String inverter - one array vs multiple arrays (pros vs cons)

[arf88]

So i got confirmation today that it was wired correctly from the start on this new inverter to two different strings but for some reason the portal showed only a single logical string. They did something in the portal to fix this and now it shows two different logical strings. But here is the connection and its right there are multiple strings.

I thought the portal and inverter would automatically relay this to the servers, guess not? IMG_20180725_174332549_HDR.jpg

[ButchDeal]

Now, the Solaredge may indeed not do that; I don't know how it handles faults like that. But given that installers may occasionally make errors (and as exemplified by this thread) it seems like cheap insurance. It's a lot easier to have an installer fix a wiring error than to replace a damaged inverter.

There are different optimizers with different limits. Maybe you should get a jobworking for an installer manufacturer they might benefit from your insight.
maybe enohase could use you. Promisse to let us know which inverters you worked on....

[jflorey2]

Most have a specification on the Isc of the array (short circuit limit). Some list max wattage but what they are really trying to limit is usually the Isc.
An array of 1GW is going to have a very high Isc even when Voc and Vpp are within range (will also have one hell of a combiner) and arc fault is going to be a problem.

Agreed. But that's a problem with array protection, not inverter operation.

The 5 watt converter in my phone. It is feed from a wall wort which has about 5 watt capacity.

Your 5 watt wall wart is connected to a source that can give it 1800 watts continuous and several kilowatts for a few milliseconds. Nor could its electronics handle an internal short that would put full fault current (hundreds of amps) through the switching device. But it works OK.

Do my knowledge there is no mode where an optimizer doesn't turn on if it sees 16V. Why would it. It is possible to have strings much larger than 16. Residential string size is max of 25 optimizers single phase, and 50 in 3 phase 480V.

Right! And thus each inverter can know how many optimizers are connected in series. Can it handle only 8? Then it wouldn't turn on if it sees more than 8 volts. Can it handle 50 in series? Then it would turn on if it sees 49 volts in safe mode - but not if it sees 51.

Now, the Solaredge may indeed not do that; I don't know how it handles faults like that. But given that installers may occasionally make errors (and as exemplified by this thread) it seems like cheap insurance. It's a lot easier to have an installer fix a wiring error than to replace a damaged inverter.

[ButchDeal]

Right. So if that billion watt panel has the right voltage range (right Voc and Vpp) it will work.

Most have a specification on the Isc of the array (short circuit limit). Some list max wattage but what they are really trying to limit is usually the Isc.
An array of 1GW is going to have a very high Isc even when Voc and Vpp are within range (will also have one hell of a combiner) and arc fault is going to be a problem.

Again, switchmode converters have overcurrent protection on their outputs. Keep in mind that the 5 watt switchmode converter in your phone charger can handle a short circuit on its output, despite being fed with a source that is capable of hundreds of amps for a few milliseconds. It does this not by "handling" a dead short across the 120 volt AC input (which could be hundreds of amps) but by limiting its own output power to safe levels.

The 5 watt converter in my phone. It is feed from a wall wort which has about 5 watt capacity. But this is a bit different from a large DC voltage source like an array, with arc issues.

Each optimizer outputs 1 volt in the de-energized state, right? So if it ever sees 16 volts before turn-on - it doesn't turn on.

1 volte in safe mode yes.
Do my knowledge there is no mode where an optimizer doesn't turn on if it sees 16V. Why would it. It is possible to have strings much larger than 16. Residential string size is max of 25 optimizers single phase, and 50 in 3 phase 480V.

And of course there's the usual way - seeing the input go past the rated current or voltage limitations for that inverter. For example, the SE7600 works at 400 volts, but is rated up to 480 volts. So if it ever sees 450 volts, something is definitely wrong, and it initiates a rapid shutdown. This brings the voltage down to n volts (16 volts for the 16 device string) which protects the inverter and optimizers.

OP has SE6000H not SE7600H, the older inverters tend to operate around 350V depending on the AC voltage it sees, newer around 380V.
They often jump up to higher voltages to handle clipping though. Remember the optimizers have a DC to DC buck/boost so they tend to keep their own modules within MPPT range.
Optimizers to my knowledge don't see a problem if there is 450 V on the DC string and the inverter very well likely would do a shutdown if it sees the voltage getting out of spec.
Voltage isn't the only problem here though amperage is, DC amps can cause arcs. The optimizers are equiped to handle a set amount of amps.

Definitely agreed there. If the inverter shuts down due to too many (or too few) panels that would be a good outcome; it would give the installer a chance to correct the problem.

There are many specs on the arrangement of SolarEdge. See the spec sheet for the optimizers again:
In this case the OP has P400s and SE6000H so:
Minimum number of optimizers per string is 8
maximum number of optimizers per string is 25
Maximum STC power installed per string is 5.7kw

https://www.solaredge.com/sites/default/files/string_sizing_na.pdf

https://www.solaredge.com/sites/default/files/se-P5-series-add-on-power-optimizer-datasheet-na.pdf

[jflorey2]

Yes of course they will but only if the input is within the specification range.

Right. So if that billion watt panel has the right voltage range (right Voc and Vpp) it will work.

Possibly but I doubt many systems would be able to handle the short circuit load on a GW array.

Again, switchmode converters have overcurrent protection on their outputs. Keep in mind that the 5 watt switchmode converter in your phone charger can handle a short circuit on its output, despite being fed with a source that is capable of hundreds of amps for a few milliseconds. It does this not by "handling" a dead short across the 120 volt AC input (which could be hundreds of amps) but by limiting its own output power to safe levels.

How would the inverter know that the system is designed incorrectly till the adverse input occurs?

Each optimizer outputs 1 volt in the de-energized state, right? So if it ever sees 16 volts before turn-on - it doesn't turn on.

And of course there's the usual way - seeing the input go past the rated current or voltage limitations for that inverter. For example, the SE7600 works at 400 volts, but is rated up to 480 volts. So if it ever sees 450 volts, something is definitely wrong, and it initiates a rapid shutdown. This brings the voltage down to n volts (16 volts for the 16 device string) which protects the inverter and optimizers.

SolarEdge wants to maintain some level of safety which is why they recommend 8 and not 7, getting things ore into the middle of the operating range instead of the edge.

Definitely agreed there. If the inverter shuts down due to too many (or too few) panels that would be a good outcome; it would give the installer a chance to correct the problem.

[arf88]

To the OP, are you sure they put them all on one string? The reason I ask is that I have 2 strings (I know there are physically 2 strings wired). However, the installer set them up in the reporting SE tool as a single string. There is no effect on the inverter or optimizers as the single string is only seen in the reporting tool. The reports seem to be correct for the most part. The only thing is the Module Mismatch Analysis report in the SE portal shows a wide range of +20 to -20. I know this is simply a reporting error.

100% sure, like ButchDeal showed when you go to the logical view you can see the different strings by the numbering. 1.1.x and 1.2.x

For the OPs case of too much wattage, I don't know how much safety SolarEdge has worked into the equipment but like stated as installed now it VOIDS the warranty, and certainly increases the likely hood of a failure to some degree.

Fortunately there have been some cold days in the past few months before it started cooking and there is some safety built in otherwise there would have been obvious problems already by now I would suspect.

[ButchDeal]

Simply not true for current. They will limit their output current to a safe level. Try it; a great many other people have.

Yes of course they will but only if the input is within the specification range.

It will just clip. Not because of anything that has to do with a bridge, but because of basic switchmode converter design principles. I can explain it in more detail if you like.

Possibly but I doubt many systems would be able to handle the short circuit load on a GW array. As stated, once you get out of the manufacturers operating range the likelihood of failure increases and eventually gets to the point of certainty.

In any case, I am surprised that the Solaredge system cannot handle a common mistake like that. It would be simple to design the system to simply not power on, or to power on and then limit. But I agree that you should not implement a system that exceeds the ratings on the Solaredge inverter (or any inverter) and expect it to produce power similar to rated array power.

How would the inverter know that the system is designed incorrectly till the adverse input occurs?
As the OP has pointed out and most know just because the STC DC voltage is over the specified limit, it is not going to hit it in hot weather....

They do what they can, with handling adverse conditions and poor installations. For example it will try to work as best it can with too few or too many optimizers but they will be at their limits during some conditions (usually cold days). I have had installers install 6 modules on a string which the older optimizers do not support. It is minimum recommendation of 8 optimizers in a string but technically 6 will in many conditions barely work as long as the AC line voltage is just right and all the modules are working, but as the AC voltage increases, 6 can not maintain high enough DC voltage to keep up. We had the installers change the wiring to correct the condition almost immediately even though as the installer stated "It is working", I could show how each optimizer what at its limit for DC to DC conversion. SolarEdge wants to maintain some level of safety which is why they recommend 8 and not 7, getting things ore into the middle of the operating range instead of the edge.
For the OPs case of too much wattage, I don't know how much safety SolarEdge has worked into the equipment but like stated as installed now it VOIDS the warranty, and certainly increases the likely hood of a failure to some degree.

[jflorey2]

An mppt charge controller has specificcations as well and they will fry if you connect mor than they are speced for.

Simply not true for current. They will limit their output current to a safe level. Try it; a great many other people have.

if you think you can connect a 2billion watt array to a residential inverter and it will just clip then you are likely a good candidate for purchase of a bridge.
clipping is where an inverter has more input than it can handle but still within the operating range.

It will just clip. Not because of anything that has to do with a bridge, but because of basic switchmode converter design principles. I can explain it in more detail if you like.

In any case, I am surprised that the Solaredge system cannot handle a common mistake like that. It would be simple to design the system to simply not power on, or to power on and then limit. But I agree that you should not implement a system that exceeds the ratings on the Solaredge inverter (or any inverter) and expect it to produce power similar to rated array power.

[ButchDeal]

To the OP, are you sure they put them all on one string? The reason I ask is that I have 2 strings (I know there are physically 2 strings wired). However, the installer set them up in the reporting SE tool as a single string. There is no effect on the inverter or optimizers as the single string is only seen in the reporting tool. The reports seem to be correct for the most part. The only thing is the Module Mismatch Analysis report in the SE portal shows a wide range of +20 to -20. I know this is simply a reporting error.

In the monitoring site you can set them up as a single string but usually the system will determine the strings for itself and show up in correct arrangements.

Here is an example of one that I set up with information from the install. The installer did not in this case provide the actual strings just the layout information so I put it in as a single string
But the system determined that the top row was one string and the rest were the second string.
You can tell by the numbering, and the logical layout. The ones where the second number is 1 are string 1, the ones with the second number is 2 are string 2.
Actually the first number is inverter number, then string number, then optimizer/module number; so 1.2.6 is inverter 1, string 2, optimizer 6

Layout1.jpg

Here is the logical layout showing the two strings:
Layout.jpg

[DrLumen]

To the OP, are you sure they put them all on one string? The reason I ask is that I have 2 strings (I know there are physically 2 strings wired). However, the installer set them up in the reporting SE tool as a single string. There is no effect on the inverter or optimizers as the single string is only seen in the reporting tool. The reports seem to be correct for the most part. The only thing is the Module Mismatch Analysis report in the SE portal shows a wide range of +20 to -20. I know this is simply a reporting error.

[ButchDeal]

Of course something can be done about that. MPPT charge controllers do it every day. If the charge current declines, the voltage doesn't have to go through the roof. The DC/DC converter within the MPPT controller limits the voltage and current at the same time. They do this by reducing the DC/DC converter drive pulse width until the current drops, while voltage remains the same.

No, it doesn't - I've tried it. So do many other inverters. It's called "clipping" when you see it during operation. It stops drawing more current from the array, even though the current is available. The voltage can only rise to the Voc of the panels, at which point you are simply throwing away the unused current. And since you have to design for Voc anyway, you are safe from overvoltage too.

In the case of the optimizers, if the optimizers are really 'working together' to set a fixed voltage at the input of the inverter, then they can limit the current to some maximum by doing the same thing - operating at a point that "throws away" the excess available power, just as MPPT controllers and overpaneled string inverters do. Now, maybe the Solaredge converters can't do this, and there is a failure mode where they can't back off any further than X. But that's not because it's impossible, it's because someone decided not to cover that possibility.

OK, now compare this to your statement "The optimizers can convert voltage but if you cut voltage then amps increase. Nothing can be done about that."

An mppt charge controller has specificcations as well and they will fry if you connect mor than they are speced for.

if you think you can connect a 2billion watt array to a residential inverter and it will just clip then you are likely a good candidate for purchase of a bridge.
clipping is where an inverter has more input than it can handle but still within the operating range.

The optimizers have an operating range, when yu are outside of that two things happen, your warranty is voided, and the second, you risk failure. The further outside of the iperating range the higher the risk of failure, at some point it becomes certain.

[jflorey2]

The optimizers can convert voltage but if you cut voltage then amps increase. Nothing can be done about that . . .

Of course something can be done about that. MPPT charge controllers do it every day. If the charge current declines, the voltage doesn't have to go through the roof. The DC/DC converter within the MPPT controller limits the voltage and current at the same time. They do this by reducing the DC/DC converter drive pulse width until the current drops, while voltage remains the same.

NO the SMA inverter would go up in a ball of fire, because the AMPERAGE is way out of specifications (~2.5MA).

No, it doesn't - I've tried it. So do many other inverters. It's called "clipping" when you see it during operation. It stops drawing more current from the array, even though the current is available. The voltage can only rise to the Voc of the panels, at which point you are simply throwing away the unused current. And since you have to design for Voc anyway, you are safe from overvoltage too.

In the case of the optimizers, if the optimizers are really 'working together' to set a fixed voltage at the input of the inverter, then they can limit the current to some maximum by doing the same thing - operating at a point that "throws away" the excess available power, just as MPPT controllers and overpaneled string inverters do. Now, maybe the Solaredge converters can't do this, and there is a failure mode where they can't back off any further than X. But that's not because it's impossible, it's because someone decided not to cover that possibility.

The optimizers are very efficient and no place to put the power other than conversion of voltage. for permitting and the way the devices work is they are either ON and producing or off (shorted) and spit out 1 volt almost no amps.

OK, now compare this to your statement "The optimizers can convert voltage but if you cut voltage then amps increase. Nothing can be done about that."

[Mike90250]

You measure the damage by how often you replace failed units. They will self-protect to some extent, My MPPT controllers do it by allowing the array voltage to rise and that shifts things into less efficient ranges, and they can revert to PWM mode. But they can be overwhelmed. I'm not going to push the limits, I'm staying where the devices are specified to work.

[ButchDeal]

Right. But if you install a small electric motor with a good inverter on the dump truck it won't burn out or anything - it will limit the current to the motor, and just not move the truck. It sounded like you were suggesting it could be damaged.

ok let me give a better example. When I was in NJ, I took my bike to a shop with a mechanic and a painter (co-owners) the painter had a very extreme bike that started out as the same model as mine. He first added a turbo and nitros to the the engine which nearly doubled the power. Took it on his first ride and ripped the clutch apart. The clutch was fine for the 80ft/lbs of torque it originally had but nearly double that and it was not sufficient. He then put on a competition clutch and on the first drive he ripped the final drive belt apart. The belt was fine for the original specs but when he added the added power and the clutch that could handle it, the next weekest part took the abuse.

Right - but from your descriptions in the past, they are fully functional and protected DC/DC converters that can create whatever voltage/current is needed by the system. And almost all DC/DC converters incorporate current limiting to prevent damage due to shorts and overloads. (Which I think is what you meant by "limiting device like a fuse." If they do that, then the inverter is protected.

The optimizers can convert voltage but if you cut voltage then amps increase. Nothing can be done about that, lower voltage equals more amps, higher voltage lower amps.
The optimizers are very efficient and no place to put the power other than conversion of voltage.
for permitting and the way the devices work is they are either ON and producing or off (shorted) and spit out 1 volt almost no amps.
They are considered limiting devices by code and thus you do not need additional fuses between the strings and inverter, but (just like fuses) ONLY when installed correctly and to manufacturers specifications. The manufacturer has limited the wattage that can be on a string!

I wouldn't expect it to work and convert 1 billion watts. But if I connected 1 billion watts of solar to an SMA inverter (for example) it would just produce up to whatever it was rated to produce, and then limit. (As long as you keep DC voltages within its operating limits of course.)

NO the SMA inverter would go up in a ball of fire, because the AMPERAGE is way out of specifications (~2.5MA). There are MANY MANY specifications on the input. Wattage is one, amps and volts is another. String inverters are going to limit based on volts and amps. Optimizers because they can handle more are limited on wattage which is a combination of volts and amps. This makes the optimizers more flexible but you still have the limitation.

In the case of the Solaredge inverter I could see it shutting down in that case, if the optimizers couldn't throttle back enough or something. But where would the damage come in? The optimizers have to be instructed to turn on and produce X volts. Why would the inverter instruct them to provide a damaging voltage or current?

Optimizers are not "micro" managed like this. They work more as a collective and group think. Each puts out what it can best put out, and others come down to keep in range. The net result is quickly getting to optimal levels for all but not centrally controlled by the inverter. In any case ALL systems have limits and specifications. When anything is installed outside of its limits things break.

I really am having a hard time believing that you are not trolling this discussion.

[arf88]

Right. But if you install a small electric motor with a good inverter on the dump truck it won't burn out or anything - it will limit the current to the motor, and just not move the truck. It sounded like you were suggesting it could be damaged.

Right - but from your descriptions in the past, they are fully functional and protected DC/DC converters that can create whatever voltage/current is needed by the system. And almost all DC/DC converters incorporate current limiting to prevent damage due to shorts and overloads. (Which I think is what you meant by "limiting device like a fuse." If they do that, then the inverter is protected.

I wouldn't expect it to work and convert 1 billion watts. But if I connected 1 billion watts of solar to an SMA inverter (for example) it would just produce up to whatever it was rated to produce, and then limit. (As long as you keep DC voltages within its operating limits of course.)

In the case of the Solaredge inverter I could see it shutting down in that case, if the optimizers couldn't throttle back enough or something. But where would the damage come in? The optimizers have to be instructed to turn on and produce X volts. Why would the inverter instruct them to provide a damaging voltage or current?

Well said, in addition isn't the point of clipping to prevent damage as well. This in theory should happen before the over voltage and beyond peak wattage is reached, all assumptions on my part?

ButchDeal by measurement I mean the damage to the inverter. There is an error page in the portal but beyond that how can one measure the damage and degradation caused. My assumption is there is no way unless its instrumented at the PCB level but that is not realistic.