Effects of partial shading

[adaviel]

I have some unbranded 3x3 cells from eBay (might be evergreen or similar, giving 0.5V). When I connect them in series, although the voltage seems unaffected, the current seems strongly affected by shading. If one cell is 50 or 80% shaded, there's little difference, but if the cell is over 95% shaded the current falls substantially from the entire string. So if, say, a bird sat on the panel and covered an entire cell, the performance would be affected far more than the simple area involved.
Is this normal ? If so, how do commercial systems cope with it apart from ensuring that installations are never shaded at all ?

[inetdog]

I have some unbranded 3x3 cells from eBay (might be evergreen or similar, giving 0.5V). When I connect them in series, although the voltage seems unaffected, the current seems strongly affected by shading. If one cell is 50 or 80% shaded, there's little difference, but if the cell is over 95% shaded the current falls substantially from the entire string. So if, say, a bird sat on the panel and covered an entire cell, the performance would be affected far more than the simple area involved.
Is this normal ? If so, how do commercial systems cope with it apart from ensuring that installations are never shaded at all ?

The answer to that is quite simple.
Commercial panels contain what are called bypass diodes. The panel is divided up into two or more series connected groups of cells and a bypass diode is put in parallel with each group.
The diode is arranged so that when the cells in its group are producing as much current as the rest of the panel they do not conduct any current. But as cells (as few as one) in the group become shaded the voltage on that group goes down as the rest of the panel tries to force current through the shaded cell. At some current level the voltage of the group will drop to zero and start to go negative. At that point the diode starts to conduct, allowing the current to bypass that section of the panel.
The result, if you have three bypass groups, is that you lose one third of the panel voltage (plus a fraction of a volt) but get full current.
The Vmp of the string will change as a result, which is why you do not normally run two strings of panels in parallel into an MPPT input if one of them is subject to partial shading.

[SunEagle]

Answer to your first question is yes even a little shading will reduce the output of solar pv cells.

Answer to your second question is that with micro inverters you might lose the output of one panel (60 to 72 cells) but you would get power from all of the other panels that are not shaded.

Answer to a question you haven't asked. How good are those eBay cells? They are not good for a grid tie system and never will be because you are required to us UL listed solar equipment for a grid tie system.

[adaviel]

The answer to that is quite simple.
Commercial panels contain what are called bypass diodes. The panel is divided up into two or more series connected groups of cells and a bypass diode is put in parallel with each group.

Zener diode I guess ? How many cells in a typical group ?
I was wondering if it was worth connecting say 2-3 spatially separated cells in parallel, to reduce the effect of an object shading all the cells in the group. I'm looking at an installation for a boat where random shading is probable and good performance per square meter desirable. One of my prototypes has pairs in parallel, then in series and it surprised me at first that shading 2 in a row had no effect while shading 2 in a column did.

[inetdog]

Zener diode I guess ? How many cells in a typical group ?
I was wondering if it was worth connecting say 2-3 spatially separated cells in parallel, to reduce the effect of an object shading all the cells in the group. I'm looking at an installation for a boat where random shading is probable and good performance per square meter desirable. One of my prototypes has pairs in parallel, then in series and it surprised me at first that shading 2 in a row had no effect while shading 2 in a column did.

Nope just a standard rectifier diode. Often a Schottky diode to minimize the voltage drop when it is conducting.
The work involved in paralleling widely spaced cells would be prohibitive if using the foil tab and bus wire method. You would be running foil over foil in several layers with insulation in between.
Since it is a DIY panel, look at what it would take to wire it that way, then decide.

[adaviel]

Nope just a standard rectifier diode. Often a Schottky diode to minimize the voltage drop when it is conducting.

Thanks. Finally got around to testing that. I'm getting about 160mV reverse voltage with a 1N5822, 30mA from a 6-cell string with one cell shaded cf. 12mA without the diode (3" cells and a table lamp).

How common is it for panels sold for marine/RV use to incorporate per-cell bypass diodes ? The shading effect seems to be huge.

[Mike90250]

here's an article about using biased FET's as an active diode with better performance than Schottky diodes

http://www.digikey.com/en/articles/techzone/2012/dec/active-bypass-diodes-improve-solar-panel-efficiency-and-performance

[bcroe]

The best way to improve a panel output under random shade, is if possible to have fewer cells associated with a bypass diode; more
bypass diodes. Bruce Roe