If only the moon reflected enough sunlight… Maybe someone should put some mirrors on the moon to focus sunlight at night directly onto my panels…
Haha! I wasn’t asking from a practical perspective, more from an “understanding” perspective.
Talking about strings in parallel:
This is true.
This is an unsafe assumption, the maximum power means exactly that, it will find its own voltage equilibrium, which will tend towards the voltage of the panels producing the most power at the time.
If you put a blanket over the low voltage string the output voltage will be the Vmp of the high voltage string. If you put a blanket over the high voltage string the output voltage will be the Vmp of the low voltage string.
In normal running conditions, the voltage will settle somewhere in between those extremes.
One string doesn’t circulate current through the other, which is what it would need to do to clamp the voltage. ( Unless extreme voltage differences)
Under load, without an MPPT a charging battery would accept current and clamp the voltage, but not necessarily at the maximum power point voltage. This is why we use MPPT’s.
When the battery was fully charged and accepting only minimum current the voltage would rise.
1 Like
But presumably very close to the lowest Vmp? Otherwise surely you will basically get no power from the low Vmp string and the MPPT wants to optimise total watts generated.
2 Likes
No, why?
My understanding is solely based on what Izak said above, that the power generated (assuming you have half your panels at 60V and the other half at 70V) if you try and take more than 60V from the 60V string, will drop rapidly. So it seems reasonable that the maximum power point would be very close to 60V should moving above it result in almost no power from the 60V string.
This makes sense to me intuitively, as current flows from a high potential to a low potential but if your lowest string’s potential is less than where you want the current to go, surely no current can flow (where you want it to flow)?
The story would change if you have only one panel at a low voltage but many at a high voltage.
Again, I cannot reiterate this enough, last time I did physics or chemistry was at high school. I have a reasonably advanced understanding of mathematics and statistics, but that is obviously not really applicable here.
Firstly, we are purely talking about strings in parallel.
I am a bit unclear whether you are switching from series to parallel and back in your reasoning.
The parallel strings are at the same voltage, not one string at 60V and one at 70V.
They are at the same voltage.
They are in parallel they must be at the same voltage, whatever that voltage is.
Ah, I see where the confusion lies:
I want to remind you that a PV panel is a bunch of diodes, not a bidirectional linear resistance.
The characteristics of a PV panel is that it doesn’t pass “*forward” current at its open-circuit voltage(Voc), by definition.
But it also doesn’t pass current in “*reverse” unless the voltage across it is a good few volts higher than its own Voc. So each string presents as an open circuit to the other respectively at close but dissimilar Voc’s respectively.
- I am using forward and reverse, without considering the diode directionality to hopefully match your directional convention.
So the higher Voc string is not circulating current through the lower Voc string. The combined output current of both strings can only circulate through the load. If there is no load, no current will flow.
True. Again, I can easily construct “thought experiments” where I can get one or the other result.
For example, if I have 100 modules that all have an Imax of 10A, of which 99 makes their peak at 45V and one makes its peak at 42V. If I put them all in parallel (which is not exactly optimal, but this is my thought experiment so go make your own!), the MPPT will likely pick 45V. That one module that wants to run at 42V will make no power at all, but 99 x 10 x 45 = 44.5kW, while 100 x 10 x 42 is only 42kW.
Or in short, the higher power to be made at the higher voltage completely obliterates the losses from losing one module.
But conversely, if I do something a bit more sane, like running a 380W module (let’s say 42V at 9A) in parallel with a 450W module (45V at 10A). Well now 19A times 42V is 800W, but if it pushes above 42V it loses the entire 380W and makes only 450W. So in this simple two-module case… the MPPT is almost certainly going to pick 42V if it is worth anything.
Of course these are thought experiments. In practice, there may well be things I have not thought of.
Edit: Well, thinking about that a bit… in the second example it will probably pick something a little above 42V. Wherever the extra power made from the 450W module cancels out the loss on the 380W module, that is where it will settle. But that 380W module in my example is on the “back cliff” of the P/V curve, so for all practical purposes, I would expect around 42V.
1 Like
Not sure where I’ve done that? In my example I had one string with a Vmp of 70 and another string with a Vmp of 60. I connected these two strings in parallel.
So I’m making the case that that voltage will be 60V in a reasonable scenario (for example 6 West plus 6 East panels in parallel on an MPPT)
Sorry, I wasn’t implying that the above was going to happen. The MPPT “requesting” (whatever the technical term would be) 70V as the maximum power point in my previous example would cause the 60V string to provide 0W while the 70V string would provide 70*(x amps). This will very likely be less than 60*(y amps in 60Vmp string + x amps in 70Vmp string) so the MPPT will choose 60V.
Back to my original point that adding (in a realistic scenario) a X Vmp string in parallel with a Y Vmp string will result in min(X,Y) being the resultant voltage at which the panels will deliver their maximum power.
Phil is probably right that that is somewhat of an oversimplification, but unless there is a huge disparity in the power levels of the two strings, probably close enough to accurate. The MPPT will see a massive power drop-off as it goes over 60V (cause it is losing the one string) and it will then “reverse” back towards 60V until it finds a peak. That peak will likely be very slightly above 60V.
2 Likes
This is a typical PV Power vs Voltage curve. You can imagine if you parallel two strings with similar VMPP 's then it will find a point in-between the 2 peaks, but if the VMPP’s are vastly different, then it will pick the lower one, or slightly above the lower one, since at the higher one, the lower voltage string will do nothing.
1 Like
That blue line accurately describes what I called the “back cliff” of the P/V curve (power vs voltage). You go over that, it loses power so quickly that most MPPTs will recognise it as a (local) maximum.
Good MPPTs scan the whole range, perhaps there is another maximum later on. That can happen when there is shading for example.
One optimisation I know about is that if the power drops by more than 30% as you scan upwards and you don’t find another peak… you can stop wasting time and go back to the peak you found.
I don’t know what voltage the MPPT will settle at. What I do know, is that the MPPT doesn’t know that it’s dealing with two discrete power curves. There is no binary choice posed to the MPPT. It is dealing with a combined power curve. It will unknowingly compromise, to optimise the maximum power.
Perhaps this little scenario will make it clearer:
Firstly, I want you to note from the curve kindly posted above that a PV string can achieve high voltage under low power conditions.
It is afternoon and the lower Voc PV string faces east, (it is in low light conditions so low power output), and the higher Voc string is in strong light conditions, it faces West. (higher power output).
The MPPT is oblivious that there are two strings in parallel or their respective orientations.
It just varies voltages searches to find a voltage/ current combo that gets the most power.
In this situation, it will vary the voltage away from the Vmp of the lower Voc panel, because it gets more power by increasing the voltage towards the Vmp of the high power West panels than it sacrifices from the lower power East panels.
The situation will reverse in the morning and the Vmp of combined panels will be lower as the East facing panels will be the major power contributors.
The voltage will tend towards the Vmp of the string that has the greatest influence on the power output. That can be a moveable feast, there isn’t a hard and fast answer, that it will be lower or higher.
It will be the optimum voltage of the combined power curves and that as in my example can easily be both lower and higher at different times in a single day.
1 Like
This makes sense, thanks!
But would you agreed that, if all panels are orientated the same direction, the Vmp will likely sit close to the Vmp of the lowest Voc?
What you said above is indeed why I like Victron’s higher amps based MPPTs more than those forcing you to put all your panels in series. You have more options. Especially when you want to combine different orientations on one MPPT.
Maybe, maybe not, it depends on the relative wattage of the strings, not the voltages.
There are many applications, some suit one thing, some suit another.
A square peg in a round hole, or a round peg in a square hole, one tries to get the best fit for the application.
Indeed. I assumed in all my argumentation that both strings are facing the same way 
When you are also dealing with different directions, or shading, then it could be any voltage between the two. If for example the string with the lower Vmp is entirely shaded, the MPPT will most likely pick a higher voltage (since that is an advantage to the string that’s in the sun, and no disadvantage to the side that’s shaded).
More specifically, I think people ask this question because they want to know what the performance penalty is. Well, the worst case scenario is that everything runs at the lower Vmp. So the maximum performance penalty is roughly the lower Vmp divided by the higher one.
Thank to all who contribute to this answer!
My takeaway:
worst cast, the new string will preform at the specs of the lower string.
No more 405w, i can now only get 420w.
Genade, but the solar panels are flying at the moment! Little to no stock of the popular sizes!!!
Yes, that is a good rule of thumb to use. Unless you give much more detail with regards to orientation, strings, etc. and some of the smart guys here can do some calculations. But I think if you don’t want to go that route based on the rule of thumb, you probably shouldn’t regardless of what more accurate calcs tell you.
I love to make my MPPT’s to the ceiling of their current capability.
I also try and keep them at a high current for as long as possible during the day.
I can use all the power I can get, even if the batteries are fully charged.
To this end, I began experimenting with mixing and matching arrays at different tilts and directions.
I didn’t use fancy brackets, I just worked with the roof surfaces I had.
I have four MPPT’s and 2 PV inverters ( with another 3 MPPTs in them).
I did some combining on the roof, but brought string down on separate cables. I realized very quickly that working with what an on-line solar calculator was hopeless.
Make equivalent voltage strings and actual measurements.
I reckon most people will get 60%-100% more panels on their roof without maxing out their MPPT’s
Obviously the more strings, the more MPPT’s, the more diversity of tilts and directions creates more opportunity to fine tune things.
There is a guy (I think it was in the Northen Cape) that took the East/West and all side panels to the extreme. He has panels mounted on the walls that get the very early/late sun, and then all angles in between. They were all on just a few MPPTs and he got still decent power even in heavy rain just from ambient light on the amount of extra panels. Most MPPTs will limit the current when the panels produce too much power to what they can handle. (Got to love those blue products 
)
Calculator never give you all the results. They just indicate the most efficient most of the time.
1 Like
Artsolar (ZA) has just teamed up with PV giant Longi. Longi are the world leaders in bifacial PV. Apparently bifacial manufacturing is coming to South Africa soon.
I am seriously thinking of bifacial E-W PV fencing, for the very early/late sun.