Solar Panel System Troubleshooting - Charging Battery

[Wolst73]

How does the PWM controller do that ?
Does it open a switch that's between the battery and the Flexblock ?

I believe it does.

 

[mheslep]

..

So we see that independent of panel aiming, the watts/m^2 of sunlight at the Earth's surface varies enormously with latitude, approaching zero at the poles.

Russ was largely correct for a *clear* summer day in most of the US. The collector and the "earth's surface" are two different things. If the collector is tilted south at latitude, it corrects for the oblique angle of the Earth's surface which otherwise reduces power per area.

The factor that can't be removed by panel pointing is path length of attenuating atmosphere for solar radiation, which grows larger at higher latitudes. When the sun angle to Earth surface determined by time of day or latitude is 75 degrees, the path length is four times greater than at an angle of 0 degrees.

https://en.m.wikipedia.org/wiki/Air_mass_(solar_energy)#Definition

The measured effect of latitude can seen by looking up the NREL solar radiation data for collectors "tilted at latitude" against "flat plate". In mid summer, some US northern cities under good conditions compare well with southern cities.

 

[jim hardy]

Surely the charge controller is temperature compensated ,
and located so as to be at same temperature as battery ?
Looks like an uncompensated charger would undercharge a cold battery.
Surely nobody would sell such a thiing in Canada...

http://www.tekrispower.com/pdfs/xantrex/Batteries%20-%20Temperature%20Compensated%20Charging.pdf

 

[anorlunda]

Russ was largely correct for a *clear* summer day in most of the US. The collector and the "earth's surface" are two different things. If the collector is tilted south at latitude, it corrects for the oblique angle of the Earth's surface which otherwise reduces power per area.

The measured effect of latitude can seen by looking up the NREL solar radiation data for collectors "tilted at latitude" against "flat plate". In mid summer, some US northern cities under good conditions compare well with southern cities.

What you say is correct, but there are other factors that can't be compensated by attitude tilt angle of the panel. The seasonal variation in daylight hours is one. The azimuthal variation in the sun's direction is another. Anyone who has lived in the far north knows that the sun does not rise in the east and set in the west, it goes around you in a circle. My daughter grew up at 65 degrees latitude. Later in life we moved to 42 degrees. She heard the phrase "the Sun rises in the East." and she said, "Says who?"

Heliostats are cool, they can follow both the sun's altitude and azimuth.

Nobody nibbled at my little riddle in #12 about a panel at the north pole. Up there, the panel's altitude angle should always be nearly 90 degrees, but the azimuth must swing 360 degrees in summer. But the riddle part is that all those azimuthal angles are South when starting at the North Pole.

Also, nothing in the OP limited the question to mainland USA or to summer months only. I always try to remind myself that PF threads can be googled. The posts can be discovered and read by anyone from any country at any time in the future. When posting on PF, I try to visualize that broad audience rather than a private conversation between the participants.

 

[Wolst73]

When multimeter hooked directly to panel it shows around 19v. When I hook multimeter up to where panel goes into solar charge controller it shows 11v? That is also the reading you get when hooked up to the next terminals which are the battery connections.

A PWM charge controller does absolutely nothing when the voltage is below 13.8. The controller is not a big source of inefficiency, it doesn't contribute inefficiency at all as long as the battery is undercharged.

Is this the reason for the slow charge? Where/why is there a dropoff in voltage? I would have expected the first terminals to still show 19v and the battery terminals to show whatever the batteries current voltage is?

 

[OmCheeto]

OmCheeto, post: 5510606, member: 103343"]Simply hook your panel directly to the LED system, without the battery.
Then let us know what the voltage is.

Solar Panel hooked directly to DMM showed 19.4V.
Panel hooked directly to Vin on led driver and DMM hooked up to Vout showed 40.4V

And I thought my numbers were goofy.
40.4 volts * 0.7 amps = 28.3 watts

Didn't you say your LED's were 18 watts?
And didn't you say your panel was rated at 20 watts?

Is it a fair guess, that you didn't hook up your LED's?

Maybe this will shed some light on the problem?

You really need another meter...

ps. Fun thread.

Found out my battery was not sulfated, like my last one.

charging a battery, with a 40 year old, DIY battery charger:
time    amps     volts     watts     wh cumulative     SOC     notes
00:00    0.0     12.46     0                           75%     resting voltage
00:00    2.9     12.86     37                                  started charge
00:01    2.7     13.09     35            0.6       
00:02    2.6     13.26     34            1.2       
00:03    2.5     13.40     34            1.8       
00:04    2.4     13.48     32            2.3       
00:05    2.4     13.53     32            2.8       
00:06    2.4     13.55     33            3.4       
00:07    2.4     13.57     33            3.9       
00:08    2.4     13.57     33            4.5       
01:00    2.4     13.58     33           33       
02:00    2.3     13.65     31           65       
03:00    2.2     13.71     30           97       
06:10    2.1     13.82     29          192       
07:30    2.2     13.85     30          232       
10:00    2.1     13.89     29          306                      ended charge
16:20            12.68                                  100%    resting voltage

 

[Wolst73]

Is it a fair guess, that you didn't hook up your LED's?

Sorry. You are dealing with an extreme rookie when it comes to electrical. Thought you meant to see what voltage was coming out of the driver. Where do you want me to hook up meter? 6-3w led wired in series.

Did the results above with regards to the charge controller make sense to you?

P.S. I googled Hall effect meter and saw what it is but can't find where to get one.

 

[jim hardy]

i guess it's this charger
they say it's temperature compensated

http://solar-thailand.com/EN/Product/20A12V24V/

Code : 20A12V24V

CMTB-20A Solar Charge Controller for Street light

CMTB-20A - CMTB series light & time controller（PWM) has light and time control Ideal for public lighting area, such as street light, path way, garden lights, parking area, bus station etc.
·High efficient Series PWM charging, can prolong the life span of battery
·Use MOSFET as electronic switch, without any mechanical switch
·With high precision temperature compensation
·Widely used, automatically recognize day/night
·With light and time control, intelligent timer function with 1-15 hours option
·Digital LED menu with simple setting and easy using
·Electronic protection: over charging, over discharging, overload, and short circuit

When multimeter hooked directly to panel it shows around 19v. When I hook multimeter up to where panel goes into solar charge controller it shows 11v?

Right there's a mighty strong clue...

 

[anorlunda]

Is this the reason for the slow charge? Where/why is there a dropoff in voltage? I would have expected the first terminals to still show 19v and the battery terminals to show whatever the batteries current voltage is?

Think of the PWM controller as an automatic on/off switch. At say 13.7 battery volts and below, it is ON 100% of the time. At 13.75 volts it is ON only 50% of the time (switching ON/OFF perhaps 10 times per second). At 13.8 volts and above, it is OFF 100% of the time.

You should have nearly zero voltage difference between the panel and the panel-side of the controller. If that is not the case, then there is a bad connection, or a damaged wire, or a seriously undersized wire.

FYI, but not part of your problem: The panel itself is highly nonlinear. The following graphic shows the voltage-current curves for a single cell. Each curve is for a specific sunlight strength. Panels are made with many cells, and open-circuit voltages 14-45 volts. The wiggly line shows the maximum power point which is the optimum operating point. MPPT controllers (which cost $100-$150 more) automatically track the optimum.

 

[Wolst73]

You should have nearly zero voltage difference between the panel and the panel-side of the controller. If that is not the case, then there is a bad connection, or a damaged wire, or a seriously undersized wire.

Same part of wire used for testing direct from panel that is in the controller so shouldn't be that. Bad connection inside controller? This is the second controller I have bought. First one didn't power on at all. Aargh. Does anyone have similar controller that can reproduce my test and get the expected results?

 

[mheslep]

What you say is correct, but there are other factors that can't be compensated by attitude tilt angle of the panel. The seasonal variation in daylight hours is one.

Sure, forcing solar collection down to useless in high latitude winters, but since Russ posted and you responded to him personally we're not discussing the annual case, we're discussing the ideal summer day case Russ posited. I added the qualifier of in the US to not waste time with extreme latitude cases (where nobody will be installing solar farms)

 

[jim hardy]

Same part of wire used for testing direct from panel that is in the controller so shouldn't be that. Bad connection inside controller? This is the second controller I have bought. First one didn't power on at all. Aargh. Does anyone have similar controller that can reproduce my test and get the expected results?

19V at panel, 11V at charger ?

Time for a sanity check.
To expand on anorlunda's suggestion

You should have nearly zero voltage difference between the panel and the panel-side of the controller. If that is not the case, then there is a bad connection, or a damaged wire, or a seriously undersized wire.

take one of your voltmeter probes in each hand.
First: place one probe on + terminal of solar panel and other probe on +panel terminal of charge controller (leftmost one)
If you read 0.0 volts that's good
if you do not read 0.0 volts you are indicating voltage across that wire which isn't really possible. A few millivolts maybe but not hundreds or even tens of them.

Second: place one probe on - terminal of solar panel and other probe on - panel terminal of charge controller (next to leftmost one).
If you read 0.0 volts that's good
if you do not read 0.0 volts you are indicating voltage across that wire which isn't really possible. A few millivolts maybe but not hundreds or even tens of them.

Assuming both those tests come out okay
repeat your voltage readings across the solar panel terminals and across the charger controller's panel terminals.

If you once again do not get the same reading both places then please advise so we can help you figure out why Kirchoff has abandoned you.

old jimFor next post so i don't misplace it...
Well, I'm getting to long shots.
This is from datasheet at
http://en.ueiua.com/supply/11.html

Sequence of hookup ? Panel before battery, sees 19 volts so sets itself for 24V ? Try disconnect everything, hook up battery first.
also it's a PWm so may draw its current in gulps instead of constant. Try a few hundred microfarads across its panel terminals.

 

[jim hardy]

AHA ! the plot thickens...

found directions for a CMT controller (has white face) made by same folks here
http://www.tradekey.com/product-free/Outdoor-20a-Overload-Protection-Pwm-Solar-Charge-8288618.html
but have been unable to find directions for CMTB(has black face like your photo) as yetCMT appears to be sensitive to sequence of hookup
here's from that instruction leaflet linked above

did any instructions come with your device? Do they differ from this ?

old jim

 

[Wolst73]

did any instructions come with your device? Do they differ from this ?

Same hookup order.

 

[Wolst73]

Somewhat. I'm guessing you did not have it hooked to the battery.

Battery was hooked up. Red light was on on controller (couldn't see in pic). Green light showing power from solar panel.

Lux drive is hooked up to charge controller. It will show 0 because controller won't switch to load when battery is low and is getting enough power from solar panel. I will unhook it from controller and wire directly to battery tomorrow and report.

 

[OmCheeto]

Sorry. You are dealing with an extreme rookie when it comes to electrical. Thought you meant to see what voltage was coming out of the driver. Where do you want me to hook up meter? 6-3w led wired in series.

Measure the voltage everywhere, with EVERYTHING hooked up: Solar panel, charge controller, battery, Luxdrive, and lights.
Lights on. Sun shining around noon. Panel facing sun, with NO shadowing.

According to my diagram, you should have 4 different voltages: V1, V2, V3, V4

Did the results above with regards to the charge controller make sense to you?

Somewhat. I'm guessing you did not have it hooked to the battery.
The controller would sense that, and is probably designed to provide "dead battery" voltage, which according to Battery University, is 11.89 volts.
Which is very near the 11.67 volts you have displayed.

Battery was hooked up.

Then I'm now guessing your battery was like the wicked witch, really most sincerely dead.ps. Sorry about the time traveling, but I noticed my original schematic was wrong.

 

[Wolst73]

First: place one probe on + terminal of solar panel and other probe on +panel terminal of charge controller (leftmost one)

Positive terminal was 0.23v
Negative terminal was 0

According to my diagram, you should have 4 different voltages: V1, V2, V3, V4

Panel by itself: 20v
Panel hooked up to charge controller (V1): 11.6v
Battery terminal/battery charge controller (V2): 11.6v
Led drive connected to battery, lights on (V3): 19.2v
V4 I can't measure because controller won't switch to load if battery is low (or I don't know how to force it to).

Tried the other charge controller (which I said earlier didn't work). Apparently this one you have wait a few minutes before it turns on. Other one was instant. Same results for V1 and V2.

Jim mentioned this above. This is from the manual that came with the controller. Does that mean on my 12v controller it can only handle maximum 17v before it shuts power down? That could be why battery is so slow to charge because voltage has to fall before circuit to battery is turned back on? Also explain why V1 is lower than panel by itself? Meaning panel is too big? Just a thought.

 

[Wolst73]

Tried covering half the panel with some cardboard. Reduced panel voltage to around 15v. Now V1 shows only 9v and V2 is still 11.6v.

 

[jim hardy]

Our posts crossed in the E-wind. You're headed in the right direction.

jim hardy said: ↑

First: place one probe on + terminal of solar panel and other probe on +panel terminal of charge controller (leftmost one)

Wolst73 said:
Positive terminal was 0.23v
Negative terminal was 0

Hmmmm... 0.23 volts something is bad wrong with that + wire... Look for a tiny piece of insulation under one of your screw terminals. late edit turned out the 0,23 volts was the blocking diode in series with red wire
Might be an intermittent ? Wiggle the wires.
( i hate to ask this- don't be offended...
You did skin the wires, didn't you? My sister had an MG automobile with repeated taillight problems until we figured out the factory guys hadn't stripped the insulation off the wires before attaching the push-on terminals. Taillights only, musta been a new guy.)
Sometimes a wire will get stretched and break most of its internal strands. Feel it for a spot that's thinned or soft.

Does that mean on my 12v controller it can only handle maximum 17v before it shuts power down?

It is very unclear just what they meant.
The symptom you describe is so confusing i didnt know what to make of it.
So i took it at face value - if it sees more than 17 volts it powers down for self protection.
If you hook up the battery first, the controller has someplace to put the current from solar panel
if you hook up solar panel first the controller has noplace to put the current so panel voltage will rise, and you measured 19 volts. Does that make it shut down ? I don't know either

Another experiment ?

Cover the panel with a tarp or something
disconnect panel and battery
reconnect battery , wait for computer to wake up
then reconnect panel
connect voltmeter to panel terminals on controller
gradually remove tarp uncovering panel a little at a time, watching voltmeter.

Every experiment is worth a thousand speculations !

 

[Wolst73]

Cover the panel with a tarp or something
disconnect panel and battery
reconnect battery , wait for computer to wake up
then reconnect panel
connect voltmeter to panel terminals on controller
gradually remove tarp uncovering panel a little at a time, watching voltmeter.

When removing cover slowly volts rise to max 11.6. When remove cover all at once, volts rise to 19.9 then drop instantly to 11.6.

When panel is connected to controller these terminals measure about 12v. Unconnected they are 19v. Same as what I measure with the wire leads.

 

[anorlunda]

If you look at these curves from #39 again, you can see that dropping the panel volts from 19 to 12 means that it is pulling nearly short circuit current. (The difference between rated current and short circuit current is small. You can see that on the nameplate sticker on your panel.)

That raises two questions.

1. If the panel is sending 1.2 amps, and the voltage drop is 0.23 volts, then the resistance of the connection is 0.19 ohms. That's too much resistance. Either there is a bad connection or an inadequate wire. 0.19 ohms is the resistance of about 12 feet of 22 gauge wire. 22 is far too small a gauge. 18 gauge wire gives about 1/3 as many ohms per foot. 16 gauge wire 1/4 as much.
2. If more than one amp is going to the battery and the battery voltage is still less than 12, the battery is definitely bad (probably a cracked lead plate). Don't you have any current measurements? If you buy a replacement, make sure it is deep discharge battery. Also make sure all the wires in your project up to the Luxdrive are also at least 18 or 16 gauge, and make them only as long as necessary.

Edit: On second thought, if you buy a larger deep discharge battery, it too is likely to become sulphated. To minimize supplication, you would have to fully recharge the battery on the first 2 hours of peak sunlight each day, and leave it on a float charge for the other 3 hours per day. (At 49 north, you may not have many peak hours of solar in December).

I repeat my earlier recommendation, Nicad or Lithium batteries are better suited to your application.

 

[jim hardy]

oops i missed anorlunda's post. Crossed in the e-wind...

When removing cover slowly volts rise to max 11.6. When remove cover all at once, volts rise to 19.9 then drop instantly to 11.6.

Wow !
That is a good piece of information.
It says the controller does not immediately accept any current . One presumes the computer is in control of that..
And - your photo explains the 0.23 volt drop in + wire, it's that black diode with red wire soldered to it.

So - why does voltage drop to 11.6 , same as battery, yet we think battery is not getting charged ?
old troubleshooter's fallback ------------
When nothing will show itself faulty
you have to make everything demonstrate that it's good.

Do we know the panel is capable of delivering its full current ?

Hmmm nameplate says short circuit current is only 1.23 amps

What i would do next is figure out a way to test that.
I have a 3 step test in mind, read it over and see what you think :
......................

I blew up that picture with your DMM in it.
Your DMM looks like its highest current scale is 0.2 amps and we don't want to blow the fuse that's inside of it.
So we can't just set meter for amps, hook it across the panel and uncover like you just did for volts (which was a great test by the way)

But we could put a shunt across the panel and measure voltage across the shunt.
Does your meter have a 200 millivolt scale (or 100 millivolt) ?

Do you have access to some small wire, # 20 or #22 ? That'll make a primitive shunt.
#20 is 10 milliohms per foot, #22 is 16 milliohms per foot.
Ten feet of #20 would make 100 milliohms, a 1/10 ohm shunt
Just a couple inches over six feet of #22 would make 100 milliohms, 1/10 ohm shunt.Step 1:
I would disconnect panel from controller
cover it with your cardboard
connect my shunt across its terminals
connect meter to read voltage across my shunt
uncover panel and observe voltage.
You should see about 0.123 volts in the sunlight for which that panel was designed.
However much less than that you get is how much shy of panel's rating you get.

That experiment is a sanity check for broke panel.

Step 2:
Then i'd move my shunt to the far end of those two wires intended for panel controller, twist it to the red and black from panel but don't connect them to controller
and repeat the test, again measuring voltage right at the shunt
That checks that the diode and wires are capable of carrying an amp from the panel to their far end where the controller is

Step 3:
Then i'd hook the positive red wire back up to the controller but not the negative
i'd hook my shunt in place of the black wire , one end to controller and other to panel
hook meter to measure voltage across my shunt , that is between its ends at panel and controller just as when you measured voltage drop across wires (reported 0 and 0.23)
That voltage across shunt X10 will be your current in amps. 0.05 volt = 0.5 amp.
Uncover panel and read the current.
That current X volts across panel is the watts you are getting out of your panel.

..............................
Those three steps should answer a couple of questions:
1. Is tmy panel capable of delivering an amp or two to the panel ? What's the best it can do in my local sun?
2. Does my controller accept current from the panel? How much ?
3. How many watts does my panel really deliver in vivo ?

That should point you toward the reason your system isn't doing what you want done.

And you'll no longer be a noob !

Buy some extra shunt wire, your friends will want one too. Hardware stores sell #20 bell wire and the lowest grade of el-cheapo speaker wire is #22.

Have fun ----

old jim

 

[Wolst73]

Your DMM looks like its highest current scale is 0.2 amps and we don't want to blow the fuse that's inside of it.

0.2 Amp? I'm sure I've blown the fuse a while ago then. What's the point of making a DMM that measures so little? Granted it was only about 30$.

I may need a diagram to explain the shunt connections if possible.

 

[jim hardy]

0.2 Amp? I'm sure I've blown the fuse a while ago then. What's the point of making a DMM that measures so little? Granted it was only about 30$.

It's cheap. A ten amp high energy fuse that's safe at your meter's 350 volts is expensive so they don't put a ten amp scale on many DMM's
When somebody competing for the Darwin award wraps the fuse they just blew with tinfoil then commences to use the meter in their electrical panel - that's a setup for a fireball big as a weather balloon.Here's what makes me think your meter doesn't have a high amp scale
but it's right at the limit of readability for me

Here's how to make the shunt out of plain old wire

be aware it's not terribly accurate and as the wire warms it changes resistance by copper's temperature coefficient, about 0.4% per degree C
but this ain't the Steinway factory... ±a few percent will do us fine.

For step 1
you'll connect the two ends of your shunt across your solar panel terminals and measure voltage across the terminals

For step 2
you'll twist the two ends of your shunt to the far end of the panel's red and black wires and measure voltage there

For step 3
you'll hook the two ends of your shunt to where you unhooked the black wire's two ends from panel and controller
use the shunt instead of the black wire so you'll know how much current is flowing there. You'll have to unwrap a few turns and untwist them to reach I'm sure.

I hope it tells us a lot

old jim

edit oops i see the red and black wires are soldered to the panel
so you won't remove the black wire just tape its loose end
and use your shunt instead of the black wire
(i thought they were landed under the screws... doh)

 

[jim hardy]

wrote in a hurry hope it's clear

late in day anyhow you'll have to wait for sun to get up high again

 

[Wolst73]

Will have to wait until next weekend. Back to work.

Probably dumb question, but is the shunt bare wire?

 

[OmCheeto]

Will have to wait until next weekend. Back to work.

Then you'll have to wait 2 weeks for a reply from me, as I'll be out of town next weekend.

And given that anorlunda and I both think your battery is nearly, or already toast, I would recommend setting everything up to charge the battery until then.
And NO Lights!

And it's supposed to rain on Thursday, so it would be wise to make sure your electronics are covered.

Panel by itself: 20v
Panel hooked up to charge controller (V1): 11.6v
Battery terminal/battery charge controller (V2): 11.6v
Led drive connected to battery, lights on (V3): 19.2v

These voltages look like what I would expect.

And - your photo explains the 0.23 volt drop in + wire, it's that black diode with red wire soldered to it.

Good catch!

Granted it was only about 30$.

hmmmm... (google, google, google)

nearest Walmart:
860 Langford Parkway
Langford, BC, V9B 2P3
...
HVTools Digital Multimeter (HV830B) $17.53 (Canuck bucks)

Capable of measuring 10 amps.

Probably dumb question, but is the shunt bare wire?

Yes. That is a dumb question.
But one would expect that from someone who said;

Sorry. You are dealing with an extreme rookie when it comes to electrical.

We were all dumb rookies, once.

IMHO, "shunt" can have two meanings. Which, IMHO, means that it's an overburdened term.
Though, wiki lists 7 different meanings, so I guess I'm a bit "dumb" for not knowing about the other 5.

Shunt (electrical)

Contents:

1 Defective device bypass
2 Lightning arrestor
3 Electrical noise bypass
4 Use in electronic filter circuits
5 Diodes as shunts
6 Shunts as circuit protection
7 Use in current measuring​

Actually, one of my two definitions was "short circuit".
hmmmmm...

 

[jim hardy]

Probably dumb question, but is the shunt bare wire?

not at all

you'd best use insulated wire. Else it'll short out every place it touches .
Bare would be best for heat dissipation, but with any length keeping the copper conductors from touching one another as we twist and wind it becomes impossible..

A proper shunt is made from a short length of some metal with zero temperature coefficient, manganin is popular. You can buy low ohm resistors made for current sense but i assume you're not close to an electronics supply house.
http://powerelectronics.com/passive-components/current-sense-resistors-heed-call-more-power
The reason for grabbing our wire in the middle to make it a double conductor is so it will be non-inductive when we coil it.. After you wrap it around the ruler , current will flow around the coil clockwise in one half the turns and counterclockwise in the other half, so the magnetic fields cancel. See :Right Hand Rule".
Twisting them together makes it easier to handle.

 

[Wolst73]

...
HVTools Digital Multimeter (HV830B) $17.53 (Canuck bucks)

I will get new meter. Of course that link shows its out of stock at that Walmart (hate Walmart).

Can diagram be adjusted to show where I am to measure with multimeter? (sorry Jim)

 

[jim hardy]

Can diagram be adjusted to show where I am to measure with multimeter? (sorry Jim)

test 3: Borrowing from OM's diagram of post 46
Green is usually negative... so i drew shunt there

sorry for low quality
i'm not much good with graphics and MSPaint is torture.

Good Luck !
Sears may have an inexpensive meter...