Solar PV strings in parallel, blocking diodes or not.

[SpeedyDave]

**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

[NiHaoMike]

Don't even need the blocking diodes under realistic conditions, unless the charge controller doesn't have that built in.

[SpeedyDave]

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.

[SpeedyDave]

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?

[bdunham7]

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.

[SpeedyDave]

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.

[Someone]

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.

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.

[digsys]

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 :-) )

[f4eru]

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.

[The Soulman]

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

[SpeedyDave]

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"

[SpeedyDave]

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.

[Kleinstein]

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.

[bdunham7]

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.   

[Seekonk]

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.

[fourfathom]

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.

[jnissen]

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. 

[bdunham7]

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.

[Seekonk]

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

[bdunham7]

Common misconception?  What does the forward voltage (drop, I assume you mean...) of the cells have to do with blocking diodes???

[fourfathom]

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.

[bdunham7]

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.

[SpeedyDave]

 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 ??

[SpeedyDave]

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.

[Kleinstein]

.... 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.

[bdunham7]

 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 ??

Across the diodes.  But as Kleinstein says, the MOVs are probably overkill and are just my idea.  One diode and a fuse is also enough, with an MOV if you are as concerned about lightning as I am.  My main issue with these HV strings is that you don't always know when failures occur--you don't go up on the roof with a multimeter every month and test stuff--so critical parts, especially safety related, need to be very robust.

[SpeedyDave]

**Update**

Things are progressing and I have acquired a few panels, diodes, breaker etc.

Let the fun begin !!

Here is how I have decided to include blocking diodes for each string.
The way that the DC breaker (4 pole) is configured, I am going to use 2 diodes per string.

Diodes are 1000v 10A

Looking at the picture, the string positive enters the first breaker, passes thru the diode, then the next (second) breaker and onto the +ve input of the inverter.

The string negative enters into the fourth breaker, passes thru the diode, then the previous (third)  breaker and onto the -ve input of the inverter.

The strings on my system are configured exactly the same, but have buss-bars connecting the two breakers together.

I will also replace those buss-bars with diodes.

Easy to monitor and no fire hazard.

Should work a treat !


*note.  I was speaking to a solar installer the other day and he showed me some inline blocking diodes ( modular with MC4 connectors each end )  and he was like "  yeh, so what, use them all the time "  !!!

[Technomaniac]

Hi.  If you are talking about a diode in the conducting direction in series with each string of panels, the panel producing the higher voltage will reverse-bias the diode in series with the low voltage producing panel string. IT WONT CONDUCT AT ALL. So effectively you are switching off the weaker panel altogether. If you want all panels to contribute all day, you can't use series diodes. I have a 24v system and I have experienced this. I got around it by using separate regulators and battery pairs (12v car batteries) on each panel, or parallel groups of panels, and then bringing the groups together through series Schottky diodes.Even then, only the battery pair with the highest voltage is in use - same theory, reverse-biased diodes dont conduct. .But I have all panels horizontal, 8 feet off the ground, as I'm in a dusty area and have to clean them at least once a week. Incidently, if any readers of this are doing the 24v thing like me, put some reverse diodes across each 12 volt battery. It took me a long time to work this out, second hand batteries, stronger one is reverse charging the weaker one through the night (if the batteries are being discharged more than is healthy.)
We had a retirement home fire started in a panel on its roof a couple of weeks ago. I don't like so many panels in series, if one fails (its only a coat of semiconductive paint) then you have the 400 volts to jump across the gap and because its DC the arc keeps right on. I read once that when Edison had 110v DC mains around the houses, if a lamp hanging from the ceiling would blow, sometimes an arc was created and it would move up the wire to the ceiling and along the wire, burning down the house.

[SpeedyDave]

Nearly ready for final connections and testing.
I understand that reversed bias diddes dont' conduct, but my theory is that both strings will be contributing due to the fact that  PV panels are a high impedance current source. ie the voltage varies according to load.

In my existing system, OCV is roughly 440V, but that drops to around 330V when loaded by the inverter setting the MPPT load.

If a string is not under load, its voltage will rise to its OCV spec.
This will then be higher than the LOADED voltage (MPPT determined) of the second string, and as such, the diode will conduct and start suppling current.
This load will pull the voltage of the first string down to be the same as the second string, so that BOTH strings will contribute, although I recognise that they will not be contributing equally.

Once my setup is ready, I can easily add an extra panel to one of the strings ( 11 on one and 12 on the second ) and physically measure what is going on.
I am confident that my theory will be proven correct.

[bdunham7]

Hi.  If you are talking about a diode in the conducting direction in series with each string of panels, the panel producing the higher voltage will reverse-bias the diode in series with the low voltage producing panel string. IT WONT CONDUCT AT ALL. So effectively you are switching off the weaker panel altogether. If you want all panels to contribute all day, you can't use series diodes.

The blocking diode will only be reverse-biased and non-conducting if the MPPT (loaded) voltage of the first array is higher than the OCV of the second.  And in that case, you still would want the blocking diode because without it the second array would absorb power from the first.  In a small 24V system as you mentioned, with similar panels and a very light load, I suppose it is possible that the differences could be the same or even less than the forward voltage drop of the diode, but that would be best remedied by increasing the load and lowering the voltage a bit.  Any competent MPPT device should do this.

[SpeedyDave]

***update***

Wiring done.

some system tests etc over the next few days to get some base line numbers etc. , then a full system test planned for Sunday.

Once base line numbers are done, the big theory test will be on Monday, where I will test 2 strings (east and west) with two differnt outputs ( 11 panels vs 12 panels ) thru blocking diodes, feeding into one MPPT at the inverter.
I can then directly measure voltages at the breaker, either side of the diodes, of both paralleled strings.

If my theory is correct, the loaded voltage out of each string will be the same ( current will be different of course )

[SpeedyDave]

The numbers are in !

Data valid as of 8:30am, 22 degrees C, clear skys.  ( both strings are at 20 degrees inclination)

Due East facing string of 11 panels,  OCV 455v
Due West facing string of 12 panels, OCV  495v

Parallelled, thru blocking diodes into a single MPPT on grid tied inverter.

Both panels loaded voltage now reads 358v each ( obviously, as they are parallelled ! )


Volt drop over the blocking diodes.  about 0.8v on the east string, about 0.7v on the west string. Yep, sounds right as the current flow is differnt.


current draw,
East string = +4.2 amps

West string = +2.3 amps


So there we have it, my theory is correct 
Two enirely different strings, facing different directions can positively contribute to a solar PV grid tied installation. 

[SpeedyDave]

Actually fascinating to watch  the MPPT adjust the load on the east/west string to regulate the total output of the inverter (5kW)

Very clever stuff 

[bdunham7]

What are the system voltages during that long flat spot?

[SpeedyDave]

Mppt raises voltages (reduced current draw) to about 410v

Under full load (earlier) the voltages are pulled down to about 340v. (higher current though)

So this is how the inverter manages its max output, even when being overdriven.

Clever stuff !

[sibianul]

Hello everyone,
Dave, thanks for you posts, I'm in almost the same situation as you, in the future I want to connect 2 different strings (EST String has 7 x 290w mono panels, SOUTH String has 7 x 295w mono panels).
Those strings are now connected each to a different ONGRID inverter.

This year I decided to try an OFFGrid system, I purchased an MPP Solar 8048 MAX inverter (8kw) and a Hyundai Ioniq Lithium Battery (28kwh), I also mounted a third String facing WEST, 10 x 370w Mono TrinaSolar Panels. The inverter has 2 MPPT inputs, each 4000w and max 18A. TrinaSolar string will be connected on MPPT1 and if that string won't be enough to recharge the big Lithium battery, I will extend the SOUTH string wires (~45meters) and connect it to the second MPPT on the offgrid inverter.

But I was wondering if it will be an even better idea, at least in the winter, to connect the EST and SOUTH strings in parallel, and than connect both to the second MPPT input of the offgrid inverter.

Regarding your picture, I don't quite understand how you connected two strings in parallel with that breaker, but I want to ask you regarding the heat, do those diodes get hot ?

I wanted to buy MC4 connectors with embeded diodes, and also an Y MC4 cable to join the 2 strings, as I want to join the strings on the roof, under the pannels, and directly from that point to extend the cable, using a thicker one: 10mm² compared to the existing  one which is 6mm² . Would this work too ?

The strings have each maximum 9Amp, should I get a 10Amp Diode MC4 connector, or a 15Amp diode connector, to be on the safe side with that diode?



Bellow you can see a drawing with my strings placement, and how I would extend  the cables to reach the OFFGRID INVERTER, that is placed inside my garage. The ONGRID inverters (green and yellow) are IP65 and are placed outside on my terrace. I'm not able to use a shorter cable from East adn SOuth string to the offgrid inverter, as there are windows all around the house, I can only go under the roof tiles, and on the back of my house, where I already cut the termoisolation of the house to insert a big tube inside wich I have the DC cables.



Any advices, from anyone, are welcome.

Thank you

UPDATE. Is this the way you connected your string in paralel ?

[SpeedyDave]

Lots to go thru, but sounds like you are on the right track.

In my system, the diodes get warmish.  Maybe 50 degrees C ?

The end point of the parallel strings is at the MPPT at the inverter ( inverter has 2 sets of inputs for each tracker )

on my inverter (sunny boy) ALL of the negative inputs are commoned within the inverter.

The 4 pole breaker is a specific solar breaker ( magnetic arc quenching) and it is polarity conscious, meaning that it has +ve and -ve terminals.

i'll try to explain; 

the +ve from the PV string enters the (first of 4 ) breaker at the top left. current flows thru the (closed) breaker down to one diode, thru that diode and then back up thru the next  ( second of 4) breaker to the +ve input on the inverter.

The -ve from the PV string enters the (fourth of 4) breaker at the top right. current flows thru the (closed) breaker down to the second diode, thru that diode and then back up thru the next  ( third of 4) breaker to the -ve input on the inverter.

I have 4 strings of PV

I have 4 x 4pole breakers ( 2 diodes per breaker, one on the positive lead and one on the negative lead ) 

Yes, I could have just used one diode per PV string, but it was easy to use two as the bridges between the positive pair and the negative pair of breakers.


The current sharing "magic" happens when the inverters MPPT loads up the PV circuit thus dropping the voltage. When the loaded voltage goes below the open circuit voltage of the SMALLEST PV string, it will be able to forward bias its diodes and thus contribute power to the system.  So obvioulsy the two strings that you want to parallel need to be fairly close in OCV.

On my system, I have noticed MPPT voltages between about 280 - 380 volts, so as long as the OCV of each string is higher that that, they will both contribute.

Now I can almost hear the naysayers keyboards clattering !


YES !!!   I do know that this is not the most efficient way of harvesting solar power.
YES !!! individual MPPT's or optimisers or even inverters for each string WILL harvest more power !!!!!


but that was not the intention of this experiment   ( i got a bunch of second hand panels and had a spare input on one MPPT on my inverter. this was an experiment to see HOW to parallel two different PV strings, that are on different orientations into one MPPT on my inverter)

On my system, I have gone from a daily max of about 40kWh per day to about 47kWh per day.
The added PV string was 2.2kWh and daily irradiation of 6hrs.
So the potential of 12kWh ended up being 7kWh, these arew west facing panels anyway so not a bad outcome.
Here in Australia, we are limited to a 5Kw inverter anyway, so now, instead of having a 5Kw peaking bell curve at midday, I have a nice big 5Kw flat-top curve from about 9:30 to 3:00 PM

great outcome for a couple hundred bucks in second hand panels.

[bdunham7]

I have a nice big 5Kw flat-top curve from about 9:30 to 3:00 PM...
great outcome for a couple hundred bucks in second hand panels.

Excellent!  I think more and more systems will go to this 'clipped' model as the panels get cheaper and more people install them.  It's not just your inverter or your panel that limits the practicality of peak power, it is the grid itself.  Google "CAISO duck curve" and see what I mean.

[sibianul]

In my system, the diodes get warmish.  Maybe 50 degrees C ?

Ok, not to hot at 50deg C

The end point of the parallel strings is at the MPPT at the inverter ( inverter has 2 sets of inputs for each tracker )

on my inverter (sunny boy) ALL of the negative inputs are commoned within the inverter.

I think its normal to have the negative input common inside the inverter, all GND are common, no matter the POSITIVE voltage as I know. In my MPP Solar 8048 MAX inverter the GND are common too, and I think I opened an Growatt inverter and it also had common GND

I have 4 x 4pole breakers ( 2 diodes per breaker, one on the positive lead and one on the negative lead ) 

This made everything clear to me Now I understand your setup, you have diodes on negative wire also, I think I will use an diode on positive wire only, would you think this might be a problem ?

Regarding the diodes, anyone knows what are the disadvantage of using a bigger diode than the current generated by my string? My panels have a maximum 9A current, will it be more safe to get a 15A or 20A diode, instead of the 10A ? Is there any disadvantage going with bigger diodes ? Is the loss trough the 20A diode much higher than the loss trough the 10A diode ?

Thank you

[SpeedyDave]

you have diodes on negative wire also, I think I will use an diode on positive wire only, would you think this might be a problem ?

Regarding the diodes, anyone knows what are the disadvantage of using a bigger diode than the current generated by my string? My panels have a maximum 9A current, will it be more safe to get a 15A or 20A diode, instead of the 10A ? Is there any disadvantage going with bigger diodes ? Is the loss trough the 20A diode much higher than the loss trough the 10A diode ?

Thank you

One diode on the positive would be fine.
If your strings are 9 amps max, then a 10 amp diode is a bit close to max, so a bigger ( 20 amp ? ) would offer a margin of safety.  ( no real difference in "losses" in a bigger diode , just cost )

My strings max out at just over 5 amps, hence the 10 amp rated diodes ( 1000 volt blocking rated )  were plenty ( and cheap )

And yes, I only used 2 diodes ( one in +ve and one in -ve ) per string because             a; I had spares        and       b;  it was an easy and neat way of bridging the two beakers together.



The circuit breaker I used is this one  https://www.ebay.com.au/itm/124216950191?hash=item1cebe83daf:g:hKgAAOSwm3BazuIn

The diodes were these https://uk.rs-online.com/web/p/switching-diodes/7512709/   but they have been discontinued.

This would be fine, but you can choose what suits you !   https://americas.rsdelivers.com/product/on-semiconductor/ud1006fr-h/on-semi-600v-18a-diode-2-pin-to-220f-ud1006fr-h/1867562

Diodes must be rated to withstand the full (reverse)  voltage of your string, as well as to safely pass the full max current of the string.
So just figure out what those two values are and add say a 50% safety margin (or a bit more )  and then select a diode that can handle those values.