Long term follow-up on solar shed lighting

[papaof2]

The battery used in the solar lighting system in the equipment shed is AGM and they have a 3 to 5 year life when exposed to the year's varying temperatures - in particular the heat of summer. That battery's been in place for 3 years, 8 months, 25 days so it may need to be replaced soon: $$

I found a good deal on a possible replacement for that battery so I went out to make some pictures in the shed to have a "hard copy" of the configuration and the space available for the battery - the replacement LiFePO4 battery is NOT the exact same size as the original AGM battery so it won't fit in the original "ammo can" - thus the need to "see" and rearrange things.

While there, I discovered that the lights have developed a very short "blink off" about once a minute. Back in the house for a digital voltmeter to check the battery's voltage - 12.6 volts under a 2 amp load is OK but should be higher when the lights have not been on for a week or more - the battery is 3+ years old so this may be a sign of age. The solar panel was at 13.4 volts at 11AM on a mostly cloudy day so it's fine. The most likely culprit for the blink is the XL6009 boost/buck converter which holds the voltage on the LED strips at a constant 11 volts, regardless of the battery voltage (experimented with to provide enough light in the shed but not overheat the LEDs).

I'm positive I purchased a spare XL6009 board when I put the system together but I can't stand on concrete long enough to hunt for the right box :-( Two XL6009 boards (less than $5 each, delivered) should be here from eBay in a week or less so they will be a replacement and a spare. The question is whether I'll be able to stand one step up on a short stepladder for long enough to unsolder the old board and solder in the new one (4 wires in a very small space, but I do have a rechargeable soldering iron and a good head light ;-) Standing on a narrow step is harder on your body than the same amount of time standing on a larger level surface because on the step you're moving more to keep your balance. One electronic component needing to be replaced after three plus years in non-conditioned space isn't bad for a shadetree engineer ;-)

The possible replacement battery is here ($29.99 + tax + ship):

batteryhookup.com/products/relion-rb10-pc-12-8v-10ah-lifepo4-battery

It's a 12 volt 10AH rated LiFePO4 battery with a 10 amp BMS on the smaller terminals - but it also has a second set of larger terminals connected directly to the battery (no BMS, no fuse) and that connection is rated for 50 amps. Think of the location - an equipment shed with a riding mower (electric start) and a self-propelled walk-behind mower (electric start). There would always be a fully charged battery available for jump starting or boosting the battery of either of those mowers. I think it might be worth buying this battery instead of building a replacement just to get that extra bit of backup - I'll know the first time I need to do a jumpstart - and the jump start/boost wiring might not need to be bigger than #10 - just let the weak battery charge a minute or so and then hit the starter.

These LiFePO4 batteries are used but tested to deliver 85% capacity or better (about 8.5AH). The new price from the manufacturer is just over $300. ( relionbattery.com/products/lithium/rb10-pc ) I'll take 85% of the original capacity for 10% of the original price ;-) With the multi-thousand cycle life of a good LiFePO4 battery at 80% dischage, this might be a "lifetime" battery for this application. In lighting the shed, the battery isn't discharged much on a day the lights are used (typically less than 20% discharge) and the lights are not used every day. Maybe 10 or 20 year battery life at that level of use? If I needed extended lighting in the shed, the LiFePO4 battery could provide more than 3.5 hours of light before hitting 80% discharge.

The terminal arrangement is because the batteries are designed for use with wind turbine pitch control systems so you know they're built like tanks and for use in varied weather conditions. Those type backup batteries get replaced frequently and usually with a lot of life left in the "old" batteries.

Still working on recovering from surgery :-( Might be in PT in another 8 weeks or so. It seems I've been doing almost nothing but eating, sleeping, reading and taking Rx pain meds - I need to be freed from this cage ;-)

PDF datasheet for the battery:
ceb8596f236225acd007-8e95328c173a04ed694af83ee4e24c15.ssl.cf5.rackcdn.com/docs/product/RELiON-Data-Sheet_RB10-PC.pdf
Very nice long. flat discharge curve - 12.5 volts to better than 95% DOD. 24,000 cycles to 50% DOD. 12,000 cycles to 80% DOD. I see why the original price was $300+. Yup, I think I've found a lifetime battery ;-)

[biggkidd]

Just don't let a LifePO4 try to charge when frozen. That's the only part that would bother me. Where you are you don't get a lot of cold I wouldn't think.

[papaof2]

From the manufacturer's datasheet, it can be charged down to -4F (5 amp charge rate must be reduced to 0.5 amp at 14F) Since the charge source is a 30 watt solar panel, it can't reach the maximum 5 amp charge rate (Isc is 1.84 amp) and with the panel facing southwest it won't have full sun until after noon - which means the battery will have had time to warm up from the overnight low before appreciable charge is available.

This battery has all kinds of built-in protection so I think I would have to make a real effort to cause damage - $300 new for a 10AH LiFePO4 battery says a lot for the quality of the components and the in-built level of protection ;-)

I do have some long term records of yearly high and low temperatures: Record high 106°F in June 2012. Record low -8°F January 1985. Although it has happened in the past, I'm not likely to have temperatures outside the safe charge range for this battery (-4F to 113F) or the safe discharge range (-4F to 140F).

I'll probably try charging the battery using a current limited source with the battery in the freezer drawer of the big fridge - that's around 0F and would be a good place for testing. Nice to have proper equipment for testing used batteries ;-)

[papaof2]

After maybe a dozen iterations, I think I have the proper PWM charge controller configuration for the LiFePO4 battery. The battery and the controller need several days in place in the real world to see if there's enough sun in winter for a PWM controller to keep up with the use of the lights in the shed. If not, I can double up on the battery or I could get a new EpEver XTRA1206N 10 amp MPPT controller for $48 + tax (free shipping).

Based on the one discharge test I've run (more to follow), the 10AH (new), 8.5AH (used) battery tested to 7.64AH - that'a 90% of 8.5AH and I have the charge controller set very conservatively - which explains having 90% of the battery's advertised capacity. On the other hand, I'm looking to beat the manufacturer's promised "3500 cycles to 100% DOD" and "12,000 cycles to 80% DOD". If I can keep the battery between 20% SOC and 90% SOC (preferably in the 50% to 90% range), it should last even longer (number of cycles).

The testing I've done today is to determine what % SOC the lights in the shed use in a typical 20 minute activation and how long it will take the 30 watt solar panel to recharge that use with full sun (an electronic load serves in place of the light string so I can do this inside and sitting down). Then take that number and multiply it by a fudge factor because the 10 day forecast has partly cloudy, cloudy or rain. That's probably the worst possible test case so I hope to have enough bits together to be able to do the install on one of those "partly cloudy" days and then do some testing in the real world.

If my PWM charge controller configuration works as well as I think it will, I'll post that on this forum for others to explore. Might be a week or more - my body doesn't respond well to rain and I don't climb even short stepladders when using serious Rx pain meds.

Some of the PWM controller parameters match the MPPT controller one-to-one but some are different. I might also play with an MPPT controller, but only after the shed lighting system is again fully operational.

[papaof2]

It's Spring in January - today's high was 72F and it was sunny most of the afternoon. That doesn't last long - tonight's low is 39 and tomorrow's high is 56 which will be followed by a low of 34.

For the first time in almost 4 months, I tried lifting the boat door (garage door on non-sealed basement space so not valid as garage space). Didn't seem like much more than the 10lb limit I've had since surgery 3+ months ago.

Anyway, out to the equipment shed with some test equipment in one hand and the key in the other hand. After a short discharge (lights on long enough for me see to get the test equipment connected), at 4:15PM the 30 watt solar panel was delivering 0.2 amp at 13.2 volts with the shed lights off. When the shed lights were turned on and drawing 1.6 amps, the battery voltage fell to 11.9 volts (all numbers from the EPEver MT50 meter). 11.9 volts is the "1C discharge load" voltage of a 50% discharged battery so I think my decision to replace the AGM battery pack with an LiFePO4 battery came at the right time.

I have an eLog01 device (records all parameters every 11 minutes) on the charge controller to monitor it during the next few partly cloudy, cloudy, rainy, whatever days so I can see where the battery gets to with a little sun (partly cloudy), skylight only (cloudy) and during the rain that's forecast for two days (probably even darker sky than cloudy).

The solar panel does have a warranty (20 years to 80% of rated output) so I expect it's fine and any problems are with the 3+ year old AGM battery pack that's been out in winter cold (below 20F) and summer heat (above 95F) for 3+ years.

I'll know a lot more when I connect the eLog01 to a computer and read the collected data.

[papaof2]

The eLog01 device did not record data,although it was showing a green LED. It did contain the factory test data from 2017. Pressing the reset button for 5 seconds got the light to flash red - but it also seems to have bricked the device as far as reading from it. I tried resetting it again, removing and replacing the battery, etc. and then connected it to a known-to-be-working charge controller to see if it can collect and replay data. It's been in place more than 12 hours so there should be some data if it is working. If it's not working, I have designs for a couple of DIY data recorders that can poll and read RS-485 data and I just might build something. If I could buy another $80 Dell Latitude 3180 tiny Win 10 laptop, that might be a good choice for a data logger - it can run 8+ hours when fully charged and I have 12 volt to 19 volt adapters for Dell laptops to keep it charged.

Along with the LiFePO4 batteries for the solar shed lighting, I'll also be testing a new MPPT charge controller. An EPEver XTRA1210, rated 12/24 volts, 10 amps, 130 watts solar power to the battery (the controller can handle 195 watts of solar input for 12 volt systems, it just throttles it back to the maximum of 10 amps out to the battery).

Most older MPPT controllers (and some PWM controllers) have a "User" battery type, where the user defines the perhaps two dozen or so charge / discharge / protection values the controller needs for the specific battery type and size being used.

The XTRA series still has the "User" option, but it has settings for various configurations of lithium ion and LiFePO4 batteries. Just select the battery type - "F" for LiFePO4 or "N" for lithium ion - and the number of cells in series. Options include F04, F08, F15, F16, N03, N06, N07, N13 and N14 depending on the configuration of the battery. The controller has low battery temperature charge protection for LiFePO4. The top end of the XTRA series, the XTRA3415/4415N (30 and 40 amps, respectively), can handle 12 / 24 / 48 volt battery banks so the F15, F16, N13 and N14 settings only apply to that controller.

For the 12 volt LiFePO4 I'll be using, that's "F04" and all the bits and pieces of configuration I put together to use an older charge controller are supposed to auto-magically set for me ;-)

Will it work as well as I hope? I bought the EPEver controller because I've been satisfied with the ones I've used in the past (approaching 5 years in service on one of them).

I'll soon have this controller down in the basement recharging a battery that's been on an AH capacity test with discharge to 80% DOD (about 12.8 volts).

Will it live up to my expectations? I certainly hope so. At $49.99 + tax (free shipping) from Amazon, it's a great price and not much more than some of the PWM controllers.

Now to see if the product lives up to its advertising...

I still need to get several days of data from the PWM charge controller to see how much power the shed lighting system getting with our partly cloudy / cloudy / rainy weather... I do have a wifi dongle for monitoring remotely - need to see how much power that uses.

[papaof2]

Success! After letting the shed's solar lighting "burn in" for several days, I put the eLog01 logging device in place for about 48 hours to see how well the overall charge - use - recharge process worked.
The spreadsheet from that two day run looks exactly as it should.

Daytime: battery voltage drops when the lights are turned on (voltage drops under load), the charge controller sends power from the solar panel to the battery and load; when load is removed the charge controller keeps sending power to the battery until it's been at it's "full" voltage for 30 minutes and then the controller goes back to sleep.

After dark: battery voltage drops when the lights are turned on (voltage drops under load), the charge controller recognizes the voltage drop but can do nothing until there's enough sunlight on the solar panel to produce power so it waits for the solar panel voltage to rise with the sun. When there's enough solar voltage, the charge controller connects the solar panel to the battery and lets the battery charge until it's been at "full" voltage for 30 minutes and then the controller goes back to sleep.

Unless a limb damages the solar panel or the shed (and the battery & charge controller inside), this project should be good for about 10 years ;-) The 30 watt solar panel peaked at 16.39 watts at 3PM (expected peak time because the shed roof faces southwest). If that roof faced south, the peak time would have been around 1PM and the peak power would have been higher. The angle of the sun changes through the day and the sun's angle on a solar panel is critical in determining how much of the panel's rated power is available: that's what solar trackers do for you - ensure the panel(s) always have the best possible alignment with the sun. Depending on the location on the Earth and the sun's angle to the panels, a tracker can provide 20% or more increase in available power. I haven't built a tracker (yet) but I do have most of the parts... As long as the current configuration keeps up with the usage of the shed's lights, I probably won't make any further changes.

[papaof2]

For anyone who's interested, the LiFePO4 battery pack in the shed is one of the Fuji battery packs that can deliver 40 amps (from batteryhookup.com and that battery sold out in a couple of days of being listed :-( The advantage is that battery pack fits inside the ammo can I used for the AGM batteries and the PWM controller is mounted on that ammo can - disconnect 6 wires and bring that box in for testing or reconfiguration or use with another solar panel or just to have another 12 volt source available in the house.

I'm considering the LiFePO4 batteries from the wind turbines as power for some in-house backup lighting: basement first and then various rooms on the main level. Not looking for the light level of several "40 watt equivalent" or bigger light bulbs, just "safe passage" lighting - something someone outside might think is a flashlight aimed at the ceiling. A few 6" sections of a "12 volt" LED strip placed on the wood trim above the windows would probably provide the level of light I'm looking for and a 12 volt, 7Ah battery should be able to run that level of lighting for several days. I could also use a DC-DC buck converter to drop the voltage to the LED strips to something "adequate" - as I did with setting the buck regulator for the shed lights to 11 volts: bright enough to be used but the LEDs don't get hot. You need enough power to the LED strip(s) to get the minimum needed lighting level. There are also some "1 watt" LEDs that would need a buck regulator to drive them but they provide "reading level" light or a very impressive hand lantern ;-)

And be sure to have some of the "squeeze" and/or "crank" flashlights out if you allow any visitors in. We have both types in the bedroom and near the rack the backup power system is on (also crank and wind-up radios). We have candles and a lighter near the backup system. I think the candles are more correctly called "backup for the backup" ;-)

My EPEver LS1024 / LS2024 PWM charge controller settings for 12 volt LiFePO4 battery (combined from multiple sources but they work for me - probably OK for the Tracer series 1210 through 4210 but haven't tested that yet):

BatType: 0 (User)
BatCap: 10 (AH)
TempComp: 0
OVCutOff: 14.60
ChargLmt: 14.50
OVRcv: 14.20
EquilV: 14.45
RaisV: 14.45
FloCharg: 13.90
RaisRcv: 13.30
LVRcv: 11.20
WarnRcv: 11.30
LVWarn: 11.20
LVCutOff: 11.00
DischaLmt: 11.00
RatedLv: 0
EquilDur: 30
RaisDur: 30
BatDisch: 30
BatCharg: 100
ChargMode: 0

If you're looking at the device configuration screen on a PC, you'll figure out the ones that may seem oddly named ;-)
Remember, they're designed and built in China so even the abbreviations are in Chinglish.