Choice of inverter

[jrweiss98020]

IIRC, bypass diodes work between individual solar cells or groups of cells within a panel, NOT for the entire panel. They prevent a partially shaded panel from fully shutting down.

However, if a panel is fully shaded, it may bring down the string if optimizers are not installed.

 

[nwdiver]

IIRC, bypass diodes work between individual solar cells or groups of cells within a panel, NOT for the entire panel. They prevent a partially shaded panel from fully shutting down.

They work for each cell group or all cells groups (the whole panel). There are no cell groups that can't be bypassed => the entire panel can be bypassed.

 

[jrweiss98020]

Can you post some references, please?

 

[nwdiver]

Can you post some references, please?

A screen shot of 2 panels in a string clearly being fully bypassed wasn't enough?

You don't need optimizers to prevent this. Bypass diodes also allow unshaded panels to continue operating. All solar panels have bypass diodes.

Here's a screenshot of unshaded panels operating in a string with shaded panels. Current is bypassing the shaded panels.

View attachment 632922

Here's the innards of a solar panel J-box. You can see that there's a clear path in, through 3 bypass diodes and out... If voltage drops across a cell group due to shade the bypass diode becomes reverse biased and creates a circuit for current flow. If all three cell groups have shade the entire panel is bypassed.

 

[zonela]

I'm in NorCal, full crew came over to install my solar system today. However, they only got the Tesla inverter in contrary to the SolarEdge I had on my BOM. We had some discussion and the team lead mentioned he would rather wait for the SE inverter if it were him. They had quite a few Tesla inverters installed and are seeing more than expected issues already. He personally would stick with SE over Telsa/Delta inverters, at least for now. Really appreciate his candor. Wait time is estimated 3-4 weeks for the SE inverters to be in stock. Given that peoples experience here on waiting for Tesla trying to fix any solar issues after installation, I decided to wait for SolarEgde inverter instead of proceeding with the installation today.

 

[jrweiss98020]

A screen shot of 2 panels in a string clearly being fully bypassed wasn't enough?

Here's the innards of a solar panel J-box. You can see that there's a clear path in, through 3 bypass diodes and out... If voltage drops across a cell group due to shade the bypass diode becomes reverse biased and creates a circuit for current flow. If all three cell groups have shade the entire panel is bypassed.

No; not enough. A picture with your description of its operation is not a reasonable reference, especially since your opinion is counter to every professional reference I've seen.

Per Power Optimizers: Everything You Need to Know,

The optimizers increase the overall energy output of your PV array by consistently tracking the maximum power point (MPPT) of each individual module in your system. Tracking the modules MPPT increases the efficiency of DC power from the solar cell, and down to the central inverter where that power is converted to usable AC power.

Also, per What are DC power optimizers?,

However, if you have a roof with partial shading issues or multiple orientations, your power output will suffer. That’s because the production of the string of panels is limited by the performance of the worst performing panel.

In the challenging solar installation conditions outlined above, the maximum power you can create with each panel varies. That’s because the amount of power each panel generates depends on temperature, amount of sunlight, and special characteristics of the power load. When panels face multiple directions or some are shaded, these factors differ across the array.

Power optimizers account for this problem by using a technology called maximum power point tracking (MPPT). Smart researchers figured out that if you can track the maximum power of each panel in real time, then smooth and condition the electricity each panel generates before sending it along to the inverter, your inverter can process much more electricity.

And finally, for now, String Inverters, Power Optimizers, and Microinverters | EnergySage:

Like microinverters, power optimizers are devices located at each panel. However, instead of converting the DC electricity to AC electricity at the panel site, they “condition” the DC electricity and send it to a string inverter. In scenarios where your roof is shaded, the panel level optimization afforded by power optimizers results in higher system efficiency than using a string inverter alone.

Similar to microinverters, power optimizers reduce the impact of panel shading on system performance, and also offer panel-level performance monitoring.

 

[pdx_m3s]

I'm in NorCal, full crew came over to install my solar system today. However, they only got the Tesla inverter in contrary to the SolarEdge I had on my BOM. We had some discussion and the team lead mentioned he would rather wait for the SE inverter if it were him. They had quite a few Tesla inverters installed and are seeing more than expected issues already. He personally would stick with SE over Telsa/Delta inverters, at least for now. Really appreciate his candor. Wait time is estimated 3-4 weeks for the SE inverters to be in stock. Given that peoples experience here on waiting for Tesla trying to fix any solar issues after installation, I decided to wait for SolarEgde inverter instead of proceeding with the installation today.

As with any new tech, avoiding rev1 is advised.

 

[zonela]

I'm in NorCal, full crew came over to install my solar system today. However, they only got the Tesla inverter in contrary to the SolarEdge I had on my BOM. We had some discussion and the team lead mentioned he would rather wait for the SE inverter if it were him. They had quite a few Tesla inverters installed and are seeing more than expected issues already. He personally would stick with SE over Telsa/Delta inverters, at least for now. Really appreciate his candor. Wait time is estimated 3-4 weeks for the SE inverters to be in stock. Given that peoples experience here on waiting for Tesla trying to fix any solar issues after installation, I decided to wait for SolarEgde inverter instead of proceeding with the installation today.

Chatted with my project advisor and is being pushed to accept Tesla invertor, claiming Tesla is ending contracts with SE/Delta soon and it's showing 180 days to wait for any SolarEdge inverters. Anyway, gonna end up delaying a month or two for another installation time slot due to this surprise today..

 

[nwdiver]

No; not enough. A picture with your description of its operation is not a reasonable reference, especially since your opinion

Not really an ‘opinion’... you can see the wiring for yourself. It’s not complicated. Just 3 diodes allowing current to bypass shaded sections. How... how would it not bypass the panel if all 3 sections have shade? Look at the wiring... that’s not possible.

What you’re referring to as ‘professional references’ are just marketing misinformation. I posted panel performance. There’s no real benefit to module level over string level if installed correctly.

 

[pdx_m3s]

Not really an ‘opinion’... you can see the wiring for yourself. It’s not complicated. Just 3 diodes allowing current to bypass shaded sections. How... how would it not bypass the panel if all 3 sections have shade? Look at the wiring... that’s not possible.

What you’re referring to as ‘professional references’ are just marketing misinformation. I posted panel performance. There’s no real benefit to module level be string level if installed correctly.

Is there any benefit to optimizers for panels in different orientations? Say you have 12 panels facing west, and 3 facing south, whereby the 3 facing south aren't big enough to be on their own string (i.e. all 15 panels on 1 string). What about that?

 

[nwdiver]

Is there any benefit to optimizers for panels in different orientations? Say you have 12 panels facing west, and 3 facing south, whereby the 3 facing south aren't big enough to be on their own string (i.e. all 15 panels on 1 string). What about that?

Yes; That’s the one situation where there’s a benefit to optimizers but with multi-string inverters it’s rare that you can’t just have 2 or 3 strings. Each string can face a different direction just all panels in each string need to have the same orientation for best results.

 

[Vines]

No; not enough. A picture with your description of its operation is not a reasonable reference, especially since your opinion is counter to every professional reference I've seen.

There is some grey area here and this answer is not black and white.

It all comes down to cost. PV panels are not a race car, so the cool factor is over. You want it to make you money, and to do that you have to first spend money. This creates a payback period and the efficiency that matters is $ per kWh, with some concern for Aesthetics and PITA factor.

The solar industry is rapidly changing, and just 3 years ago the answer was not the same. Its still not always the same answer. String inverters today are way more flexible.

String Voltage: If you string was on the low end of the MPPT voltage window, or startup voltage window, then bypassing a module or 2 can be catastrophic. However inverters these days have huge windows. I can remember when you really wanted target voltage within 150-200V wide windows. These days you can have inverters that fire between 175VDC and 900VDC with 600V wide MPPT windows. The MPPT technology and multiple channels really makes strings way more flexible than they were in the past. Way less often will a single module shaded or even 3 will take down a string completely.

Design constraints: With High Voltage/Low Current modules (Sunpower) this problem is a bit worse, as each loss of a module is a large portion of a string. When a string had a 200V target window, and your customer didn't buy exactly enough PV to be at the top of the window, then easily losing 1-2 panels can drop MPPT target or even startup voltage.

In exchange you didn't buy a $60-80 piece of electronics and install it on the back of your $200 module, increasing its price by a whopping 20%. How many years will those Optimizers cost to payback? Will they, or if the rest of system dies before your payback then you lost, since you had to mount them and wire them too.

The final consideration is Rapid Shutdown, and I believe most successful DC inverters have a DC keep alive signal (See Sunspec Compliant Inverters) Then you install a box on the roof one per module or 1 per 2 modules with 80V maximum which accomplishes the rapid shutdown.

There are probably other technologies, but as I have dealt with microinverters mostly for the last few years I am not as close to this as I once was so my knowledge might be a bit old.

 

[nwdiver]

String Voltage: If you string was on the low end of the MPPT voltage window, or startup voltage window, then bypassing a module or 2 can be catastrophic. However inverters these days have huge windows. I can remember when you really wanted target voltage within 150-200V wide windows. These days you can have inverters that fire between 175VDC and 900VDC with 600V wide MPPT windows. The MPPT technology and multiple channels really makes strings way more flexible than they were in the past. Way less often will a single module shaded or even 3 will take down a string completely.

This is even more simplistic than that. A debate around whether the inverter will be operating outside its MPP band is fine... at least that's based on reality. The argument @jrweiss98020 is making (unless I'm misunderstanding his position) is that...

However, if a panel is fully shaded, it may bring down the string if optimizers are not installed.

...a fully shaded panel isn't bypassed. Which is a conceptual misunderstanding that is seemingly intentionally perpetrated by Enphase and SE. It's simply not true. To some extent it's the opposite. With an optimizer or micro if a panel is fully shaded you have the ability to NOT bypass the panel and squeeze a few watts from a panel in which all cell groups are shaded (yay?). But if a single cell group has full sun the other two are bypassed regardless of whether it's an optimizer, string or micro.

 

[pdx_m3s]

Yes; That’s the one situation where there’s a benefit to optimizers but with multi-string inverters it’s rare that you can’t just have 2 or 3 strings. Each string can face a different direction just all panels in each string need to have the same orientation for best results.

Good to know. And obviously, must meet the minimum string voltage. I believe it's usually around 8 panels.

 

[nwdiver]

Good to know. And obviously, must meet the minimum string voltage. I believe it's usually around 8 panels.

With modern inverters 3 might work. With the right panel even 2 panels can work. Most MPP bands are now 100v to 480v and most newer solar panels have a Vmp of ~35v, higher voltage panels have a Vmp of ~59v.

 

[dareed1]

Let's take two extremes, using approximately correct numbers for Q.PEAK DUO BLK-G6+ 340W panels that Tesla is currently installing here in NorCal. These are so-called half-cell panels with 120 cells--3 substrings of 20 in the upper half of the panel and another 3 substrings in the lower half. There are three diodes as usual. Let's say that at noon each panel produces 9 Amps at 33 Volts or 297 watts. Assume 12 panels to correspond to a 4.08kW system. If an entire panel is shaded, the power output using either a string inverter or a optimizer system will be close to 3.27kW instead of 3.56kW. In both cases the panel is bypassed. Here I've assumed that the diffuse radiation isn't intense enough to allow conversion in the optimizers. Practically there is no difference in the two type of systems.

On the other hand, if the lower or upper half of a panel is shaded, then it is capable of still outputting 33V, but now at only 4.5A. Remember that the current running through the the panels (string inverter) or through the output circuits of the optimizers has to be the same for all panels. The half shaded panel is capable of generating 148W in this situation, and an optimizer system would be able to capture most of it, because the DC to DC converter can transform the panel output to 16.5V at 9A (with some loss in efficiency) in the optimizer output. A string inverter would find that it could either consume 4.5A from all the panels at 396V total (1.78kW) or consume 9A (which will cause the shaded panel to be bypassed) at 363V (3.27kW, the same as if the entire panel were shaded). Of course it chooses the higher power.

So in this case the optimizers make about 5% more power than the string inverter. If more than one panel is shaded, the advantage increases, but not linearly: if all 12 panels are half shaded, the two systems give the same power.

Is this advantage significant? I suppose that in most installations, shading occurs in the early morning or late afternoon, when power generation is already a fraction of peak output. At our house, once shading starts, it progresses from none to covering the whole array in about 30 minutes. For this situation, the power generation difference is truly minimal. I suppose that in some instances shading occurs at midday e.g. a tall telephone pole whose shadow slowly traverses the whole array, and here the advantage of optimizer system could be significant.

I think the ability to individually monitor power output per panel is a nice perk for us nerds, and optimizers/microinverters or rapid shutdown systems with monitoring can provide that while a plain string inverter (Delta or new Tesla) with middle circuit interrupters can't.

BTW, the Delta M8 specifies that it can do MPPT from 50 to 480V, so in principle 2 panels is enough.

 

[nwdiver]

On the other hand, if the lower or upper half of a panel is shaded, then it is capable of still outputting 33V, but now at only 4.5A.

That's not how shade across all cell groups behaves. The current of the cell group is restricted by the lowest current allowable by the most heavily shaded cell. The current of a solar cell is proportional to the light it's exposed to. If a solar cell 'wants' to push 6A and you try to push 8A that cell become a load and voltage drops. That's HOW bypass diodes function. It doesn't take much. I built my first solar array almost from scratch. I installed the bypass diodes so I got to experiment with how much shade it took to trigger them. Usually only a cell or two.

You don't lose half power with a half shaded panel. You lose ~80% because the shaded section restricts current flow. This is an optimized panel getting hit with shade across all cell groups. You can see the optimizer tracking to the higher MPP at ~38v. Once the shade hits current drops from ~7A to 1.2A. Sure... with a string inverter the output of this panel would be 0w instead of 45w. Is that worth it? IMHO... no.

 

[dareed1]

nwdiver,
I am specifically referring to half cell solar panels, and where the shade happens to cover one half of the panel so that the other group of 60 cells are fully illuminated. That's why half cell panels have become popular: one half of the panel is able to drive current independently of the other. I did describe it as the extreme case; typically shade won't fall so nicely on just one of the two groups.

 

[nwdiver]

nwdiver,
I am specifically referring to half cell solar panels, and where the shade happens to cover one half of the panel so that the other group of 60 cells are fully illuminated. That's why half cell panels have become popular: one half of the panel is able to drive current independently of the other. I did describe it as the extreme case; typically shade won't fall so nicely on just one of the two groups.

Interesting. The two sides of a half cell panel are in parallel not series. They're promoted as being more shade tolerant and this may be true in some cases but if 1/6th of the panel is shaded wouldn't you lose half the panel instead of 1/3rd with a regular panel? You can't have strings of dissimilar length in parallel which is effectively what would happen if 1 cell group of a half-cell panel was shaded...

 

[dareed1]

No, if 1/6th is shaded 1/3 is lost. Here's a pretty good description of how they work. The graphic that shows how the voltages and power are proportioned is by REC, an Australian solar panel maker. As far as I understand it, these panels are always at least as good as a standard panel in shaded situations.