hard wiring solar panels

Cutting the MC4 connectors will generally invalidate the panel warranty. Not a big issue with the low $$$/Watt prices of most panels this days).

If you go through the panel j-box, make sure you have a good seal. Water in j-boxes will generally take a panel down in something like months-year. Also, when wiring the panels, try to have a drip loop before any j-box. The loop being lower than the j-box entry will drain water off the wire before it can migrate into the j-box seals.

-Bill

Shading on solar panels can drop output from 50% to 100% (yes, 1/2 to zero harvest energy).

A "shaded" solar cell "goes high resistance" (or technically, its current flow goes to whatever the amount of sunlight/skyshine/etc. falls on the cell). Since the cells are in series in solar panels (i.e., 36 cells * 0.5 volt per cell = ~18 volts Vmp)--That means that one cell limits the current from Imp (based on sunlight) to that one shaded cell (1/10th to 1/100th of the current available from the rest of the cells in the series string).

There are different solar panel configuration (36 cells in serie is 18 Volts Vmp; 60 cells in series for 30 volts Vmp, 72 cells for 36 volt Vmp, etc.). And some panels have several strings in parallel (i.e., a 72 cell panel can be 2x strings of 36 cells for Vmp 18 volts at 2x the current per cell/string).

Modern panels all pretty much have "bypass" diodes.. They are in parallel with ~18 to 24 cells or so. If you have two solar panels in series, and one panel has some shade, the bypass diode will let current flow "around" the dark segment.

Solar cells are really just "giant" diodes. And if you put much more than ~12 volts across a shaded (high resistance) "dark" cell, it will ruin the cell (over voltage the reverse biased dark cell/diode). The bypass diodes allow that current to "go around" that dark cell.

The whole question of (for example) for panels in parallel, and one is shaded... That one panel will effectively stop producing Vmp, and its current output will drop to zero. And the other three will perform normally.

If you have the 4 panels in series, and one cell is shaded... Then you will lose (for example if 18 cells per bypass diode), the Vmp of that panel will drop from 18 volts to ~9 volts--And the rest of the string will still produce full current at 9 volts... So that is a 50% reduction of that shaded panel.

But we cannot usually control shade to one cell in a panel (or just allow one column of cells)... Generally the shade falls across both/all strings and that means the whole panel loses output voltage... And you are down to 3 panels producing full voltage and one panel not producing any voltage--So you still have only 3 panels producing energy (assuming the Vmp-array >> the minimum input needed for that MPPT controller + Battery Bank voltage.

In rare conditions, you can make it all work for you.. For example, if you install panels in portrate (sideways) in a snowy environment... The snow collecting on the bottom edge of the panels covers 1/2 the cells in one string---The upper 1/2 of the panel can still be in sun and produce some useful power (again, depending on the rest of the panels series/parallel connections, and the MPPT+Battery bank configurations).

Otherwise, modern mono/poly crystalline panels--There output current is pretty much proportional to the amount of solar energy hitting the cells/panels. In fact, a solar cell/panel driving into a current meter (more or less a dead short) is a pretty accurate solar energy measuring tool.

When solar panels were $10-$30 per Watt--Almost anything you could do to improve harvest was cheaper than adding more panels. With solar panels down in the $1 to $0.50 per Watt range, it is usually cheaper to add 10% more panels than to change from PWM to MPPT charge controller to increase harvest by 10% (mostly in winter--Summer harvest is not "improved" with MPPT over PWM because Vmp-array falls as panel temperature rises in hot weather).

I still am a big fan of MPPT... But this is because they work with "cheap" GT type solar panels (panels that do not "match" PWM to Battery Bank voltage such as Vmp-30 volt and higher panels) and efficiently take the higher voltage/lower current array and down convert to lower voltage and higher current for the battery bank. "Cheap panels" and "expensive MPPT" make for less expensive systems vs "12 volt panels" (more expensive) and "cheap" PWM controllers.

I also like MPPT controllers because they can supply full voltage and current to a battery bank in very hot weather. An 18 volt panel will drop to near 14.4 volts--And you need >15 volts for "full" voltage and current to charge using a PWM controller on a lead acid battery bank to 14.8 or 15.0 volts. If you are in a very hot climate, a PWM controller can be handicapped a bit.

And even if an MPPT controller could produce 2-5% more power vs PWM in low light conditions (i.e., a 100 Watt panel in 10% light would output 10 Watts--5% increase of 10 Watts PWM harvest with an "improved" MPPT harvest would be 10.5 Watts... You would be hard pressed to prove it with standard DMMs and Current Clamp meters (very close to the accuracy limits of the meters and variability of test conditions/temperature/sunlight/battery bank voltage, etc.). I could add a 10 watt panel (10% more array) much easier and probably get more "useful" harvest.

I try to avoid the PWM vs MPPT vs Low Light etc. efficiency arguments. While there are real physical effects and differences--They are typically so small, an otherwise properly designed system--The differences pretty much disappear when compared to overall operation.

These days, I don't even bother with different efficiency numbers between PWM and MPPT controller designs. They are different, but for most folks, it just clouds the other issues (siuch as matching array configuration to controller type to battery bank voltage, proper wire AWG, etc.).

Other issues usually take president... Panel angles (vs trackers vs virtual tracking), getting rid of any shading, choice of batteries and panels, etc. need to resolved to get a reliable/cost effective system.

-Bill