-
**sorry for the cross post, but I thought my initial post would get better traction in this thread ***
Dave,
Would it be possible to do a EEVblog video on the how and why of paralleling solar PV "strings" that would have different orientations (i.e. one string east and 1 west) connected to the same MPPT input on the inverter.
Some info states that the strings must be "exactly the same number and spec" panels, or on a roof angle less than 15 gerees, or exactly east and west etc etc.
I do not believe it!
Given that the inverters MPPT basically is a variable voltage sink, where it clamps its input voltage (the PV output voltage) and then just lets the amps flow, and given that the PV panels/strings have a relatively high impedance, i reckon that two strings of different open circuit voltages on different orientations can easily be paralleled into on MPPT if each string is fed via suitable blocking diodes, as each string voltage will rise to the point of forward conduction of its blocking diode. Yea, the current flow will be different between each string, but I believe that each string would contribute to the total power in the same way that paralleled caps of different sizes do.
I would REALLY like your highly qualified technical opinion on this ! ( this could also be highly relevant to solar pv design / install people, as it would allow for more diverse installations )
Cheers
Speedy -
Don't even need the blocking diodes under realistic conditions, unless the charge controller doesn't have that built in.
-
Thanks NiHaomike.
The video you posted clearly shows that a lower performing panel in parallel with a another panel is indeed reversed biased, and definitely requires blocking diodes to protect the cells.
As I said, I believe that each string of panels ( that are exposed to different levels of solar irradiance) will generate the MPPT (set by the MPPT in the grid connected string inverter) determined voltage, and the current will be determined by the irradiance on each PV string.
( note, I am referring to a typical domestic solar PV installation that would typically have 10 - 12 panels connected in series to create a "string", with a typical MPPT set voltage of ~ 350V - 400V )
So.
If we have one string of 10 PV panels due east and one string of 10 PV panels due west, connected in parallel, with blocking diodes, and it is 10am, I would expect that the east facing string will be outputting at basically full power ( say 350 Volts and 8 Amps ) whilst the west facing string will be outputting much less, as it is not in full direct sunlight. Even so, the west facing string will still make 350V easily, just not many Amps.
Even so, I THINK the west facing string will still contribute SOME power to the inverter.
The balance of power from eats to west will shift to the west panels as the day progresses into the afternoon.
The diode blocking voltage would only need to be lowish (?) as a shaded string is still able to produce Voltage, but the current would be very low.
Hopefully Dave can set the story straight with some actual technical info. -
And therefore, if we extrapolate this out, can we use two completely different strings of panels (within reason) , connected in parallel (via blocking diodes) and still get a satisfactory result?
-
And therefore, if we extrapolate this out, can we use two completely different strings of panels (within reason) , connected in parallel (via blocking diodes) and still get a satisfactory result?
What do you call satisfactory?
The concept is simple enough--if the MPPT voltage in-use exceeds the OC voltage of the weaker string, it will be reverse biased, need blocking diodes and will produce no power. This wastes the power it could have been producing. If the MPPT voltage is less than the OCV of the weaker string, it will contribute some power but will suffer an inefficiency that you can calculate easily enough at the panel level, but unless all the panels in each array are pretty close in both lighting and efficiency, I don't think you can easily quantify it for a string without actually testing (perhaps by switching off the other string).
Just from your description and my limited experience I think the output of a system built this way will be significantly lower than a more efficient system with a dual-input inverter, dual inverters or microinverters. My question is why do you want to do it this way? Is it a system cost issue? There are already many "diverse" systems installed with varying orientations or varying shading (mine, for example, before some tree trimming) that work fine. -
My question is why do you want to do it this way?
Good question.
2 reasons.
1. To broaden my knowledge and understanding
2. my personal circumstance.....
I have a 5.7kW system ( 5.7 PV and 5kw dual MPPT inverter)
My system is now 8 years old.
I have 2 strings.
String one is 2 strings of 10 panels in parallel, all grouped together and facing north.
String two is 1 string of 11 panels, grouped together, facing east.
There is no spare roof space available north or east.
I would like to drive the inverter a bit harder and flatten out my daily production bell curve. (and increase my daily output)
My daily production is biased to late-morning due to the east facing string. Power production generally peaks around 11am and then tapers off as the east facing string gets less and less direct sunlight.
So the obvious solution is to add panels onto the west facing roof, and parallel them onto the second MPPT ( with the east string = 11 panels )
The current "advice" is that this can be done is using the exact same panels and number of panels. (( i.e. 11 panels the same as what I have (suntech 195W ))
The issue is that these panels are not available any more, and I still think that blocking diodes will be required to prevent reverse biasing the cells if there was any shadowing.
So, my options are to replace ALL of the panels, so that the 2 paralleled east and west facing strings are exactly the same
OR
just create a second string that is near enough to the OCV of the existing string of 11, add a couple of dollars worth of blocking diodes and BAM, I will have string #2 (east west) producing ~2kWh for nearly the whole day, instead of peaking at 11am and then tapering off to not much by mid-arvo.
Hope that makes sense.
-
just create a second string that is near enough to the OCV of the existing string of 11, add a couple of dollars worth of blocking diodes and BAM, I will have string #2 (east west) producing ~2kWh for nearly the whole day, instead of peaking at 11am and then tapering off to not much by mid-arvo.
Thats as simple as it gets but there is a surprising overlap in insolation between the east and west facings of typical roofs. Both panels will be illuminated substantially at the same time. Do you have any daily data for the string at the moment? And when you say east-west how close exactly? They need to be matched reasonably well for simple MPPT systems to work which may be the advice you are getting.
Hope that makes sense. -
Every combination of coupling - whether series, parallel with diodes, orientation etc will LOSE efficiency. The amount depends on the mismatch and can be quite high !
The ONLY interconnection method is with a micro-inverter on EVERY panel or panel segment. That is because the relationship between the "power generator" and the "storage system" (or inverter) is tightly matched to create an MPPT profile. Micro-inverters then handle the "differences" by synchronous adjustment (via direct data link to each other) and achieve MAX I/O transfer. You'll find ALL newer "competent" installers now use micro-inverters. We have been doing that on our solar cars for 30+ yrs, for that very reason. It's taken a while for home installs to catch up - but then, there's a lot of bad practices in that area (which is another long story :-) )
-
Yep.
Two different panel strings, with different angles -> two MPPT (at least) needed.
if you don't do that, you lose up to 50% of peak power. -
Should be fine, SMA calls it "polystring", I don't know if they're using blocking diodes but they probably do.
http://www.smainverted.com/files/2012/12/Article_Polystring_East-West-Orientation-TEN122510.pdf -
Thanks all.
Yes I do understand that I will lose some efficiency.
If I have 2 x 2.5kW strings, and face them east / west, I know that I will never get a peak output of 5kWh.
BUT, a string will only ever output maximum when solar irradiance is optimum (ie at right angles to the panel in the x & y axis, and this typically only happens for a very short period each day.
By having 2 strings of different orientation paralleled, that power peak will be higher than the 2.5kW(of each string) and be above the 2.5kWh of a single string for a longer period of time.
A bit like being "tracked" -
Should be fine, SMA calls it "polystring", I don't know if they're using blocking diodes but they probably do.
http://www.smainverted.com/files/2012/12/Article_Polystring_East-West-Orientation-TEN122510.pdf
Thanks for the link.
It looks like the theory is correct, I just need to understand how this works (real world) from an electrical perspective. The actual theory of parallelling high impedance variable voltage /variable current sources into a variable low impedance load.
I am pretty sure blocking diodes of sufficient power rating will work fine.
I am sort of sure that they do not need to have a high reverse bias voltage rating as both strings will be able to output near OCV even when in shadow, though obviously there will be little power(amps) generated. -
It probably helps to have the blocking diodes in series for the east / west modules. Worst case (e.g. early in the morning) the east string may produce something like 80% of the OCV, while the west facing panel is still near zero. The details depend on the cells used. Especially polycrystalline cells may not work well at low light and really go down on voltage.
So the blocking diode should still be good for the full string voltage (one string). Anyway higher voltage rating for the diodes is usually not that expensive.
The efficiency will be a little lower - how much depends on the cells.
With the cells now relatively cheap it can still make sense not to buy another inverter. -
I would like to drive the inverter a bit harder and flatten out my daily production bell curve. (and increase my daily output)
My daily production is biased to late-morning due to the east facing string. Power production generally peaks around 11am and then tapers off as the east facing string gets less and less direct sunlight.
So the obvious solution is to add panels onto the west facing roof, and parallel them onto the second MPPT ( with the east string = 11 panels )
The current "advice" is that this can be done is using the exact same panels and number of panels. (( i.e. 11 panels the same as what I have (suntech 195W ))
The issue is that these panels are not available any more, and I still think that blocking diodes will be required to prevent reverse biasing the cells if there was any shadowing.
So, my options are to replace ALL of the panels, so that the 2 paralleled east and west facing strings are exactly the same
OR
just create a second string that is near enough to the OCV of the existing string of 11, add a couple of dollars worth of blocking diodes and BAM, I will have string #2 (east west) producing ~2kWh for nearly the whole day, instead of peaking at 11am and then tapering off to not much by mid-arvo.
Hope that makes sense.
Yes, if total production and efficiency are less important than extending your production into the evening, then you have the right idea. As the panels get cheaper, these less-efficient and sometimes curtailed installations make more sense. And if you do it on the cheap, it makes even more sense.
In your case, I wouldn't worry about "matching" the west array to the east. If it is more powerful and even if it has a higher OCV, it will never actually take away performance from the east array, it will simply take over when it overpowers it. IOW, in your parallel-with-diodes setup, a larger west array will always give you the same or more combined power than with a smaller one. So I would put the most and largest panels you feel comfortable with buying (limited by your inverter, of course) on your new western array. Keep in mind that once the panels are there, you can always upgrade your inverter or add an extra if you decide you want greater efficiency.
One more thing--you mentioned earlier that you could use a "lowish" voltage blocking diode. To the contrary, I think in parallel systems these blocking diodes should be extraordinarily robust. Significant reverse current may be a fire source and solar panels are subject to surges, stray voltage from lightning, etc. If there is a commercial product with the right connectors, use that. If I were making my own, I would use 2 diodes rated at least 40A, 1000V, in series with a high-energy 10A 600V minimum fuse between them and robust 600V MOVs across each one. That may seem extreme, but it isn't that expensive. Keep in mind that if your blocking diode fails short, you will have the power being produced in one array dissipated in the other--and not necessarily uniformly--even if the power is turned off. -
I thought that extremely poorly done video proved that blocking diodes are bullshit. Yet, people were left with the idea that they were needed. I have 1500W of panels facing every which way going into a 500W charge controller. Panels even with minimal ambient light still produce the same open circuit voltage.
-
I thought that extremely poorly done video proved that blocking diodes are bullshit. Yet, people were left with the idea that they were needed. I have 1500W of panels facing every which way going into a 500W charge controller. Panels even with minimal ambient light still produce the same open circuit voltage.
I sort of agree with you. I have three identical panels on my boat, and they used to be just tied in parallel. Being on a sailboat, one or more are usually shaded, so the current from each can vary wildly. But as you point out, the voltage remains pretty constant. And, even a completely shaded panel will not draw any appreciable reverse current when in parallel with a similar fully-lit panel. I've measured this myself.
But there is a potential issue that might warrant blocking diodes: Panels have been known to fail with shorted cells. In this case, the reverse current in a paralleled system can become significant and perhaps dangerous. The blocking diode will prevent this (uncommon) problem.
In any case, to successfully parallel panels, they need to have the same number of cells. Not all panels do.
Now, I have individual MPPT controllers for each panel. It's more of a science experiment, but I do seem to get slightly more power than when the panels were in parallel. -
Every combination of coupling - whether series, parallel with diodes, orientation etc will LOSE efficiency. The amount depends on the mismatch and can be quite high !
The ONLY interconnection method is with a micro-inverter on EVERY panel or panel segment. That is because the relationship between the "power generator" and the "storage system" (or inverter) is tightly matched to create an MPPT profile. Micro-inverters then handle the "differences" by synchronous adjustment (via direct data link to each other) and achieve MAX I/O transfer. You'll find ALL newer "competent" installers now use micro-inverters. We have been doing that on our solar cars for 30+ yrs, for that very reason. It's taken a while for home installs to catch up - but then, there's a lot of bad practices in that area (which is another long story :-) )
100% agree. I went with micro-inverters for this exact reason many years ago. Recently had a single panel show reduced output on really hot days and that would have been impossible to diagnose easily. Still working it out as it's under warranty but seems to be panel related and not micro-inverter related. -
I thought that extremely poorly done video proved that blocking diodes are bullshit. Yet, people were left with the idea that they were needed. I have 1500W of panels facing every which way going into a 500W charge controller. Panels even with minimal ambient light still produce the same open circuit voltage.
I didn't bother watching the video, but apparently the reason for needing blocking diodes in HV systems hasn't emerged...
The cells in the panel ARE diodes, more or less. I don't have data on what their PIV and leakage numbers are, but I'll bet they are neither good nor consistent. I think the PIV might be 12V or so, leakage highly variable. Just like using multiple diodes in series for higher voltage doesn't work well unless you balance the leakage, so applying reverse voltage to a long string of cells is going to result in the greatest voltage across the cell with least leakage, until it breaks down and then the next one, and so on. You might get away with this on a small system--how many cells do you have in series?--but on a larger system, even with the same model panels in the same orientation, you are betting heavily that nothing goes wrong. -
This is a common misconception. It isn't a case of PIV at all. All those cells have a forward voltage. If you want to melt snow off your panels, just provide a forward voltage sufficient to get them conduction. This voltage is higher than they can produce so it isn't a factor normally. As long as you don't supply more current than the panel is rated for it is safe. Just like putting a bunch of diodes in series and passing current thru them. Of course, if you have one entire panel short out the bypass diodes as in a lightning strike there will be a problem. I had a hit and had to replace several diodes.
The only point of that video was to make a video and get someone to watch. That is the purpose of almost every youtube video. That and to buy something from his store. He has no serious knowledge of electronics -
Common misconception? What does the forward voltage (drop, I assume you mean...) of the cells have to do with blocking diodes???
-
Common misconception? What does the forward voltage (drop, I assume you mean...) of the cells have to do with blocking diodes???
I'm not sure I understand your question, but look at the equivalent circuit of a solar cell:
That diode starts conducting at around 0.5V (it's kind of squishy), and this is what establishes the cell open-circuit voltage. If you put a voltage on the panel that exceeds the sum of the cell diode drops then reverse current will flow. The series blocking diode prevents this reverse current.
Note that we have been discussing blocking diodes, not shunt diodes. -
That diode starts conducting at around 0.5V (it's kind of squishy), and this is what establishes the cell open-circuit voltage. If you put a voltage on the panel that exceeds the sum of the cell diode drops then reverse current will flow. The series blocking diode prevents this reverse current.
Note that we have been discussing blocking diodes, not shunt diodes.
OK, he's right, it's not PIV. Doh! It's the bypass diodes that deal with the PIV! But the need for blocking (not bypass) diodes is still there on parallel strings. The fact that the panel can absorb some reverse current and convert it to heat is not adequate protection against faults on high power systems. I realize it is commonly done this way on small-to-medium systems, but that doesn't make it a good idea. -
If I were making my own, I would use 2 diodes rated at least 40A, 1000V, in series with a high-energy 10A 600V minimum fuse between them and robust 600V MOVs across each one.
Thanks
could you please explain exactly where you would connect the MOV's ?? -
I thought that extremely poorly done video proved that blocking diodes are bullshit. Yet, people were left with the idea that they were needed. I have 1500W of panels facing every which way going into a 500W charge controller. Panels even with minimal ambient light still produce the same open circuit voltage.
@6:40 in the (poorly done) video, the shaded panel is reversed biased and drawing current from the unshaded panel. Blocking diode would prevent this. -
.... Panels even with minimal ambient light still produce the same open circuit voltage.
They will usually not produce the same open circuit voltage, but it may not take much light to get as much open circuit voltage as the other panel has at the MPP, that is under load. Here it also helps that the cells in the sun usually are hotter and this reduces the voltage.
The details depend on the type of cells. Especially polycrystalline cells may not behave very much like an ideal diode.
There is another aspect to the diodes: they also help if the panels are slightly different, or something like a some cells (or a shunting diode) fail short.
I don't see much need for adding MOVs - they are more like a fire hazard on there own. Except for lightning strike, there is not much chance to get over-voltage spikes.
With a low voltage system die extra loss from the diode drop can be more than saved from no applying the voltage to the second (shaded, less sunny) panel.