Partial shading of string

[Xplode]

I'm searching around and looking to get a clear answer.

if I install 3 panels in a string, and some shade is sitting on one of the panels, will modern (ie 300w Longi 60cell) panels deal with that? Will the one panel that is shaded drop production but still let the other two panels produce full power? I believe this is the fundamental purpose of the bypass diodes right?

Now, what if there is a parallel string of 3 identical panels connected as well. How does that shade affect production?

Been reading articles but they're all very generic and most of what I found was a few years old.

Thanks!

[Mike90250]

in a series string of panels, as shade progresses across panels, the cells cut out, and the bypass diodes kick in.
bypass diodes are a last resort to keep some sort of production going. After a while, they fail, hence the market for replacement diodes.

As each segment of cells is shaded, about a third of the panel's voltage is lost. if it was a 30V panel, it's now 20V. Another segment shaded, and you are at 10V
and so on....
At some point the voltage drop is enough the controller cant function anymore and all charging stops.

If the panels are in parallel, the diodes still function and as a panel drops to 20V, ALL the panels in parallel are forced to follow that voltage, sometimes dumping power backwards into the shaded panel. That's why there are fuses for parallel panels, to keep fires from starting.

Remember, these are SOLAR PANELS not moonlight panels. Full sun is required.

[Xplode]

Ok that all makes sense, except the fire part. Do the bypass diodes not prevent backflow of the current through the shaded panel's string?

Even if the voltage is pulled lower on the string, I'm looking at approx 114Voc per string on an mppt charging 24volts. So it can drop quite a lot before the controller would drop out.

I've been reading and I've seen the terms "diffuse shading" and "hard shading"
If a panel is just not in direct sunlight, but isn't completely dark, doesn't it generate full voltage but reduced current? Currently my other setup with 2 panels stays at 77V output right up to sunset and then rapidly drops off. They're tilted slightly East, so they're only really receiving indirect light at this point.

[Ampster]

There is another thread that also discussed this. Depending on where you are located and whether you need to comply with NEC 2014 or NEC 2017 the choice may be limited to SolarEdge Optimizers or Enphase microinverters. I am working through the Rapid Shutdown issues with an alternative to the above. Where are you located? Do you know which version of NEC your local AHJ (Authority Having Jurisdiction) follows? Are you grid tied or off grid? Is yours a roof mount or ground mount?

[inetdog]

Ok that all makes sense, except the fire part. Do the bypass diodes not prevent backflow of the current through the shaded panel's string?

Even if the voltage is pulled lower on the string, I'm looking at approx 114Voc per string on an mppt charging 24volts. So it can drop quite a lot before the controller would drop out.

I've been reading and I've seen the terms "diffuse shading" and "hard shading"
If a panel is just not in direct sunlight, but isn't completely dark, doesn't it generate full voltage but reduced current? Currently my other setup with 2 panels stays at 77V output right up to sunset and then rapidly drops off. They're tilted slightly East, so they're only really receiving indirect light at this point.

1. The bypass diodes are in parallel with each bypass group of cells. It is there to conduct forward current instead of trying to force it through the shaded cell group with a voltage which might exceed the breakdown voltage of the cells. Backflow, on the other hand, is generally prevented by the diode characteristics of the cells themselves. The fuses are there to prevent backflow in a damaged panel or string of panels, not in the simple case of shaded panels.
2. If the voltage is pulled lower on one string going into its own MPPT output, you do have a much greater tolerance for voltage loss, as long as the string is not close to the dropout voltage of the MPPT device initially. Panels or strings feeding a PWM device, on the other hand, are likely to be operating close to the dropout voltage. But if two strings share one MPPT input, then a loss of voltage in one string may either cause the unshaded string to operate at a point below its Vmp, with a loss in power. The shaded string will lose only the potential power of the bypassed cells plus the voltage drop of the bypass diode(s). Or, with a different MPPT algorithm, the unshaded string may operate at its Vmp while the shaded string contributes nothing. (The MPPT algorithm having stopped at a local power maximum rather than the absolute power maximum.)
3. As the light on the panel (insolation) decreases, the Vmp and Voc also decrease, but very slowly compared to the decrease in Imp and Isc. A diffuse shading will reduce the current but not necessarily trigger the bypass diodes, while a hard shadow covering an entire cell will definitely force the bypass diode to conduct, dropping the output voltage by 1/3 in the case of three bypass groups.

[Xplode]

There is another thread that also discussed this. Depending on where you are located and whether you need to comply with NEC 2014 or NEC 2017 the choice may be limited to SolarEdge Optimizers or Enphase microinverters. I am working through the Rapid Shutdown issues with an alternative to the above. Where are you located? Do you know which version of NEC your local AHJ (Authority Having Jurisdiction) follows? Are you grid tied or off grid? Is yours a roof mount or ground mount?

I am off grid and further north then the NEC applies to, so I am not worried about these challenges. And this particular thread is more of a learning question for me, as I would like to truly understand the behavior to help in my decision making.

[Xplode]

1. The bypass diodes are in parallel with each bypass group of cells. It is there to conduct forward current instead of trying to force it through the shaded cell group with a voltage which might exceed the breakdown voltage of the cells. Backflow, on the other hand, is generally prevented by the diode characteristics of the cells themselves. The fuses are there to prevent backflow in a damaged panel or string of panels, not in the simple case of shaded panels.
2. If the voltage is pulled lower on one string going into its own MPPT output, you do have a much greater tolerance for voltage loss, as long as the string is not close to the dropout voltage of the MPPT device initially. Panels or strings feeding a PWM device, on the other hand, are likely to be operating close to the dropout voltage. But if two strings share one MPPT input, then a loss of voltage in one string may either cause the unshaded string to operate at a point below its Vmp, with a loss in power. The shaded string will lose only the potential power of the bypassed cells plus the voltage drop of the bypass diode(s). Or, with a different MPPT algorithm, the unshaded string may operate at its Vmp while the shaded string contributes nothing. (The MPPT algorithm having stopped at a local power maximum rather than the absolute power maximum.)
3. As the light on the panel (insolation) decreases, the Vmp and Voc also decrease, but very slowly compared to the decrease in Imp and Isc. A diffuse shading will reduce the current but not necessarily trigger the bypass diodes, while a hard shadow covering an entire cell will definitely force the bypass diode to conduct, dropping the output voltage by 1/3 in the case of three bypass groups.

Ok, that makes a lot of sense. Some of that I'm gonna have to re read after my morning coffee to fully understand what you're saying.

If I'm understanding the bypass diodes better now, it seems like I could add a diode on each strings at the combiner to prevent any backflow. That should prevent any backflow issues if I discover undesirable behavior.

In this scenario, i'd have the choice to use two Tracer BN controllers or one Outback FM60/80. So better controller or better configuration?

[littleharbor]

I believe what the bypass diodes are trying to accomplish is to keep the current of the panel consistent with it's rated current so that, in a series string, it doesn't pull down the whole string. As long as the shaded panel can continue to produce consistent current, albeit at lower voltage, the only loss is a minimal amount of voltage from the entire high voltage string.

[Mike90250]

The Midnight charge controllers had a ICE MELT feature where high voltage could be applied to the panels to create enough reverse current through dark panels, to heat and melt ice, in theory. Turned out the voltage needed to backfeed was tougher to obtain than expected.

[inetdog]

Ok, that makes a lot of sense. Some of that I'm gonna have to re read after my morning coffee to fully understand what you're saying.

If I'm understanding the bypass diodes better now, it seems like I could add a diode on each strings at the combiner to prevent any backflow. That should prevent any backflow issues if I discover undesirable behavior.

In this scenario, i'd have the choice to use two Tracer BN controllers or one Outback FM60/80. So better controller or better configuration?

Modern silicon-based panels typically will not conduct in the reverse direction at the voltage levels they would be exposed to. I do not know of anybody who would recommend the use of blocking diodes. Remember that the fuse is there to protect a damaged panel or string. The energy loss from the forward voltage drop across a blocking diode is small ( ~1V times the string current), but unnecessary. It is just one more component to fail.

[Xplode]

I don't like the idea of adding a diode for that very reason. More custom DIY means more failure and complexity down the road. The "modern" is why I was wary of older writeups I had found.

So fusing the parallel strings, do I aim for 10A fuse (just over the 9.x Amps the label should produce) or do I fuse for wire size? I don't really understand how a fuse is going to protect when the power produced can't really ever exceed the wires' rated capacity.

Appreciate all the input!

[inetdog]

I don't like the idea of adding a diode for that very reason. More custom DIY means more failure and complexity down the road. The "modern" is why I was wary of older writeups I had found.

So fusing the parallel strings, do I aim for 10A fuse (just over the 9.x Amps the label should produce) or do I fuse for wire size? I don't really understand how a fuse is going to protect when the power produced can't really ever exceed the wires' rated capacity.

Appreciate all the input!

The fuse is not there to protect the wire, it is there to protect the (failing) string. The panel specifications will list a maximum series fuse size. Follow that recommendation, and know that the fuse is not likely ever to fail. But keep spares on hand anyway. Make sure the fuse and holder you get is rated for DC at the string voltage. Standard glass fuses are often rated for AC only or for a relatively low DC interrupting voltage compared to the AC rating.
Along those same lines, fuses or breakers on the AC side of grid tied inverters are not there to blow from excess current coming from the inverter, they are there to protect against utility current feeding back into a shorted wire or failed inverter.