Solar Panel System Troubleshooting - Charging Battery

In summary, the conversation discusses a small solar panel system used to power garden LED lights. The system includes a 20w solar panel, 12v 30a pwm solar charge controller, and a 12v 18ah battery powering 6x3w LED Cree lights. The lights work for one night, but then are off for 2-3 days, leading to trouble with charging the battery. The panel specs say it produces 1.23A but this has never been measured. The conversation also includes suggestions for improving the efficiency of the system, such as using a larger panel and properly sizing the battery. Additionally, the importance of regularly charging the battery to avoid sulphation is mentioned. Overall, it is recommended to monitor
  • #106
OmCheeto said:
still don't see where we've done measurements for the following configuration:

I will get those measurements this weekend.

I emailed the seller of the solar panel last night explaining the low amp readings we found. They responded this morning "asking for the test date of the voltage(?)Will replace if panel is bad." Are they checking my weather?? I never mentioned any voltage findings just amperage.

Also, I see we are 6 pages into this thread. Just want to thank everyone for their time and advice in helping me out on my little project. Very much appreciated and I have learned a lot (should have bought a clamp meter!). It's a challenge now to make it succeed
 
Engineering news on Phys.org
  • #107
Tom.G said:
Although a useful number, I don't recall anyone mentioning 'cycle count' in this thread. Can you point me to where it was brought up?
Thanks.
When a poster confused the "cycle use" charging information as printed on the battery as meaning it was a motorcycle starting battery.
 
  • #108
Wolst73 said:
I will get those measurements this weekend.
Another useful measurement would be the resting voltage of the battery each day. I would do this measurement in the morning, before you go to work.
Assuming you have the system set up now to charge the battery, and no lights on in the evening.

I emailed the seller of the solar panel last night explaining the low amp readings we found. They responded this morning "asking for the test date of the voltage(?)Will replace if panel is bad." Are they checking my weather?? I never mentioned any voltage findings just amperage.
I'm still not convinced that there is anything wrong with your panel.
We have too little data, IMHO.

Also, I see we are 6 pages into this thread. Just want to thank everyone for their time and advice in helping me out on my little project. Very much appreciated and I have learned a lot (should have bought a clamp meter!). It's a challenge now to make it succeed

Threads like this actually prompt me to do lots of testing, and figure out what's going on with my panels.
Somewhere in the forum, in a very old thread, I claimed my panels were all putting out 100%, even though they are 25 years old.
One person questioned that, and since I'd never actually checked their output, I kind of shut the hell up.
I did a load test of them on May 2nd this year, and finally figured out how wrong I was.
Their average peak output is only 22 watts. They are rated at 50 watts.

But anyways, in looking for that post I found something in one of the bazillion solar panel threads, which I had forgotten about.

Dollar Store Pumpkin Light Solar Conversion [physicsforums]
Post #8, Oct 2, 2011
Younger less senile Om said:
From Battery University:

100% depth of discharge on a starter battery = 12–15 cycles
What that means, and probably why Tom.G made the query;

Tom.G said:
Although a useful number, I don't recall anyone mentioning 'cycle count' in this thread. Can you point me to where it was brought up?
Thanks.

is that, if you've discharged your battery 12-15 times, it is most likely a very good candidate to be turned in and recycled.

Ah ha!
Found that post. {Feb 2010}

mheslep said:
clueless Om said:
ps. My 1981 1991 Kyocera panels are still operating at 100% rated capacity. Trust in quality, not statistics from the 1950's.
Yep, trust in quality. Let's also trust in physics. How do you know the panels are still 100%? Have they been in daily use since 1981?

This is one of the best things about PF, IMHO.
There is always someone around, who will call you on your, ummmm... , male-Bos taurus-excrement.

In my defense, for being a bit late, it was February.
 
  • #109
OmCheeto said:
Assuming you have the system set up now to charge the battery, and no lights on in the evening.

You assume wrong! I didn't think it would be good to have the panel hooked up if it wasn't sending enough charge to the battery. Didn't want to harm the battery any further. I will hook it back up tonight.
 
  • #110
Wolst73 said:
You assume wrong! I didn't think it would be good to have the panel hooked up if it wasn't sending enough charge to the battery. Didn't want to harm the battery any further. I will hook it back up tonight.

hmmmm... My passport expires in 2 months.
Perhaps I should make my third road trip of the year, and come visit you in Victoria, this weekend.
I have lots of resistors and thingys.

ps. Lovely town, btw. I've been there 3 times, back around 1980.
 
  • #111
OmCheeto said:
I'm still not convinced that there is anything wrong with your panel.
We have too little data, IMHO.

The results from Jim's tests don't convince you? Do you think multimeter is not accurate enough (+\- 2% + 5 for DC current) or perhaps the operator lol? I will repeat them this weekend with both charge controllers (30A and 10A) to see if we get different results. Hopefully will be sunny as its been unseasonably crappy all week.

No response yet from panel seller.
 
Last edited:
  • #112
Wolst73 said:
I emailed the seller of the solar panel last night explaining the low amp readings we found. They responded this morning "asking for the test date of the voltage(?)Will replace if panel is bad." Are they checking my weather?? I never mentioned any voltage findings just amperage.

This should help... tell them

Open Circuit voltage in mid-day sun: 19 (do i remember right ?)
Short circuit current in mid-day sun: 0.12 amp downstream of blocking diode, 0.35 amp upstream (do i remember right?)
Voltage across blocking diode : 0.23V (ditto ?)
Voltage when connected to battery in mid-day sun: 11.8, current 0.07 amp (ditto? )

and i have got myself a Canadian sunburn trying to get this frustrating thing to work.
 
  • #113
Wolst73 said:
The results from Jim's tests don't convince you?
Nope.
I want to see the results from the test I described in post #97.
From your V3 measurement in post #47, of 19.2 volts, we know that your lighting system is actually only consuming 13.4 watts.
 
  • #114
jim hardy said:
This should help... tell them

Thanks Jim. That did the trick. They are sending a new panel. Hopefully not the slow boat. Worst case, I'll have two panels to wire together.
 
  • #115
Great!

Meantime try OM's test ?
It'd be interesting to know how much current the led driver sends to those lamps.
As you know, LED voltage is rather constant, related to how many electron volts it takes to make photons of whatever color is the LED.
White LEDS used to use a UV diode that excites phosphors inside the plastic to make visible light. So they were rather high voltage, and i think you measured 19 volts for six of them ? That's probably fairly independent of current.
3.1 volts each? Hmmm that's more like green per these folks..
What is part number of yours, or datasheet link ? I looked all over Cree for 3 watt yard lights to no avail..

https://wiki.analog.com/university/courses/eps/diode-curves
LEDPlanck.jpg
so
It'll be interesting to get a current reading on LED side, too, like OM suggested, and see if it makes its 0.7 amps.. i hope we don't find a perpetual motion machine here...:))

old jim
 
  • #117
OmCheeto said:
Another useful measurement would be the resting voltage of the battery each day. I would do this measurement in the morning, before you go to work.

11.67V Thursday and Friday morning. Checked it last night around 7pm and it was 11.76V.
 
  • #118
Wolst73 said:
11.67V Thursday and Friday morning. Checked it last night around 7pm and it was 11.76V.
Well, that gives us a couple of more clues.
Another handy clue, would be an estimation of what time in the afternoon the panel was shadowed, given that sunset is 9:10 pm.

Interpolating the "State of Charge(SOC)" to "Voltage" table provided by Battery University, I came up with the following equation: SOC = (volts - 11.89)/0.0076
Plugging in the change in voltage, and battery capacity(≈240wh) numbers, I come up with 28.4 watt hours of charge, for yesterday.
Given that you live only 200 miles away, and roughly the same distance as I from the Pacific ocean, and my observations of the color of the sky yesterday, I'd say your panel is working fine.

Of course, with your battery still reading below dead, this is just a rough guess.

Whoa! I see it's already cleared up, up there. We are still under full coastal morning cloud cover.
 
  • #119
probably there's such a table at Batteryuniversity
but i couldn't find it
this one's from

http://jgdarden.com/batteryfaq/carfaq4.htm#ocv_soc
battvstemp.jpg
 
  • #120
jim hardy said:
probably there's such a table at Batteryuniversity
but i couldn't find it...
Post #21.
 
Last edited:
  • #121
OmCheeto said:

i'd been there but didn't find a table showing temperature dependence. That's what i was after.
It's only worth fifty millivolts or so...

..
 
  • #122
OmCheeto said:
configuration:
proxy.php?image=http%3A%2F%2Fwww.europa.com%2F%7Egarry%2Fpf.2016.07.10.solar.garden.lights.rev.6.png

Ok, here we go. Bright sunny day. Measurements taken between 12:30 and 1pm. Panel pointed directly at sun.
V1: 4.42v
V3: 16.26v
Measured amps at green line(negative) between Lux and lights: 0.02amps

Repeated Jim's tests with both the 10a and 30a charge controller. Same results as before and negligible difference between the 2 controllers.

Interesting the 30a controller had the green light showing battery was good, whereas when I plugged it into the 10a it showed red still.

Hope that sheds some more light. (Ha, no pun intended)
 
  • #123
Did last post with results go through? Any thoughts? Panel issue?
 
  • #124
Wolst73 said:
Did last post with results go through? Any thoughts? Panel issue?

Had to go do social responsibilities yesterday.
Your readings look like operator error to me.
So this morning, I'm doing as close as I can to the same experiment, without a constant current driver.
In your defense, I hooked up my spaghetti wiring wrong the first time, and also came up with gibberish.
2017.07.17.spaghetti.wiring.solar.test.png

Things are looking better now.
I also got out an ancient analog multimeter I picked up at a garage sale for $1.00 this spring, and have refined and eliminated many of the myriad of unknowns.
2017.07.17.cleaned.up.solar.test.png

My 33.6 meters of wire which runs to my panel, as I have to chase sunlight around my yard, measures 1.65 ohms, which means it's 22 gauge.
I also eliminated all of my 0.05 ohm test leads, replacing them with eentsy weentsy little clothes line clips.

Still waiting for the sun to come out...

Interesting that I'm getting as much power out of the panel, given the conditions.

Code:
V no load      V loaded    amps       watts
19.27           10.83      0.712      7.71096
18.72            6.44      0.421      2.71124
(post spaghetti readings)
 
  • #125
OmCheeto said:
without a constant current drive

Does my driver replace resistor? Hook positive amp terminal to positive output of driver and -ve amp to -ve of panel?
 
  • #126
Wolst73 said:
Does my driver replace resistor? Hook positive amp terminal to positive output of driver and -ve amp to -ve of panel?

What does "-ve" mean?
 
  • #127
Negative. Don't know why I abbreviated that one but not positive.
 
  • #128
OmCheeto said:
Your readings look like operator error to me.

I'm sure you're right! All results look funny?

V1 reading was done by connecting solar panel to Lux drive and measuring voltage at that connection. V3 reading was done at connection of lux output to led lights (lights on but not bright) and measuring there.

Amp reading was done by disconnecting negative wire of lux drive output from led lights and connecting positive amp meter terminal to that wire and negative amp terminal to the negative wire of led lights.

I saw on your second picture u had the amp meter negative terminal connected to negative solar panel and positive amp meter terminal to the positive output of the resistor. Using that set up, should I have connected negative amp meter terminal to solar panel and positive amp meter terminal to positive output from Lux?
 
  • #129
Wolst73 said:
I'm sure you're right! All results look funny?

V1 reading was done by connecting solar panel to Lux drive and measuring voltage at that connection. V3 reading was done at connection of lux output to led lights (lights on but not bright) and measuring there.

Amp reading was done by disconnecting negative wire of lux drive output from led lights and connecting positive amp meter terminal to that wire and negative amp terminal to the negative wire of led lights.

I saw on your second picture u had the amp meter negative terminal connected to negative solar panel and positive amp meter terminal to the positive output of the resistor. Using that set up, should I have connected negative amp meter terminal to solar panel and positive amp meter terminal to positive output from Lux?

Because these are such simple circuits, and you are using digital meters, the polarity of the meter leads is irrelevant. All that matters is that they be hooked into the correct place in the circuits.

pf.2016.07.17.solar.garden.lights.rev.7.png
 
  • #130
OmCheeto said:
Your readings look like operator error to me.

Repeated tests:
image.jpeg


image.jpeg


Where is the error of my ways?
 
Last edited:
  • #131
I found a site for worldwide weather and insolation information.

http://eosweb.larc.nasa.gov/sse
(The site doesn't work with very old browsers 'cause they use a more recent encryption protocol than was available in the past.)

On the left side of their page, click on Meterology and Solar Energy.From there you can select which data you want for your location. Remember that Longitude is "-" (negative) for the Western hemisphere. It looks like your cloudless days occur only 10% to 30% of the time depending on season. (I used 43 to 44 Lat. and -122 to -123 Long.)
 
  • Like
Likes OmCheeto
  • #132
Looks to me like panel does not send enough power to controller for it to make full output.

What do other folks see ?
 
  • Like
Likes Tom.G
  • #133
jim hardy said:
Looks to me like panel does not send enough power to controller for it to make full output.

What do other folks see ?
Yup, definitely underpowered, it's the cause we're chasing.
Bad panel, not enough panel area, not enough Sun? I think all three! Maybe panel orientation and damaged battery by now.The second one can cover for the others, but at what cost? See post #98, then factor in cloud cover from post #131.
 
  • #134
Tom.G said:
Bottom line is if you want to carry thru from two days of rain you will need at least ten times your current panel area.:H

So given my cloudy rainy winters, solar power is not a practical option? 10x current panel area just to power lights?

Any way to test battery viability? Hook up to charger and see if it charges up fully??
 
  • #135
Wolst73 said:
Any way to test battery viability? Hook up to charger and see if it charges up fully??
Yea. And then discharge it at a known rate and see if it can supply the expected runtime. I recommend you discharge not more than 50% or so. (saves batteries that way.)
 
  • #136
Wolst73 said:
So given my cloudy rainy winters, solar power is not a practical option? 10x current panel area just to power lights?

Any way to test battery viability? Hook up to charger and see if it charges up fully??
Given the length of time the battery has been fully discharged, I would take it into an automotive battery shop and have them charge it for you.
They might also be able to do a test discharge.
 
  • #137
OmCheeto said:
I would take it into an automotive battery shop and have them charge it for you.
Automotive shops aren't likely to have a charger small enough, or the knowledge, to handle that small a battery, especially a sealed one.

If you have a battery charger, you could try hooking it to your Solar Charge Controller and charge the battery through that. Just pretend the charger is the Solar Panel. Probably should monitor the battery voltage and current to see if it's working.
 
  • Like
Likes OmCheeto
  • #138
jim hardy said:
Looks to me like panel does not send enough power to controller for it to make full output.

What do other folks see ?

I don't see the panel! :biggrin:

I stopped by http://aa.usno.navy.mil/data/docs/AltAz.php that gives the angle of the sun for a given location. (I used W 123° 30', N 48° 24')
At noon today, at the southern tip of Vancouver Island, the sun was 62.1° above the horizon.
That means a 20 watt panel, laying flat on the ground, should have been generating 17.7 watts, under optimal conditions.

Obviously, with the readings indicating 1/2 watt, this was not the case.

Wolst73 said:
Where is the error of my ways?

Not making friends with a local electrician. :oldtongue:
 
Last edited by a moderator:
  • #139
Found out my neighbor had a panel gathering dust in his garage. No idea on specs. Hooked it up to multimeters and was reading 20V and 0.8A. Motomaster eliminator brand. Battery is now up to 12.23V.

Now the problem. Hooked it up in omcheetos configuration to see if readings were better. V1 was around 5V if I remember right. Was hooking up amp meter when the lights went out. Appears Lux drive is now dead. It was rated for 10-32 volts in so shouldn't have been that?? 0.8 Amps way less than battery putting out?? Shite again.
 
  • #140
Bizarre! What a bummer. :cry:
 
Back
Top