Thursday, August 16, 2012

Will Harbor Freight Solar Panels Back Feed Each Other?

From guest blogger Hillbilly Gene......

The Inquiry: Will HF Solar Panels Back Feed Each Other?

No one, to my knowledge, has ever conducted a test to see if three solar arrays, such as included in the Harbor Freight 45 Watt kit, will interact and back feed current from one panel to another. I know that the HF charge controller has “blocking” diodes to prevent a nominal 12 volts from the battery from potentially feeding current back to the solar panels at night.

Logic would suggest that if back feeding were not possible, then HF would not have included those diodes. If back fed current were possible with a nominal 12 volts, then why wouldn’t the 23.5 Voc of one panel back feed another panel being shaded, and not producing 23.5 Voc?

Are those blocking diodes in the HF CC there for another reason? Such as if someone inadvertently shorts the solar panel input wiring? Instant smoke and fire inside the HF CC would result without those diodes; not counting melted wire back to the location of the “short”. I know this scenario would only happen if the negative of the battery was also connected to the negative of the panels, but anything is possible. Anyone concerned with grounding issues might do this; such as grounding the negative of the solar panels and also grounding the negative of the battery at the same point. That wiring would set up that Murphy’s law scenario.
I decided to conduct a test to see for myself.

Test Instrumentation:

My instrumentation was fairly simple:

Two HF 98025 Multimeters:

One was used with the 10ADC function; the other was used as a regular voltmeter.

One Harbor Freight 45-Watt Solar Panel Kit. This kit has been installed and working beyond the six-month “seasoning” period; the time when the electrical output of the 3 panels are expected to be permanently reduced by 20% from their like-new output. This HF kit is assembled per the HF manual, pointed at the sun with no particular fussing as to exact angle.

Brown corrugated cardboard: large enough to cover just one HF panel.

Miscellaneous. This is a category for anything used, but not specifically mentioned, such as jumper wires with alligator clips. Muttering and cussing belongs here.

Testing Parameters:
First, all three panels were cleaned with “Windex”

Testing consisted of reading the combined voltage of all the parallel panels and reading the amps of one panel being shaded or not shaded.

Shading was a piece of corrugated cardboard that just fit over one panel. It was placed on top of that panel, with no provision to secure it. Light may have filtered into the panel around the cardboard edge, but I made sure that no sunlight was directly on the panel. The one-panel shading is labeled “drastic” shading because I doubt any panel would receive that much individual shade in any normal installation.

Volts and amps were recorded with the cardboard on or off the panel, and with the rocker switch of the HF CC on or off. Rocker switch “on” was the battery-charging voltage; Rocker switch “off” simulated Voc, or no charging.

This test was conducted on the 17 of December. That’s close enough to the winter solstice (21 December) to be at a time when incoming solar power is weakest, because the earth’s axis is tilted as far north as it will get annually. A Photon spit out by the sun takes 8.3 minutes to reach earth and slows down when it reaches the earth’s atmosphere. Since the earth is tilted away from the sun during the winter it has effectively more atmosphere to travel through, and loses energy.

The sun’s Photons just don’t have enough energy to knock too many electrons silly in the panel this time of year, but the dazed electrons that do get whacked make up the current flow from the panels to the charge controller.

Time was about 9:30 – 10:30 am. Latitude in northern hemisphere: draw a line between Kansas City and Indianapolis, and you’re close enough. I believe anywhere mid-country between Louisiana and the Canadian border would give approximately the same results. It was a clear, sunny, day. Ambient temperature was 32°F.
The circuit wiring used is shown in Fig. 1

Test Number 1:
With a HF 98025 Multimeter on the 10A scale (reading panel #1 amps), and a second HF 98025 Multimeter set to 200 VDC (reading the parallel 3-panel-array voltage at the charge controller).

Here is the Test #1 data I collected:

The yellow-highlighted results of Test 1-D indicated that the shaded panel #1 wasbeing back fed current from the other panels. Minus amps means the current is going in the opposite direction from Test 1-B. The Volts of all the panels also apparently dropped about 0.2 volts.

Reasons for Test Number 2:
I wanted to see exactly what that –0.02 amp reading (-20ma) in Test 1-D really was. The HF 98025 meter will read up to 9.99 amps on the normal 10-Amp scale with a precision of ± two rightmost digits, meaning that –0.02 reading could be –0.12 or it could be –0.00.
So I devised another test to determine that amps value given in Test 1-D more precisely.

Test Number Two:

Start of theory:
First, the theory guiding this test: I know that E = I x R. So if I make R = 1 ohm, then E = I x one. Any voltage that is read across a one-ohm resistor can be interpreted as current, because the formula now reduces to E = I. It’ll be a cheap Hillbilly Ammeter Substitute.

A 1-watt resistor was chosen because W = I (squared) x R. Since Panel #1 has a rating of 1-amp, my equation now becomes W = (1-amp) x (1-amp) x (1-ohm) = 1 Watt, which will easily handle solar panel #1 current.

And then I assembled the circuit as shown in Fig. 2.

Using the HF #1 Voltmeter setting as an Ammeter setting:

The HF #1 Multimeter was set to either the 200m DCV or the 2000m DCV range as needed. Normally the 200m range displays up to 0.20 volts, and the 2000m range displays up to 2.00 volts. Because of my setup, I can now read and instantly interpret the meter either as a 200-milliamp (0.200 amp) range or as a 2000-milliamp range (2.000 amp, actually 1.999 amp) respectively (m = 1/1000), rather than voltage. The HF meter on these ranges has an accuracy of ± one rightmost digit, which is much better than the 10A DC range accuracy of ± tworightmost digits.

The polarity of the HF #1 meter was selected to read positive current from the panel to the charge controller. Any resistance in a working circuit will have a voltage drop, and that voltage drop will have a polarity to oppose the source voltage. The source voltage in this case is the voltage from solar panel #1. If the voltage across that resistor switches polarity it means current is going intothe solar panel.

End of sermon theory.

The testing parameters duplicated Test Number 1 parameters.
Test Number 1 is repeated here for easy reading and comparing with Test Number 2

Results of Test 2:
Comparisons of both tests:

Focusing on the yellow highlighted numbers:
Test 1-D indicated panel back feeding of -20ma intopanel #1 by panels 2 and 3, but the exact amount was either inconclusive or suspect. CC not charging, panel #1 drastically shaded, Voc = 23.1
“OL” in Test 2-B stands for overload, or more than that meter setting of 200mv could read
Test 2-C indicated back feeding of –2.2ma intopanel #1. CC charging, panel #1 drastically shaded. Note that Test 1-C under the same conditions did not indicate any back feeding, because the 10-A ammeter setting in 1-C can’t read 0.002 amps, although the polarity was –000.
Test 2-D confirms Test 1-D that back feeding isoccurring and the amount is –21.4ma.
That cheap Hillbilly Ammeter Substitute mentioned in the Test Number Two Theorysection worked as planned by verifying panel #1 back feeding of over –20ma. See test 1-D vs. test 2-D.

My conclusion is: The HF Solar Panels WILL back feed, given the right conditions. Murphy’s law rules here, and diodes are cheap. I will therefore wire my permanent kit in the following manner:

If I have to use a combiner box I will install the diodes and fuses in that box. If I don’t need a combiner box I will place the diodes and fuses somewhere convenient near the charge controller.

I want to redo this testing about June 21, when the solar energy is at it’s maximum. My thoughts are that there will be more amps back feeding because there will be more current available from the uncovered panels at that time of year. The sun’s Photons has more muscle.

As usual, I will be grateful for anyone pointing out any trashing of theory or for any sins of omission. I try to include everything, even if it means being long winded.

Hillbilly Gene

P.S. For those that don’t know, Murphy’s lawstates: “If anything can go wrong, it will”.
Some will say Murphy was an optimist.

Please join us over on the Do it yourself solar energy forum and check out all of Hillbilly Genes articles.

Everything You Need to Produce Your Own Solar Energy.....Check Out New World Solar Power!

No comments:

Post a Comment