PV life, MTBF, failure modes and maintenance schedule

[RegGuheert]

What is the expected life, MTBF, failure modes and maintenance schedule for your PV system?

For reference, here is a Wikipedia article on failure rate.

I'll start. Our system consists of 42 Enphase M190 microinverters and 42 Sharp NU-U235F3 PV modules. You can see details of the system along with pictures in the link in my signature.

LIFE

Enphase estimates the working life of the M190 to be over 20 years. I do not have better data than that, so I will use that.

Monocrystalline PV panels have an unknown life expectancy. Their electricity production gradually degrades over their long life.

MTBF

Enphase calculates the MTBF of the M190s to be 331 years, which equates to a FIT rate of 345 failures per billion hours. This number is pretty close to the actual MTBF I have calculated for systems operating all over North America. (MTBF does seem to be lower in AZ than elsewhere, BTW.) Since we have 42 M190s that gives a total FIT rate of 14,490 failures per billion hours. That means about one microinverter failure every 7.9 years.

Industry tends to feel that PV panels have an MTBF of over 600 years, so I will use 600 years or a FIT rate of 190 failures per billion hours. That gives a fit rate of 7980 failures per billion hours. That means about one PV panel failed every 14.3 years.

I don't know the FIT rates for the mounting system, wiring and breakers, but I will assume they are very low and I will use 0 FITs for my calculations.

Given all that, I expect the overall FIT rate of our system to be about 22,470 failures per billion hours, or an MTBF of 5.1 years. So during the 20 year life expectancy of the microinverters, I expect to see about 3 failures with likely most of those being microinverters.

In reality, we have experienced the failure of one microinverter very early in its life, which was replaced under warranty.

FAILURE MODES

Each of the expected failures should only affect the production of a single panel, so three failures would reduce production of the array by only about 7%, even if not repaired. In reality, one inverter replacement has already been accomplished. Future replacements may or may not happen depending upon whether or not they occur under warranty and where the failures occur in the array. Some inverter failures would require a significant amount of effort to replace while others, like the one which already failed, are trivial to replace.

MAINTENANCE

This system is maintenance free. I do not clean the panels, but rather I count on the rain to wash them occasionally and the snow scrubs them about once or twice a year when it slides off.

SUMMARY

Life: 20 years
MTBF: 5.1 years
Failure modes: Benign
Maintenance: None

If you have an Enphase-based system, you should be able to estimate your MTBF by dividing 213 years (the MTBF of a single panel/inverter combination) by the total number of panels you have. However, since our PV panels and Enphase inverters all sit on our roofs, we can expect those who live in colder climates to experience higher MTBFs and those who live in hotter climates to experience lower MTBFs, just like we see with our LEAFs.

If you have any other type of system, you can make a similar type of calculation to estimate the MTBF of your system.

 

[kentuckyleaf]

Is there a concern abour hail damage to PV panels? I know the glass is designed to withstand hail, but at what threshold? In addition, the design life of a new stick built home is about 50 years. If the PV system is located on a roof, it would need to be removed when the home is at end of life. If the PV system is 50 years old when removed from a home that is at EOL, would it be placed on a new roof? I like this thread, let the information flow!

 

[RegGuheert]

Is there a concern abour hail damage to PV panels? I know the glass is designed to withstand hail, but at what threshold?

Most PV panels are designed for 1-inch hail.

 

[RegGuheert]

Let me also do the calculation for a fully off-grid system which I designed and built two years ago, about the same time as my own array. This system consists of the following:

- 20 Sharp NU-U240F1 PV modules arranged in five strings of four panels
- 1 Xantrex XW6048 Inverter
- 1 Midnite Solar Classic 200
- 8 Concorde PVX-4050HT SLA Batteries
- 1 Maxwell BMOD0063P125 Ultracapacitor

This solar array is in a field at a 50-degree elevation and the equipment is housed in an insulated but not climate-controlled room located below the array.

LIFE

Monocrystalline PV panels have an unknown life expectancy. Their electricity production gradually degrades over their long life.

XW series inverters come with the industry-standard five-year warranty. That said, they have proven to be extremely reliable in the field, even though they have been around for about 8 years. While I'm sure some have failed, I have not heard of any such failures. I will put the life at around 12 years.

Classic charge controllers were just coming out when I assembled this system. This one had, and still has remaining, some firmware issue which affect sytem reliability. More on that later. I went with this unit based on it's capabilities and design heritage. I will also peg this at about a 12-year life.

Concorde SLA batteries should have a life of about 5 years in this application. Hopefully we will achieve that, given that the load to date has been light. Unfortunately, the batteries did not come very well matched in voltage and the charge controller did not have coloumb-counting in place last time I checked, so there has been a bit of overcharge to date.

Maxwell ultracapacitors have a rated life of 100,000 hours at rated voltage, but it goes way up for low voltages like this application. I'm going to guess 30 years is easily achieveable.

MTBF

Industry tends to feel that PV panels have an MTBF of over 600 years, so I will use 600 years or a FIT rate of 190 failures per billion hours. That gives a fit rate of 3800 failures per billion hours. That means about one PV panel failed every 30 years.

I'll put the fit rate of the ultracapacitor at abou the same level as the PV panels: 3800 FITs or 30 year MTBF.

The batteries have a pretty low reliability. I read many stories about these failing after just two years due to overcharge and losing electrolyte, which cannot be replaced. We chose the sealed batteries because the owner did not want the maintenance associated with flooded batteries and because of the high impedance of flooded batteries under cold conditions. (The ultracapacitor was added later and slightly reduces this issue, but only for transient loads.) I'll but the MTBF of each battery at 20 years, or about 5700 failures per billion hours. 8 of them gives a total FIT rate of 45,600 failures per billion hours or an MTBF of about 2.5 years.

The XW6048 inverter appears to be very reliable in the field, so I will put its MTBF at about 12 years, comparable with its life. That is about 9,500 failures per billioon hours. (It has had a pretty hard life, to date, but that's a separate story!)

The Classic Controller has also had a good field history, but it has only been out there for two years. I have noticed Midnite has lowered the current limits on the new units to below that used in the controller we installed, which implies they have had reliability issues with the unit as installed. I will put MTBF at about six years, or a FIT rate of about 19,000 failures per billion hours.

I don't know the FIT rates for the mounting system, wiring and breakers, but I will assume they are very low and I will use 0 FITs for my calculations.

Given all that, I expect the overall FIT rate of our system to be about 81,700 failures per billion hours, or an MTBF of 1.4 years. So during the 5 year life expectancy of the batteries, I expect to see about 3 failures with likely most of those related to the batteries.

So far, we have not experienced any failures of any of the equipment after over two years of operation. The system is not being loadd by its intended load, since the house is not yet completed or occupied full time. Mainly, this system is used to support construction, including several 5-HP woodworking tools which have been a challenge for the XW to start reliably.

FAILURE MODES

This system is typical of many off-grid PV systems that employ SLA batteries with the addition of the ultracapacitor. That component should help to extend the life of any batteries used while only slightly increasing the overall failure rate. Basically, any failures in the PV array, which ar unlikely, can be isolated by switching off 20% of the array. This array is oversized, so that should not result in an operational failure. The ultracapacitor also can be pulled out of the system in case of a failure, but it's presence should help the life of the batteries. But a failure in ANY of the other components in the system will result in complete loss of the PV capability and the need to run from the propane-powered backup generator. As those components fail, they will either need to be replaced or repaired to restore system functionality. While the Classic controller may be repairable, the XW inverter and the batteries likely will need to be replaced. We'll see.

MAINTENANCE

This system is designed to be maintenance free. They do not clean the panels, but rather they count on the rain to wash them occasionally and the snow scrubs them about once or twice a year when it slides off.

This was a known trade-off against system reliability. As failures occur in the system, we will reassess the choices made and make adjustments if necessary.

SUMMARY

Life: 5 years for batteries and 12 years for inverter and charge controller
MTBF: 1.4 years
Failure modes: Severe
Maintenance: None

As you can see, I have made up some of the numbers for component failure rates based on my experience and that of others that I have read about. Unfortunately, with the exception of microinverter makers the manufacturers of PV power-electronic components do not like to talk about the reliability of their products. If anyone can provide manufacturer or other credible data to support the reliability of any of these components, please feel free to link it here.

 

[AndyH]

A view of MTBF from the vantage point of a former "Professional Quality and Reliability Engineer licensed by the State of California."

Suppliers in the solar industry quoting high MTBF numbers are misleading folks into expecting long system lives. Such statements are factually wrong. An MTBF of 600 years sounds great but 100% of the products may fail in a short time. Failures within 10 years are common for solar inverters. This is because wear out mechanisms determine the lifetime of products, and these failure mechanisms are not predicted by MTBF.

Data from the European Power Supply Manufacturers Association, an independent trade body, found that MTBF figures for the same product could vary by 10:1, depending on the methodology used in calculating them. This is further evidence that an over-emphasis on MTBF as a measure of real-world reliability is just marketing smoke and mirrors.

Warranties offered by companies on their products will increasingly be in the spotlight. However, this will be credible only when data is supplied to support such claims. MTBF predictions do not constitute credible data for this purpose. The PV industry will benefit from longer warranties but it's critically important to provide the underlying body of accelerated life test data to support such warranty claims.

http://www.renewableenergyworld.com...lity-a-misunderstood-relationship-in-solar-pv

I've said this in a number of places already, but one last try - 1. warranty is about 80 or 90% marketing and 2. MTBF (widely misunderstood and misused) can be manipulated to paint a picture that improves marketing while completely obscuring the 'facts on the ground.'

One is traveling down a dark road when they present manufacturer-provided MTBF and/or warranty numbers and then suggest that has a direct correlation to real-world performance.

http://engineering.case.edu/centers/sdle/sites/engineering.case.edu.centers.sdle/files/peshek.pdf

String inverter manufacturers find the reliability claims of microinverters dubious.

So do other microinverter manufacturers...
http://www.enecsys.com/resources/whitepapers/
http://solarbridgetech.com/wp-content/uploads/2011/05/SLB_E_Design_Reliability.pdf

Thanks Reg for providing a home for this line of conversation.

 

[AndyH]

One of the challenges I have with lines of reasoning presented on this forum is what appear to me to be a desire to promote a specific technology or method without regard to how the real-world use of that technology affects the outcome. Just one example is MTBF for PV panels.

PV modules now have long lifetimes with warranties offered up to 20 years and mean time between failures in the field of up to 522 years for residential and 6,666 years for utility systems [6].

[6] A. Maish, "Defining Requirements for Improved Photovoltaic System Reliability," Progress in Photovoltaics, pp. 165-173, 1999.
http://energy.sandia.gov/wp/wp-content/gallery/uploads/Flicker-PVSC-Cap-Paper-Final.pdf
(I wonder what type of panels were evaluated - mono- or poly-crystaline, or thin film? Does it matter? The paper doesn't say.)
http://www.osti.gov/bridge/servlets/purl/2677-MUMLuV/webviewable/2677.pdf


Another example is the use of the term "inverter" in other threads. There isn't a single type of inverter - there are different types (with or without transformers, for just one example), there are a number of types of cooling (passive, active fans, water cooling), in different types of service (grid connected no battery, grid connected with battery, not grid connected, residential/business, commercial/large scale). It is not reasonable to lump these together.

It's not enough to say "Vendor A says their system will last 3 million years" without also defining exactly what type of service they modeled when determining that number. Especially when it's probable that the numbers are as toasty as a solar panel in Tucson in July.

This paper proposes a new methodology for calculating the mean time between failure (MTBF) of a photovoltaic module-integrated or module-attached inverter (PV-MII). Based on a stress-factor approach, the technique invokes the usage model for the inverter to determine the statistical distribution of thermal and electrical stresses for the electrical components. The salient feature of the proposed methodology takes into account the operating environment volatility of the module-integrated (MI) electronics to calculate the MTBF of the MII. This leads to more realistic assessment of reliability than if a single worst-case operating point were used. Measured data (temperature and insolation level) is used to experimentally verify the efficacy of the proposed methodology. The results confirm that the electrolytic capacitor is the most vulnerable component with the lowest MTBF, but more importantly provide a quantified assessment of realistic MTBF under expected operating conditions rather than extrapolating a conclusion based upon a single worst-case operating point, which may have a low probability of occurrence.

http://ieeexplore.ieee.org/xpl/logi...re.ieee.org/xpls/abs_all.jsp?arnumber=6317934

This thread can be a useful, fact-based compilation of useful data, or can simply be a parroting of sales language and anecdote. Fact or sales - which way should we go?

 

[DesertDenizen]

I have a 4k system with BP Solarex Amorphous panels. It is a 10 year old system, back then it required 120 panels! But in 10 years I have had no maintenance and it is working just fine.

 

[RegGuheert]

I have a 4k system with BP Solarex Amorphous panels. It is a 10 year old system, back then it required 120 panels! But in 10 years I have had no maintenance and it is working just fine.

That's awesome! Is your system grid-tied or off-grid? What kind of inverter do you have?

The amorphous panels were supposed to lose some efficiency over time. Do you have any feeling for how much yours have lost?

 

[DesertDenizen]

I have a 4k system with BP Solarex Amorphous panels. It is a 10 year old system, back then it required 120 panels! But in 10 years I have had no maintenance and it is working just fine.

That's awesome! Is your system grid-tied or off-grid? What kind of inverter do you have?

The amorphous panels were supposed to lose some efficiency over time. Do you have any feeling for how much yours have lost?

I am on the grid, TEP, the electric utility here, had a program that gave a 30% discount off of panels if you stayed on the grid. I still have a usual bill of zero so I don't feel any loss. My all electric home and my EV charged by my panels so I am happy. I don't recall what my invertors are. I have four, one for each 30 panels.

 

[AndyH]

Is there a concern abour hail damage to PV panels? I know the glass is designed to withstand hail, but at what threshold? In addition, the design life of a new stick built home is about 50 years. If the PV system is located on a roof, it would need to be removed when the home is at end of life. If the PV system is 50 years old when removed from a home that is at EOL, would it be placed on a new roof? I like this thread, let the information flow!

The paperwork for my Evergreen and Sun panels say they will survive the IEC 61215 Hailstone Impact Test - a 25mm ice ball at 23 m/s (83 km/h, 51 mph). To survive, there can be no broken cells or glass.

http://webstore.iec.ch/preview/info_iec61215{ed2.0}en_d.pdf
http://www.nrel.gov/docs/fy12osti/54714.pdf

I've yet to read anything about PV panel end of life. According to an NREL comment, the first panels made are 45 years old and still making power. The 'official' end of life for panels and batteries is when they degrade to 80% capacity/capability. They can remain in service beyond that though if they're still providing the power or storage they need to provide.

 

[RegGuheert]

I am on the grid, TEP, the electric utility here, had a program that gave a 30% discount off of panels if you stayed on the grid. I still have a usual bill of zero so I don't feel any loss. My all electric home and my EV charged by my panels so I am happy. I don't recall what my invertors are. I have four, one for each 30 panels.

I figured you were connected to the grid since you said you had said you had your system for ten years with no maintenance and no problems. I think it's great that you don't even know what type of inverters you have. Frankly, that's how it should be! Put the system in place and forget about it.

Well, whatever inverters you have, they have achieved a demonstrated MTBF over 40 years! Kudos to the manufacturer who fielded that product. I hope you have many more years of trouble-free operation!

 

[RegGuheert]

So far no one in this thread has conflated MTBF with service life, nor has any manufacturer mentioned done so. These two reliability issues have been dealt with as the two separate things that they are. Any suggestion that MTBF is just some sort of marketing ploy by microinverter manufacturers to confuse potential buyers is sadly mistaken. If you take such a view, it would be easy to assume that a microinverter system would last for over 25 years without any failures. While that may happen, particularly with smaller systems, Enphase's calculations indicate it is unlikely in a large system. While I believe their calculations are based upon extreme conditions and are likely overly conservative for my installation, it appears they may not be in other locations.

In any system with many parts, BOTH MTBF and service life are extremely important considerations, as should be obvious from the calculations done in the OP for a microinverter-based system and the later one done for an off-grid system.

PV systems have had to deal with the concept of MTBF for many years since many PV installations require the installation of many individual PV modules. Fortunately, PV modules achieved extremely high MTBFs many years ago to the point where their failure rate was not an issue in most applications. In order for a microinverter-based system to be a viable alternative, it is a basic requirement that the microinverter have BOTH a long service live AND a high MTBF. Unfortunately, inverters in PV systems have traditionally had neither, typically suffering frequent failures and short service lives.

Enphase Energy took on the challenge of providing microinverters into the PV market by developing an advanced design and building it using advanced manufacturing technologies. They have backed up their claim of a long service life with a detailed analysis of the life of the electrolytic capacitors in their product, explaining exactly the operation of capacitor in the inverter, the environment and the failure modes that are expected. This analysis was backed up by an independent study of the life expectancy of these capacitors that was commissioned by a potential investor as part of their due diligence. They have also discussed both the service life expectancy and the MTBF calculation they have made for their microinverters in a white paper which covers their analysis approach, testing and conclusions. In addition, that white paper explain the benefits of microinverters in overall system availability.

I have confirmed the MTBF of Enphase microinverters to be as high or higher than Enphase's calculations, at least outside of Arizona. Right now, my analysis shows an overall MTBF of over 315 years for 20 systems spread throughout North America. That said, this analysis indicates to me that the MTBF may be significantly higher that stated, especially outside of Arizona. Of the four failures seen, two are in Phoenix and one is in Tuscon. Are those failures simply due to the product being installed in an environment outside the capability of the inverter? I certainly do not know, but it is a definite possibility. Per Enphase's specifications, the inverter is for use only up to 65C (149F) ambient temperatures. If these microinverters have been exposed to higher temperatures than that, then any reliability calculations do not apply. Likewise, the one failure in my system could have been due to infant failure or misapplication. Still, I have included all four of these failures in my analysis for completeness and to allow anyone to choose to include them in their analysis or to only include the environments which better match their applications. Sadly, I don't think my analysis can be carried too far into the future as Enphase will be turning off public access to the failure information which I am using. It's too bad, since it is useful information, both for those in AZ who may wish to avoid this product until more is learned and for those elsewhere who want to get a feel for how long these microinverters will last.

Simply put, Enphase Energy has taken up the flag on the biggest failing of the BOM suppliers in the PV industry: reliability. They have picked up where SMA lead by moving to grid-tied string inverters to eliminate the batteries from the system BOM. SMA's Sunny Boy inverters have perfomed well, particularly when compared to the offerings which Xantrex fielded at about the same time. Still, with a self-proclaimed field MTBF of only 11.4 years, these components often require replacement multiple times within the life of a modern PV system. DesertDenizen's experience is an exemplary one, with a demonstrated MTBF of over 40 years. Likely his inverters are made by SMA, so this bodes well for them in the desert, where microinverters may be struggling for life on the hot rooftops there.

Off-grid PV systems have always been plagued by the reliability issues concerned with the storage batteries and still are. The off-grid BOM manufacturers have sometimes used this as their excuse to not offer long life and high MTBF in their inverter and charge controller offerings. This fact is reflected in their poor field performance and short warranty periods which remain at only five years to this day. These off-grid BOM companies also refuse to publish any information about either the measured or the calculated service life or the MTBF of their products. As a result, anyone wishing to field an off-grid system is forced to work entirely from anecdotal information about service life which is, at best, biased. In the end, neither the manufacturers nor those who lauded these systems in internet forums will pay for the maintenance and repairs that these systems will require after their short warranties expire.

So are microinverters the end-all in PV systems? Certainly not! First of all, they do not provide grid back-up capability nor are they useful in locations without grid access. As noted, I installed a fully off-grid PV system at about the same time I installed our microinverter-based system. But microinverters, with their expected life over 20 years and an MTBF of over 300 years are very attractive for homepower systems but may not be as attractive for commercial applications unless the MTBF can be significantly enhanced. As you can see from the OP, even the stated level of reliability, if achieved, would result in about 5% failures within 20 years. That is OK if you install 50 inverters, but if you install 50,000, then you have a real maintenance headache (unless you choose to ignore the failures). That is why central grid-tied inverters are the mainstay of large installations of hundreds of kW of PV. But that's a big market which is attracting other investment. It is currently being challenged by the mini-inverter. They claim to have the highest reliability and all of the benefits of microinverters (and more) at the cost of central inverters. It will be interesting to see how this segment plays out going forward.

 

[AndyH]

Reg - do you have any ties to Enphase? Because the story you're telling is 100% marketing!

Enphase did not invent the microinverter - there were at least two previous microinverter lines on the market and they failed miserably. I owned some of the 2nd generation devices. I agree completely that Enphase has done a masterful job of creating a market and dominating it. Let's not forget that they had to quickly abandon their first two iterations (the ones with 10 year and 15 year warranties) in order to increase reliability. The only thing I've been able to find that suggests these inverters are physically, mechanically, thermally, or superior in any other way is that one report suggests they doubled the critical electrolytic capacitors so that there is an on-line backup if one fails. That should ease the internal heating on the electrolytic caps, but at the price of an increased parts count and introduction of another (though non-critical) failure item - yet does nothing about wear-out.

They're not even using the latest or most efficient tech. There's a Canadian microinverter company that's moved away from high frequency switching and has eliminated electrolytic caps from their design. Were I in the market for a microinverter, I'd be importing them from Canada as that's a product that I think can stay on the roof without failing.

Yes, MTBF and other reliability metrics can be useful in the right situations. They can also be twisted out of shape and used as a marketing tool. And again - MTBF is NOT useful for predicting wear-out/end of life, or of predicting the behavior of any specific system. MTBF is like the general understanding that in Vegas, the house always wins. Yet some can beat the odds on a regular basis within that environment.

It's my opinion, based on reading Enphase's materials, info from other manufactueres, results of US National Lab tests and conferences, that Enphase's claims are...optimistic, not supportable by their tech, and that time will show that their marketing crew is more important to their financial success than the folks in engineering.

Can you show a patent, design element, or any other proof that their inverters are not simply a smaller version of a conventional inverter? Can you show anything that confirms these devices are truly better than current art?

Good luck with your system - I hope it serves you well for a very long time. But please do not force-feed any of us a marketing message without also giving out free antacids.

 

[AndyH]

http://www.pv-magazine.com/services...ility-for-100-years-_100009311/#axzz2OJPT8ou5

The low malfunction rates of the installed devices results in a MTBF value of 100 years for the SolarMax S series. The MTBF value (Mean Time Between Failures) is the expected operating time between two consecutive failures – subject to ideal installation conditions (inside installation, clean/dry environment).

"The MTBF value is only a statistical measure for the malfunction rate and only to a certain degree allows conclusions regarding the operating time of the inverters to be reached, but it nevertheless demonstrates the potential of solar technology," explains Hans-Georg Schweikardt, head of product management with Sputnik Engineering. "The MTBF value of the S series is calculated on the basis of the devices installed on site and both the plant and the grid operator are provided with continuous output performance and, thus, high reliability in terms of grid feed-in."

These are current tech transformerless grid-tied string inverters from a company making inverters for 20 years. One percent annual real-world failure rate.

Does that at least kill the myth that string inverter companies don't report MTBF or evaluation conditions?

 

[RegGuheert]

Reg - do you have any ties to Enphase?

Yes. I am their customer.

Because the story you're telling is 100% marketing!

Enphase did not invent the microinverter - there were at least two previous microinverter lines on the market and they failed miserably. I owned some of the 2nd generation devices.

So you had a bad previous experience with another manufacturer's microinverters and you believe that all microinverters are created equal? Is that it?

I agree completely that Enphase has done a masterful job of creating a market and dominating it. Let's not forget that they had to quickly abandon their first two iterations (the ones with 10 year and 15 year warranties) in order to increase reliability.

Abandon?? I think that is a very poor characterization of what happened. The second iteration with the 15-year warranty is what is on my roof and the roofs of many others. Most of the inverters that are in my spreadsheet demonstrating the impressive MTBF are M190s.

There were several reasons Enphase redesigned the M190 into the M215. First, they needed to increase the power capability of their inverters. I'm sure enhanced reliability was one of them since their stated goal was to achieve a 600-year MTBF. They will need at least that to compete in the commercial PV space. I think they will need much higher numbers than that. But they also had an issue with their wiring system. The old system required 15A circuit breakers. I had a HARD TIME finding 240V 15A breakers for my system! With the higher power rating of the M215 the 15A limit would have been even more frustrating.

The only thing I've been able to find that suggests these inverters are physically, mechanically, thermally, or superior in any other way is that one report suggests they doubled the critical electrolytic capacitors so that there is an on-line backup if one fails. That should ease the internal heating on the electrolytic caps, but at the price of an increased parts count and introduction of another (though non-critical) failure item - yet does nothing about wear-out.

Incorrect. More later.

They're not even using the latest or most efficient tech. There's a Canadian microinverter company that's moved away from high frequency switching and has eliminated electrolytic caps from their design. Were I in the market for a microinverter, I'd be importing them from Canada as that's a product that I think can stay on the roof without failing.

They will have to execute well to enter this market, regardless of their technology.

Yes, MTBF and other reliability metrics can be useful in the right situations. They can also be twisted out of shape and used as a marketing tool. And again - MTBF is NOT useful for predicting wear-out/end of life,

Why do you keep saying that? No one has ever said they were.

or of predicting the behavior of any specific system.

This is incorrect. Your statement implies that you believe that there are no failures between infant mortality and wear out and/or you believe that statistical analysis can tell us nothing about an individual design. Both of those things are patently false. While you cannot predict exactly when things will fail due to MTBF numbers, you CAN predict how often they will fail, assuming you have accurate MTBF data. I will grant that my MTBF estimates are likely conservative when applied to my system, but that is still useful for planning purposes.

MTBF is like the general understanding that in Vegas, the house always wins. Yet some can beat the odds on a regular basis within that environment.

It's my opinion, based on reading Enphase's materials, info from other manufactueres, results of US National Lab tests and conferences, that Enphase's claims are...optimistic, not supportable by their tech, and that time will show that their marketing crew is more important to their financial success than the folks in engineering.

You are welcome to your opinion. But Enphase has backed up their claim of a long life by showing through analysis how their capacitors can last at least 30 years. And I have backed up their MTBF claims with actual hard data. Unless you can explain what is wrong with these two analyses, your opinion is just that, your opinion.

Can you show a patent, design element, or any other proof that their inverters are not simply a smaller version of a conventional inverter?

This is the KEY point which you are missing! It is simply a smaller version of a convensional inverter! Simply by processing lower voltages and currents, the microinverter can achieve drastically superior reliability. At higher power levels you cannot achieve the same amount of component derating as you can at lower power levels. But you don't need to believe me. Simply read one of the references which you provided in this (and other) threads:

http://engineering.case.edu/centers/sdle/sites/engineering.case.edu.centers.sdle/files/peshek.pdf

Strong dependence on ratio of applied voltage to rated voltage, and rather weak dependence on capacitance.
Commercial eCaps only rated to 630 V maximum Common string voltages = 500-600 V

At 75 C, String FITs = 800, microinverter FITs = 30
Microinverter requires more components, but more balanced.

Basically, as a designer, you can design specifically for reliability. If you are designing a microinverter, you can design in more margin since you have access to components with higher ratings relative to your requirements.

Let's use an example that everyone on this forum can appreciate: Let's assume the Nissan LEAF is the BEST EV available for designers to apply to their task. Now, let's give two designers two tasks: One is to design an electric transportation system to carry one person 25 miles round-trip each day while the other is to design an electric transportation system to carry one passenger 70 miles each day. Both designs are for the same city, say Washington, DC. Now, the first designer could choose a cheaper EV to meet his goal, but he could also choose to pay a little bit more and use the LEAF for the task. If he does that, he can probably provide a solution that will last for 20 years while the other designer likely can only provide a solution which will work for only about three years. It literally will fail to work after three years, while the first designer can build one which will last for 20. This is true even though both designers had access to exactly the same technology. It's just that the first designer had a less stressful application, meaning his EV would wear more slowly AND it could be much more degraded before it would fail to meet its objectives.

But as you have seen, that doesn't mean that a microinverter cannot fail. Designing power electronics is very difficult because you must maintain ALL components within their safe operating area during normal operation and during abnormal conditions. It is not an easy task, since it is sometimes hard to even know what the abnormal conditions might be. The simple fact is that Enphase has executed very well on their products.

All of this has very little to do with marketing. This is engineering, but not something that most people realize engineers do.

Good luck with your system - I hope it serves you well for a very long time.

Thanks! You, as well!

 

[RegGuheert]

Does that at least kill the myth that string inverter companies don't report MTBF or evaluation conditions?

It sure does!

Who started the myth?

 

[AndyH]

Does that at least kill the myth that string inverter companies don't report MTBF or evaluation conditions?

It sure does!

Who started the myth?

Unfortunately, with the exception of microinverter makers the manufacturers of PV power-electronic components do not like to talk about the reliability of their products.

 

[QueenBee]

But they also had an issue with their wiring system. The old system required 15A circuit breakers. I had a HARD TIME finding 240V 15A breakers for my system! With the higher power rating of the M215 the 15A limit would have been even more frustrating.

Really? I actually used 15 amp breakers on my M215 circuits just because I at max output I wasn't even at 80% of 15 amps. I don't remember it being particularly hard to find Siemens (GE style) breakers though for my subpanel though.

 

[AndyH]

Reg - do you have any ties to Enphase?

Yes. I am their customer.

Because the story you're telling is 100% marketing!

Enphase did not invent the microinverter - there were at least two previous microinverter lines on the market and they failed miserably. I owned some of the 2nd generation devices.

So you had a bad previous experience with another manufacturer's microinverters and you believe that all microinverters are created equal? Is that it?

Sales tactic one: Trivialize objections

The answer is no, by the way. I'm NOT saying that all micros are equivalent or that Enphase has a poor product. Nor am I suggesting that anyone else has a superior product (well, maybe the Canadian manufacturer with the superior product , but I'm not suggesting that many/most string inverters are superior WRT lifespan). What I am saying is that there's no apparent direct connection between their marketing speak, their tech, and their life estimates.

I agree completely that Enphase has done a masterful job of creating a market and dominating it. Let's not forget that they had to quickly abandon their first two iterations (the ones with 10 year and 15 year warranties) in order to increase reliability.

Abandon?? I think that is a very poor characterization of what happened.

Sorry, abandon isn't the correct word. Superseded. Rapidly replaced on the market. Liability tail cut as quickly as possible.

The second iteration with the 15-year warranty is what is on my roof and the roofs of many others. Most of the inverters that are in my spreadsheet demonstrating the impressive MTBF are M190s.

How does the mid-day power clipping affect the energy harvest over time?

There were several reasons Enphase redesigned the M190 into the M215. First, they needed to increase the power capability of their inverters. I'm sure enhanced reliability was one of them since their stated goal was to achieve a 600-year MTBF. They will need at least that to compete in the commercial PV space. I think they will need much higher numbers than that. But they also had an issue with their wiring system. The old system required 15A circuit breakers. I had a HARD TIME finding 240V 15A breakers for my system! With the higher power rating of the M215 the 15A limit would have been even more frustrating.

The only thing I've been able to find that suggests these inverters are physically, mechanically, thermally, or superior in any other way is that one report suggests they doubled the critical electrolytic capacitors so that there is an on-line backup if one fails. That should ease the internal heating on the electrolytic caps, but at the price of an increased parts count and introduction of another (though non-critical) failure item - yet does nothing about wear-out.

Incorrect. More later.

They're not even using the latest or most efficient tech. There's a Canadian microinverter company that's moved away from high frequency switching entirely and has eliminated electrolytic caps from their design. Were I in the market for a microinverter, I'd be importing them from Canada as that's a product that I think can stay on the roof without failing.

They will have to execute well to enter this market, regardless of their technology.

Yes, MTBF and other reliability metrics can be useful in the right situations. They can also be twisted out of shape and used as a marketing tool. And again - MTBF is NOT useful for predicting wear-out/end of life,

Why do you keep saying that? No one has ever said they were.

Cooked numbers: You didn't follow the links I posted above, then. See below as well.
End of life: My desire from the start of the "MTBF chronicles" has been only service life. That's when you and others dropped MTBF on the table to 'prove' that a longer MTBF means a longer life.

or of predicting the behavior of any specific system.

This is incorrect. Your statement implies that you believe that there are no failures between infant mortality and wear out and/or you believe that statistical analysis can tell us nothing about an individual design.

No. What I AM saying is that while MTBF can provide an idea of how many failures might happen between the end of infant mortality and the beginning of break down for a large population, it cannot provide info on breakdown, and cannot provide any info on how any specific installation/system will perform. It can give odds on how many devices might make it to breakdown, but cannot say how many minutes/days/years are between the 'infant' and 'break' goal posts.

Both of those things are patently false. While you cannot predict exactly when things will fail due to MTBF numbers, you CAN predict how often they will fail, assuming you have accurate MTBF data.

Tell me then - how often will your inverter number 14 fail? All you can do is present odds. All of your equipment might fail on the same night (not likely) or all of your equipment might be providing power 40 years from now (not likely either). Your inverters might reflect three or four times as many failures on average, but it's ok because maybe the guy in North Dakota has a lower failure rate.

(end of part one...)

 

[AndyH]

(part two...)

I will grant that my MTBF estimates are likely conservative when applied to my system, but that is still useful for planning purposes.

MTBF is like the general understanding that in Vegas, the house always wins. Yet some can beat the odds on a regular basis within that environment.

It's my opinion, based on reading Enphase's materials, info from other manufactueres, results of US National Lab tests and conferences, that Enphase's claims are...optimistic, not supportable by their tech, and that time will show that their marketing crew is more important to their financial success than the folks in engineering.

You are welcome to your opinion. But Enphase has backed up their claim of a long life by showing through analysis how their capacitors can last at least 30 years.

Yes, they have a nice assortment of sales literature. Can you show any independent testing that confirms the claims?

How "reliable" is "reliable"? - Create a definition and metric for reliability
"failure" needs to be defined

Warning - MFR, can/will/sometimes - game the system. Cost and time are huge factors

Disconnect of what a performance test buys and the expected life of the product

These points are from a 2011 inverter reliability conference and were results of surveys of inverter manufacturers, integrators, end users, engineers, and researchers. There's no standard definition of failure, there are no required tests that provide insight into field life, and manufacturers game the system - and that includes cooking the MTBF books.
http://energy.sandia.gov/wp/wp-content/gallery/uploads/Inverter_Workshop_FINAL_072811.pdf

As for those 30-year electrolytic caps, how many hours is that? Can you show me a single electrolytic capacitor on the market that can stay in service that long? There must be no internal heating - no ripple - for that to happen - at the very least they'll dry out simply from being on a roof for 25 years.

And I have backed up their MTBF claims with actual hard data. Unless you can explain what is wrong with these two analyses, your opinion is just that, your opinion.

Do you mean the MTBF comparison using a dozen field installations as compared with numbers from Enphase that are under their control? No - not a significant dataset yet.

Can you show a patent, design element, or any other proof that their inverters are not simply a smaller version of a conventional inverter?

This is the KEY point which you are missing! It is simply a smaller version of a convensional inverter! Simply by processing lower voltages and currents, the microinverter can achieve drastically superior reliability. At higher power levels you cannot achieve the same amount of component derating as you can at lower power levels. But you don't need to believe me. Simply read one of the references which you provided in this (and other) threads:

http://engineering.case.edu/centers/sdle/sites/engineering.case.edu.centers.sdle/files/peshek.pdf

Strong dependence on ratio of applied voltage to rated voltage, and rather weak dependence on capacitance.
Commercial eCaps only rated to 630 V maximum Common string voltages = 500-600 V

500-600 volts is common in industrial scale systems but NOT in home-scale or off-grid service! A commercial cap with a 630V rating should be quite happy with a 150V input, right? How about if the input is kept below 90VDC? Again - tell me how this shows that Enphase should have a longer life than a string inverter?

At 75 C, String FITs = 800, microinverter FITs = 30
Microinverter requires more components, but more balanced.

The only way to de-rate a component so that it'll last tho or three times longer than its known calendar life is to not energize the circuit that component is soldered into. And don't put it on a hot roof. :lol:

Basically, as a designer, you can design specifically for reliability. If you are designing a microinverter, you can design in more margin since you have access to components with higher ratings relative to your requirements.
<snip>

In general, maybe. But component derating cannot fix all of the problems with aging - from solder joints, circuit board expansion/contraction (remember - roof top), and additional stresses from potting.

According to Sandia and other related conferences, folks in industry complain that theirs is a very small market - even when incorporating EV and hybrid inverters - without enough 'pull' with component manufacturers to make better components - and that's been an overall comment, not just one from string/central inverter manufacturers. Something not right about this puzzle.

Look at slide one in the 'supplementary material' section. Note the electrolytic capacitor failure curve starting at year 7. Also note the comment along the 100% failure line at about year 10: "Contain evaporating fluids" and "25 years warranty?" (That's 'conference speak' for "What are THEY smoking?!" :lol:
http://www.dfrsolutions.com/uploads/courses/SPI2012.pdf

This begs the question: If micros ARE just regular inverters that have gone through a shrink ray, and if, as you've repeatedly suggested, that regular (string) inverters only last 10-12 years, and if electrolytic capacitors are the first failure item, how can Enphase keep the electrolyte inside their capacitors well beyond the apparent capabilities of the other engineers?

I found this to be an interesting look at microinverter history:
http://matter2energy.wordpress.com/2012/04/09/the-great-microinverter-debate/

Sorry, had to break this up. The board didn't like the length, and there were bad characters in a Sandia quote that took some trial and error to find.