top of page
9689338E-C691-4DC7-ABA3-05930D268CFA_edited.jpg
  • Writer's pictureJeff Mathias

From string inverters to micro inverters, an 18-year history






Recently I attended a Tesla Powerwall 3 training session, and it brought me back to my roots in the solar industry.  I decided to write this blog to cover this history since 2006 when I started designing solar systems for the residential and small commercial marketplace and the changes in inverters over the many years to present.    

 

First what is an inverter, inverters convert the DC power produced by solar modules into AC power our homes and the electric grid use.  While the main purpose of inverters in solar systems has not changed, the design, installation and how they are supported has.

 

Transformer based String Inverters with a single MPPT :

 


String inverters allowed modules to be Strung together in series, usually in strings of modules between 7-13, with each module increasing the voltage of the string.  The early inverters had only a single maximum power point tracker, MPPT.   The MPPT was used to maximize power extraction from a solar system.   Shade is the enemy of solar and these early inverters and homes with heavy to moderate shade often could not be justified.  Some challenges with these early designs of a single string inverters included:

 

-              Early string inverters had to have a minimum voltage and maximum that often meant a string could have 8-12 modules per string.  Multiple strings on the same inverter often needed to be the same size. So, if you wanted to design a system with 25 modules, you would need two inverter, 1 with 2 strings of 8 and another with 9, or reduce the array size to 24 and have 2 strings of 12 if I want one inverter.

-              As mentioned above, shade often made it not practical to install a system as shading a module 10% could cause the entire string to drop 40% or more. 

-              This sensitivity to shade also showed itself in other ways.  If one module received heavier soiling (think pigeons and chimneys) it would negatively affect the entire system.  And if one module had a performance issue, say a diode went out cause 1/3 of the power on a module to fail, it would lower all the modules on the string and potentially the system.

-               Panel mismatch was also an issue often overlooked.  Modules are often made in ranges, say a 400-watt module is usually a minimum and the range can be between 400-watts to 409-watts.  If mixed modules are on the same string, say (1) 400-watt and (11) 409-watt all modules would operate at 400-watts. 

-              Orientation and pitch are also an issue.  If you have both an east (produces more in the mornings) and west (produces more in the afternoon) facing array on one inverter, the east array will be brought down in the morning by the west array’s lower production and the wests in the afternoon by the easts.

-              Roof slope is also an issue, with steeper roofs producing more in the winter and flatter roofs more in the summer.  If we mix and match strings between different roof slopes, the strings will not produce efficiently.

 

As you can see, all these requirements in trying to maximize solar production often made design more of an art form then simply putting modules where they looked the best.  Some of the issues could be addressed by adding multiple inverters, but this increased cost.  Other times the limitations just did not make sense.  Most solar companies had a highly experienced designer who could consider all of these factors in system design.  In these challenging old days, I often advise 1/3 of my prospects that solar did not make sense.

 

There were other challenges with string inverters with repair and maintenance:

 

-              Most string inverters when I started in 2006 had a 5-10 -year warranty.  Since most modules had a 25-year warranty, you could almost guaranty there would need to be at least one replacement in the life cycle of the system if not two.

-              Very few of these early systems had monitoring on them, and if they did it was system level reporting, not module level.  This made it extremely difficult to know if there were any module level issues and required customers to physically monitor their systems.

-              While some string inverters were reliable, when you had a problem with a system’s string or module, it was often like battleship on the roof to find the problem.  This required experienced service personal with good troubleshoot skills.

-              String inverters often required multiple trips to repair, one to go on-site and diagnosis the problem, remove the inverter, package it and send it off to the manufacturer, and a second to receive the new inverter and reinstall it.  Often a third trip was involved.  What compounded this is the weight, often over 100 lbs. could require multiple service technicians.

-              When a sting inverter was down, it was most often a critical failure, the entire system was not working or producing power.  This required a quick response from an installer to maintain good customer service.

-              String inverters also were high voltage, 600v or 1,000v, requiring addition caution when working on them.

 

String inverters did offer some benefits for simple installations (one orientation and slope):

 

1)        They were reliable during their warranty period.

2)        They had a lower up-front cost in most cases.

3)        And in almost all cases string inverters were installed on the ground.

 

Transformer-less string inverters with dual-MPPT’s: Around 2013, transformer less string inverters with multiple (usually dual) MPPT’s began replacing the older transformer based string inverters in the residential and small commercial marketplace.  These newer inverters did solve a few of the problems of their predecessors. 

 

Dual or multiple MPPT’s allowed more system design flexibility:



o   Different number of modules, per MPPT on a single inverter, say a strings of 12 modules and one of 6 modules.

o   Different orientations or slopes per MPPT, so in the example above you could have 24 modules on a south sloping roof (2 strings of 12) and 6 more on a west facing roof.

o   Dual MPPTs did allow a little more flexibility around shading, modules that had a lot of shade could be on MPPT while the others in full sun on another MPPT would be not affected by the shading.

-              Transformer less inverters also were much lighter:

o   Single person replacements.

o   Lower cost for shipping and handling.

o   Easier installations.

o   Came with a standard 10-year warranty.

But while it made string inverters better, we still had major issues:

-              Performance was still based by strings, shading, panel mismatch, and soiling of any panel effected the string.

-              Very few of these systems had monitoring, and if used, monitoring was by system only not module.

-              If a single module had derogating performance, it could go unnoticed unless it was severe.

-              Trouble shooting was complex and required a skilled technician.

-              System design was complicated and required an experience designer.

 

Optimizers with String Inverters:

 

-              In 2011, I designed my first optimizer system with SMA string inverters.  In this



design, each module had an optimizer, with strings of modules going to a string inverter like designs above.  The optimizers allowed each module to have semi-autonomy, so we could mix orientations and pitches while minimizing performance losses.  This addressed issues of module mismatch, rolling shade and soiling.  In addition, many of these systems allowed us to have module level monitoring so we could see how each module was performing.

 

While optimizes solved some issues, they still had challenges:

 

-              Addition equipment can lead to more failures and increased maintenance.

-              It does take additional time to install the optimizers and monitoring system.

-              Coming off the roof with High Voltage DC Power, usually 600 volts.

-              While optimizers often come with a 25-year warranty, the string inverters still usually only have a 10-year warranty, requiring replacement during systems life cycle.

 

Beginning in about 2012 we began to see most of the optimizer’s systems being install by SolarEdge and they are still the leader at the writing of this blog with about 40% of the residential marketplace.

 

Microinverters enter the market:

 




In 2008, Enphase Energy brought Microinverters to the market, and in a few short years they had released 4 generations.   While some microinverters could handle multiple modules, most were paired, one inverter per module. 

 

Let me digress, what is a Microinverter.   Micro inverters eliminate the need for string inverters and are installed under the solar modules (like optimizers).  Unlike optimizers though, their primary purpose of microinverters is to convert solars DC power to AC power the house can use, eliminating string inverters.  The quick growth of microinverters, often reminded me of the changes in the computer industry in the late 90’s.

 

Again, let me regress.  I worked in the computer industry back in the 90’s during the rise of personal computers and the internet.  Even though personal computers were not as reliable as mainframes, and they cost more than terminals, they still became the norm.  Why, end users wanted access information, without the need to go to the data processing department which often took months and required expensive programmers.  This “delay” and cost, between the need for information and receiving it, was the primary reason for this trend.

 

This is why by mid 2010, Enphase had grown to a 13% of the residential marketplace.  By 2023 this market share for residences has increased to almost 50%.  I designed and sold our first microinverter back in 2013 on a let’s try it and see basis, and by the end 2014 we were installing 100% Enphase microinverters.  The trend I saw happening at this time was microinverters were gaining market share in the residential market, while smaller string inverters were replacing large central inverters in the small to mid-size commercial jobs. 

 

While the first few generations of Microinverters were not as reliable as string inverters, they offered many of the same benefits to homeowners and installers that personal computers offered in the 90’s.  Microinverter monitoring was done not just at the system level, but down to the module level.  This enhanced reporting for both the homeowner and installer to see not just how much power their system was producing, but also how much each module was producing.  They could verify their system was running well and be alerted if it wasn’t.  

 

This also changed the way we designed, installed and maintained solar systems.

From a design standpoint my favorite advantage was the microinverters by 2014, came with a 25-year limited warranty, making it much easier to project lifecycle cost.   Since we converted to DC at the individual module level, we could design to what was needed, not the number that stringing allowed.  It also allowed many homeowners who had too much shade (especially rolling shade – shade that moves across the modules depending on time of day) for string inverters, to now go with solar since microinverters were more shade tolerant, shade only effected individual modules they touched, not the entire array.   It also eliminated the need of a bulky string inverter that often-required back boards or unique electrical runs to incorporate,

 




From an installation standpoint, it was safer for our crew since we were converting to 240 volts on the roof, not running high voltage dc to the ground. It was easier to test as we could see the voltages of each module and how they performed throughout the day.  Microinverters are light and easy to replace with monitoring pinpointing the exact module having issues.  By 2014 with the release of the fourth-generation inverters, reliability was on par with string inverters with about a 10% performance improvement.  And when we left the site, we could verify and show a customer everything was working as it should, remember this was the early days of solar so this was powerful. 

 

But one of the greatest values I found was on-line monitoring support for both the customer and the installer.  Having the ability to verify the system was producing and the ability to also monitor how much power the home was consuming was invaluable.  When a system or inverter stopped working, we would be notified so we could be proactive with customers repairs.  When an inverter failed, we could schedule the replacement since it only effected one module not the entire system.  And it improved customer service since we could assist a client from my office computers.

 

Some examples of this were:

 

-              On November 1st one year I called a customer wondering why the system was not working (I received a low production alert), asking him to check the solar breaker – a Halloween prankster turned it off so he lost only a partial day of power instead of months.

-              I called a customer noticing two modules had low production, asked them to look at the roof where someone had tossed paint can lids onto the array.

-              Many customers would call up asking why their bills had increased.  With the ability to see production I could verify the system was still producing correctly.  And using the consumption monitoring detail I could point out the time frames or see usage patters often finding the customer had added loads, kids home from school, electric car addition, using space heating, having a clogged well or a septic pump that was not resetting.  In other words, it was easy to provide good customer service with this information since we could see both solar system production and household usage.

 

While some of these issues can be determined with system level monitoring, it is almost impossible to verify all the modules are working properly without going on-site.  On-site verification is labor intensive, requires special tools, a highly trained technician and thus expensive. 

 

Another benefit is they have been very adaptable.  Grid tied solar by design is designed to have the inverter not export to the grid when it is down.  However, with DC strung systems, when the sun is shining the wires from the array to inverter are still live, even though the inverter is off.  Fire codes implemented requirements for rapid shut down, requiring special devices on string systems to turn off the power at the roof, which was already available on microinverters since the inverter was at the module.  Also string inverters we not designed for continued improvement, what you bought was what you got.  Enphase Microinverters had software and firmware that could be continually improved.



In my current consulting business at Mathias Energy Consulting, I am asked to do solar inspections on existing systems.  My main market share is for customers whose installer is no longer in business and they are not sure if the system is working.  When I run into string inverter systems, because of the high cost to evaluate each module (can easily double the inspection cost), I usually evaluate if the performance is working with-in guidelines.  I have found a few installers who will adopt these “orphaned” string inverter systems, usually at a fee to first do a detailed site-analysis.

 

This is very different then when I run into an Enphase microinverter System in need of support.  The last Enphase analysis I did was to a homeowner selling their home.   They had 52 solar modules and did not know if they worked.   Their solar installer was no longer in business and the monitoring was down.   I was able to reenable their monitoring system and the homeowner added me to the monitoring list.  From my house I was easily able to verify 51 modules were running well, and one inverter needed to be replaced.  I have a much larger list of vendors who will support Enphase and the adoption costs much lower since they can see the system’s module level monitoring.

 

And like all things, Microinverters do have some challenges:

 

-              They usually have a higher upfront cost.

-              The initial versions of the Enphase inverters – M175’s, M190, D380 and M210’s had a high failure rate.

-              While newer versions are very reliable, an installer may need to replace a few over 25 years.

-              Monitoring is required and must be maintained for warranty.

 

As mentioned above, these challenges are also an advantage.  While the upfront costs are higher, when you considered the 25 vs 10-year warranty, and the cost to replace a string inverter during the life cycle of the system, it is close.  The newer microinverters are very reliable, and Enphase did a nice job of replacing the older versions with newer more reliable replacements.  And monitoring is such a valuable tool for both the installer and homeowner it is well worth the effort to maintain it.

 

That is my history of the world of microinverter design over the past 18 years.  And to come full circle, what brought on this walk down memory lane, the newly released Tesla Powerwall 3 storage system.  In it’s DC coupled configuration, it will use a single string inverter to manage both the solar and storage system.  I wrote this blog to refresh why we have migrated to microinverters over these years.  I will write another blog specifically comparing the costs of using Enphase energy micro inverters on the roof with their storage system vs the Tesla Powerwall 3 DC coupled as a string inverter.  While I am a little biased as I cannot imagine going back to the days of string inverters the Tesla system does hold promise in some areas.

29 views0 comments

留言


bottom of page