used to measure current.

I use a separate 9.6V Ni-CD battery (and 5.3V dc/dc converter) to power Arduino. This particular battery was used in a remote control car for grandson. I think you will get a measurement error if you connect Arduino directly to house battery. You could get an isolated dc/dc converter to get power directly from house battery.

The first step with Arduino is to write code to make a led blink. It takes a few baby steps before getting to the stage of measuring and storing data.

grizzzman wrote:
Ordered 2 temp sensors and the screen. I want live and stored data.
grizzzman
Explorer
Dec 21, 2015
Salvo wrote:
Yes, I'm going to charge at a constant current for the entire 50% SOC. At a 4A charging current, it will take some 25 hours to charge. At 1 sec per measurement, there's lots of data. I may change program to 5 sec/measurement. I'm limited to 4 - 5A constant current charging because I don't have a bigger constant current power supply that can also get up to 17V.

I'm going to add temperature to the mix. I doubt battery temperature will be any different than ambient (for this upcoming test), but it will be interesting to have. Actually, just measuring bat temp is kind of meaningless. Ambient temp also needs to be measured. However, I just have one of these temp probes. The DS18B20 probe is excellent. It just uses 2-wires and outputs a digital signal! It doesn't require a power line to power its electronics. It gets its power when the digital signal is HI. Amazing!

Grizz- This display works well with Arduino. You don't have to store the data on a SD card.
. Ordered 2 temp sensors and the screen. I want live and stored data. Great job on the graphs btw.
Salvo
Explorer
Dec 21, 2015
Yeah, that would be a good test. Drop a pair of AGMs over here in Santa Barbara and I'll run the test.

Or, build a data logger yourself. $25- should cover it.

pnichols wrote:

Similar data for AGMs would be interesting in light of their lower internal resistance in all charging scenarios.
BFL13
Explorer II
Dec 20, 2015
The "only" difference, other than--1300 minutes vs 100 minutes for the 50-90. :)
Salvo
Explorer
Dec 20, 2015
Here's a chart that will make your head spin.

The only difference between the section 50 - 80% is that blue plot is charged at 55A while the pink plot is charged at 4A. There's about 1V difference between the plots. About 0.7V is due to battery resistance and the rest is due to increased surface charge.

The plots cross at 93%. At this point the Iota is charging at 13A while the other is at 4A constant current.
NinerBikes
Explorer
Dec 20, 2015
In charging with my 150W 12v solar panel, I found that setting the charge controller to 15.0V instead of 14.8V on my 150ah T-1275 on a daily basis gave much better results in terms of SOC in the morning and remaining voltage in the battery. I agree, the battery is there to serve you, it is disposable.

What your graphs are telling me just confirms what my observations were from the start... the WFCO at 13.6V was junk, the PD9245 at 14.4V bulk was better, but just "OK" for dry camping, and the MegaWatt30 set at 14.8V on the Honda Eu1000i, finished off with the solar panel and controller set at 15.0V to try to top charge daily, was the best I've tested, so far. I still doubt that, even at 15.0V, that I was getting to 100% SOC. It's very hard and expensive, in terms of energy wasted, to get to a full 100% SOC.

If I got down to a 5 amp charge rate on the solar panel at 15.0V showing on the charge controller, at the battery too, on a 150ah battery, that was as good as it was going to get, the amperage rate wouldn't drop much further, if any. Parasitic draw accounted for 1 amp of that 5 amp charge rate.

Thanks for taking the time to graph it all out.
BFL13
Explorer II
Dec 20, 2015
My "marker" for 90% has been at 5a on 110AH at 14.5v. So 208/110 x 5 = 9.5a

Does that seem in the same ballpark as that 94% at 4a on 208AH at 14.4?

Say the 4 at 14.4 is 5 at 14.5 would amps taper from 9ish to 5ish going from 90 to 94% on the 208?
Salvo
Explorer
Dec 20, 2015
Another cool plot...

I want to note that the voltage dive seen below 12.6V is not real. Just prior to starting the charge, the battery was being discharged at a 14A rate. I stopped discharge when battery voltage (under 14A load) dropped to 11.87V. The battery had not recovered to its static, no-load voltage before initiating charging. I you project the linear plot further to the left, then 50% could be around 12.4V. That may appear high, but surface charge most likely raised the voltage 0.3V.

When above 90% SOC, battery voltage rises quickly. That means when charging with an absorption voltage of 14.2V, tapering, below 4A occurs at 93.1% and if Vabs is 14.4V, tapering, below 4A occurs at 93.9%. There isn't that much difference.
Salvo
Explorer
Dec 20, 2015
I almost never use the Iota. It just happens to turn on when starting the gen to operate the MW. 130W of solar does all the charging. I don't think anybody would be happy with just a gen/converter to charge batteries. Operating a gen for 2 - 3 hours a day wouldn't be a good boondocking experience. However, the Iota is an excellent converter to complement solar. There's nothing better than running in boost mode at 14.8V to shorten gen time.

I may record data over a few 24 hr time periods to see how well the battery holds up to solar as the primary charging source. January, out in the Arizona desert is a worst case scenario for me.

I'm not completely sure at what % SOC voltage flattens out. The chart below says it's at 100%. I've read other sources that state voltage flattens when over-charging to about 105%.

My setup allowed the voltage to rise to 15.5V. I had set the max voltage to 16.5V. I stopped the test due to very strong battery gas odor in the MH. Voltage was not going to rise much more.

I'm really not going to worry if batteries are not getting their full charge. I've said it before, they are a throw-away item. The cost of two GC2's is just a couple tank-ups of fuel.