How did you decide? Microinverter vs Optimizer & PV Inverter

[entgegnen]

[QUOTE=sensij;n320402]

Sunpower's approach to installing microinverters is no different than Enphase's. Neither is more integrated than the other. Your comparison here is erroneous.

I am always happy to be corrected. But here is the thing - when you google search images of an enphase micro inverter installations and compare those photos to the Sun Power module, there is a difference. I've looked at various diagrams (from the respective companies) and numerous photos (on google images) to reach my conclusion that the physical integration of the Sun Power module onto the actual frame of the panel/module means that there is likely going to be more heat transference on account of actually being mounted to the frame (right below the wafer). By contrast, the Enphase is first attached to the rails... and based on it mounting height, it is usually .5 to 1-inch below the module frame (or at least the wafer). Heat doesn't transfer through air well. Hence, I think the boiling pot of water - making the pot handle hot as well - is an appropriate analogy.


I bet you didn't meet the person actually working on your roof until the day of the install. The slick sales guy or even the company owner is rarely the one who actually does the work, sub-contracting or not.

Tough audience. But hope springs eternal. I'll be sure to let you know what ends up happening.

[sensij]

The solar panel is likely to be hotter than the inverter, or at best, they will be the same temperature. I don't see a significant advantage in the thermal behavior for either one. I would guess most of the heat transfer will be radiative in either case.

[J.P.M.]

The solar panel is likely to be hotter than the inverter, or at best, they will be the same temperature. I don't see a significant advantage in the thermal behavior for either one. I would guess most of the heat transfer will be radiative in either case.

FWIW, and if interested: For two bodies with flat surface that have a common plane, any heat transfer between them will be totally radiative in nature if the surfaces are not in contact.

If in contact or clamped/screwed with no "visible" gap, the heat transfer between two bodies will be mostly conduction heat transfer, but not as much as if, say, glued via thermal contact cement, or, for metals, welded/soldered/braised/etc. While clamping leads to a relatively poor heat transfer (not entirely unlike increased electrical resistance for loosely/poorly connected electrical contacts), heat transfer in the small micro to panel gap case will still probably be conduction governed because the effectively zero air gap will prevent convection cooling between the two facing surfaces, effectively (assuming the panel is warmer than the ambient surroundings, and the micro is cooler than the panel) increasing the micro temp. While smaller than bonded surface conduction, the conduction that is present in the simple clamped case is still likely to be an order of magnitude or so greater than any convective heat transfer from the back/sides of the micro (unless the micro has fins on it). That will tend to get the two surfaces closer in temp. For a standard (say, no fins, clamped, and little air gap) scenario, the micro and the panel will probably be closer in temperature to one another than if an air gap exists between them. If so, radiation heat transfer between the two bodies will be less than if there is a, say, a 1 cm. gap or so between them.

That, however, is a somewhat different discussion than micro inverter temperature. IF there is an air gap (say, a cm. or so) between the inverter and the panel, convective cooling, either natural (gravity driven and probably relatively minor), or wind driven (and larger), such cooling will reduce the temp. of the micro. That will increase the radiative heat transfer due to the increased temp. diff. - panel to micro. Under steady state conditions, the relative amounts of heat transfer between the two bodies via radiation and conduction will be such that the amount of entropy increase due to each will be equal.

[DaveDE2]

Whilst solar panels run hot (say 60C max or so), that is relatively cool in the semiconductor world and that's why they can live almost forever. If an inverter living under said solar panel reaches similar temps, then its internal components will be much hotter (because they are generating heat), especially if one considers integrated circuit thermal resistance to die temps along with all of the other components, caps, etc. It would be best to decouple the inverter from solid contact with the panel and have it live on the rail where at least convective heat transfer can carry heat away better. Radiative heat transfer goes as T cubed and is probably not a factor. It's mostly all about convection imo. This all assumes that the rail is cooler than the panel which I think it probably is and that the inverter will be more exposed to air flow if it's not integrated with the panel.

[J.P.M.]

Whilst solar panels run hot (say 60C max or so), that is relatively cool in the semiconductor world and that's why they can live almost forever. If an inverter living under said solar panel reaches similar temps, then its internal components will be much hotter (because they are generating heat), especially if one considers integrated circuit thermal resistance to die temps along with all of the other components, caps, etc. It would be best to decouple the inverter from solid contact with the panel and have it live on the rail where at least convective heat transfer can carry heat away better. Radiative heat transfer goes as T cubed and is probably not a factor. It's all about convection.

I agree it's probably better to separate the panel from the inverter and give more available area for convective heat transfer. How much separation and in what ways can be an interesting design activity.

However, and contrary to what you write, the total emission of radiation from a body is proportional to the 4th power of its absolute temp., not the cube. The 4T^3 approx. may work when temps. are close, but it's an unnecessary approx. since the advent of calculators with a power function button. The radiant exchange between two surfaces is actually proportional to the difference in the 4th power of the absolute temp. of the two surfaces that can "see" one another in radiative exchange, not the third power of the average of the two. View factors and the idea that nothing is a perfect blackbody radiator (emissivity <1.0 and f(wavelength)) as well as other things also need consideration.

As for it being all about convection, often not. For example, and disregarding the relatively small amount of conduction through the racking, when doing a heat balance on the PV array sitting on my roof when under something like a full sun, depending on wind velocity and dew point (often used to estimate the effective radiant sky temp), about 40-60% of the heat rejected to the surroundings is from convection. The rest is from thermal radiation - or, the other way around if a dry day with little wind. Often, in heat transfer calcs.,the radiant heat transfer is expressed as a heat transfer coeff. per degree of temp. diff. between the body and the surroundings and then added to the convection coefficient to approximately linearize the overall heat transfer rate per degree. For a lot of work, heat exchanger design, process work, etc., where radiation heat transfer is relatively small, that's usually not a problem. For other work such as terrestrial solar or in space, or other places where thermal radiation becomes important. An example, heat loss through a single glazed window: 70 F. inside, 20F. outside, say 5 MPH wind outside. Convection coeff. ~ 1.0 BTU/hr.ft.^2/deg. F. radiation loss ~~ 44 BTU/hr. ft.^2, or ~~44/50 = ~ 0.7 BTU/hr. ft^2/deg. F. when linearized per deg. F and a .88 emissivity is considered.

The two mechanisms, convection and thermal radiation, or all 3 heat transfer mechanisms including conduction, need to be considered separately and treated as such if warranted, but it ain't a slam dunk, thermal radiation doesn't matter type situation.

[inetdog]

To a first approximation, when A is close to B, A⁴ - B⁴ is proportional to 4(A-B) (Average)³
As the temp difference gets very large the heat transfer becomes proportional to the fourth power of the higher temperature. In between, it is messier.

[J.P.M.]

To a first approximation, when A is close to B, A⁴ - B⁴ is proportional to 4(A-B) (Average)³
As the temp difference gets very large the heat transfer becomes proportional to the fourth power of the higher temperature. In between, it is messier.

Which is easier on a calculator ?

[HWD]

I have noticed a few people saying that they have compared the financials of Enphase to SolarEdge in their purchase decision. For those doing this, make sure you look at the most recent quarters. My concern is that people may be checking the last reported full (fiscal) year for both companies – the 12 months to Dec’15 for Enphase, and the 12 months to Jun’15 for SolarEdge.

As a quick summary – Enphase lost $19m in full year 2015. But if you look at 1Q16 (to Mar’16) and 4Q15 (to Dec’15), it made losses of $19m and $15m respectively each quarter. i.e. its current runrate means they are losing >$60m per year. This is for a company that had just $13m cash as of the latest quarter (versus $28.5m at the end of 2015)

SolarEdge made a profit of $19-20m in the last two quarters – i.e. run-rate profit of almost $80m versus $28.2m in its last fiscal year. It has net cash of $171m. Its average price per watt is less than Enphase’s COGS per watt.

Basically – either Enphase changes something drastic soon or it’s going bankrupt

[Engineer]

I have noticed a few people saying that they have compared the financials of Enphase to SolarEdge in their purchase decision. ..

Well they had some layoffs, have introduced storage which is getting a huge response, and are coming out with the S300 which brings cost par with strings (main reduction is in the case which is a significant component of the case). Their projection is that they have a roadmap to bring the micros below string cost and they believe the strings won't be able to lower costs equivalently.

People on this board should be cheering for Enphase/Micros. It's not a good thing actually that the industry is facing cut-throat end pricing, it can ultimately be bad for the market. What you want is several competitors who all have good profit margins - that's where new R&D funding comes from which delivers better products. If SolarEdge was the only choice they'd cut R&D funding to nearly zero*. Generally when the companies are forced to cut costs to the bone by the end customer you get bankruptcies and mergers that result in less innovation and choice, not more.

I don't know how it will turn out in this industry but I lived through the same process in a different industry (semiconductor test). It's not fun, but more importantly is often ultimately bad for the customers.

* A point first made clear to me early in my career when I complained that we had competitors. An older engineer said "yeah, and if they didn't exist you wouldn't have a job"

[ButchDeal]

Well they had some layoffs, have introduced storage which is getting a huge response, and are coming out with the S300 which brings cost par with strings (main reduction is in the case which is a significant component of the case). Their projection is that they have a roadmap to bring the micros below string cost and they believe the strings won't be able to lower costs equivalently.

where do you get the AC battery has a huge response? They have barely opened any markets with it, and it is lacking in the primary feature that people want with a backup system, BACKUP.

[Engineer]

where do you get the AC battery has a huge response? They have barely opened any markets with it, and it is lacking in the primary feature that people want with a backup system, BACKUP.

Many links in a thread on this board, both from the company and from distributors, a search will find it. Response was, I forget the number but twice what they expected or thereabouts.

It's not a BACKUP system and is not designed or advertised as such, do some research and you'll understand it better. It is opening in Australia only at first for good reasons, the same ones every other manufacture is doing the same thing. US release is later this year.

[ButchDeal]

Many links in a thread on this board, both from the company and from distributors, a search will find it. Response was, I forget the number but twice what they expected or thereabouts.

It's not a BACKUP system and is not designed or advertised as such, do some research and you'll understand it better. It is opening in Australia only at first for good reasons, the same ones every other manufacture is doing the same thing. US release is later this year.

twice a small number is still a small number They are late brining the product to market, day late, few dollars short so to speak.

I understand it very well. Yes they specificially do not mention that it backs up BUT they attempt very well to hid that it does NOT.
Almost every customer asking about batteries is interested in backup.


Expecting a limited capability battery system to save the company from poor financials is pretty far fetched.
Do some research and you will understand the product, market, industry better.

[ButchDeal]

many would disagree with you there. Most people looking for backup with battery particularly in coastal are ok with limited power (no AC or other big loads). so a 400-600ah 48V system works well. They generally have sun after the storm in some form so 2-3 days of no sun run time generally equates to pretty much indefinite.
All common backup solutions have critical load panel.

I never stated that this was enough to swing the market. Others stated that battery would save enphase, and I pointed out that currently the largest residential battery market in the US is for backup, which enphase is not hitting.
Enphase system could be adapted to be a backup solution but it would be much more complicated in that role.

[ltbighorn]

A 20-30 kWh battery backup system is considerably more massive (and expensive) than what would be needed for load shifting (maybe 4-6 kWh). I suspect that the current residential battery backup market is relatively small compared to the potential market for load shifting. Markets really open up when a system goes from being an expensive convenience/luxury, to paying for itself and generating ongoing revenue. Just my impression.

[ButchDeal]

a 20-30kwh lead acid battery will be more massive yes but definitely not cost more. Who really cares about the mass, though, it isn't being moved every day like in a car.... A load shifting battery would be more costly as it would be cycling daily. 4-6kwh isn't going to load shift much. If the backup battery is for critical loads and is lead acid so 50% dod comes to 10-15kwh critical. for load shifting you are on full load but just for part of day, so you are likely looking at similar.
You also have the other option which is grid zero situations but again a very limited market in HI, and a few Coops here and there without net metering (to be fair enphase should be able to handle grid zero too)

[entgegnen]

I for one an "cheering" for both companies and both technologies. I'd rather folks have good jobs than not. But obviously, company strength does come into play when considering a system that has a 25-year shelf life. It does seem that Solar Edge has more proprietary stuff, and that the apparent diversity of micro converter companies is a strength (if not for Enphase, for the technology). That being said, the strategic partnership between Solaredge and Tesla could also be a great longevity factor for product diversity (bolstering Solar Edge).

One comment above noted the financials. I too noticed the quarterly direction. (Although I had not fully researched any markdowns, potential litigation, or other reasons for it. There could be reasonable reasons for the direction. But this late into company existence is usually not the best sign. I am not a financial advisor.)

The discussion of heat transference has quickly gone over my head. But I did notice a comment on a panel/module reaching 60 centigrade (140 fahrenheit for us non-science types) Based upon the relative positioning of the physically integrated micro inverter I have no reason to believe it would be a temperature of anything much less than 140 farenheit. By contrast, a rail attached module like Enphase would seem to be 10-15 degrees fahrenheit cooler. (do to air, being in the shade of the module/panel etc. Maybe 10-15 degrees doesn't matter. I just suspected it would over time. I've seen anecdotal posts of panels/modules getting as hot as 170 degrees in southwest summers. FWIW, Enphase literature does mention conducting testing that included 50 days at a constant 185 degrees fahrenheit (for something called an IEC 61215 test). So maybe they are built to hold up.

[DaveDE2]

I'll bet there're guys here who live in the hot southwest who can tell us based on panel voltage pretty accurately what temps panel silicon can reach. I'd like to get an inkling of that. Silicon diode voltage decreases by about -2.5mV/C. The actual frame of the panel will be somewhat cooler.

[sensij]

The discussion of heat transference has quickly gone over my head. But I did notice a comment on a panel/module reaching 60 centigrade (140 fahrenheit for us non-science types) Based upon the relative positioning of the physically integrated micro inverter I have no reason to believe it would be a temperature of anything much less than 140 farenheit. By contrast, a rail attached module like Enphase would seem to be 10-15 degrees fahrenheit cooler. (do to air, being in the shade of the module/panel etc. Maybe 10-15 degrees doesn't matter. I just suspected it would over time. I've seen anecdotal posts of panels/modules getting as hot as 170 degrees in southwest summers. FWIW, Enphase literature does mention conducting testing that included 50 days at a constant 185 degrees fahrenheit (for something called an IEC 61215 test). So maybe they are built to hold up.

The discussion went over your head, but you are still offering estimates on how much cooler one microinverter would run than another? Nice. In your engineering assessment of the installation, are you making an adjustment for a difference in conductive heat sinking based on the contact area on the frame relative to the contact area on the rail? That 10-15 deg number is completely made-up... it is fine for you to use it in your own decision making, but I want to make sure that others reading this realize there is no substance to it.

[entgegnen]

Hi sensij,

The discussion that went over my head was this:

"To a first approximation, when A is close to B, A⁴ - B⁴ is proportional to 4(A-B) (Average)³ As the temp difference gets very large the heat transfer becomes proportional to the fourth power of the higher temperature. In between, it is messier."

Thanks.

[ltbighorn]

PLEASE NOTE: Im not trying to convince anyone that Enphase is a good investment, or tell anyone how they should weight this information in their decision process. I just spent some time as a casual observer and potential product purchaser, looking into Enphase's situation, and am sharing some info/thoughts that are not necessarily apparent just by looking at their stock price or their last couple quarter financial reports. I am by no means a financial advisor/expert.

Enphase certainly looks to be in a difficult position financially. SolarEdge looks like the hotness right now, and that's certainly a more comfortable and higher odds position, but if there's anything I've seen in technology it's that that's no guarantee either. I've seen the tides turn rapidly so many times. Going into my selection I figured good chance that neither SolarEdge, Enphase or my installer are around in 10 years to service warranty claims, and so I'd rather not pin my satisfaction on someone honoring a warranty 15 years from now. While it'd certainly be disappointing if Enphase got into a situation where it couldn't honor warranties in the next couple years, that particular situation feels less likely to me due to the number of alternative outcomes to straight-out chapter 7 and shutdown (acquisition, restructuring, etc.), even if their financials never improved.

That said, listening to their investor calls gives some more perspective on their position. To summarize for people who don't have the interest/time, here are some highlights:

- they've explicitly chosen the strategy of thinning gross margins and suffering some losses now in order to regain/grow top line revenue and marketshare as they work on executing their cost reduction plans, which should help improve margins again
- they have working capital arrangements on the order of $75M+ w/no significant covenants/stipulations that would prevent its use -- they feel these are sufficient to cover their anticipated cash needs without selling more equity etc.
- a big part of their low cash situation is from an excess inventory (they have $45M of inventory on hand). They've kept production up in anticipation of improving sales, but still intend to draw down the inventory level as they consider it to be too high
- their 6th gen product isn't released yet, but isn't quite what I'd call vaporware. It exists now and is undergoing QA/accelerated lifetime testing, etc for release later this year.
- they're currently on target to meet their cost reduction goals (cost of producing inverters) set for both 2016 and 2017. If they continue to hit those then that should help their margins, even in the face of further price reductions. it's in their targets to become cost competitive with low-cost central inverters.
- they saw good effect from their price reductions and found that people were still willing to pay a premium for the enphase product over competitors (obviously as long as the premium isn't too steep). competitors made moves to counter enphase's price reductions but it didn't mitigate the gains to enphase.
- they haven't had any issues with accounts receivable or credit with suppliers etc.

What do I think you can take from that? They're still in a tough situation, but they're not in a panic mode -- they could do several things to decrease losses in the short term, but they believe they have the reserves to execute on a better long-term strategy for growth/profits/competitiveness. How successful they'll be certainly remains to be seen.

Based on the field MTBFs I've seen from the M215 and M250s, I personally don't think 100% of them are going to suddenly keel over. Of course, no one knows for sure. Should Enphase's outlook continue to worsen, I can get some spares at a very low incremental cost ($100-$150 per), and I can easily switch to an alternative vendor for future expansion without significant sunk costs. If roof access wasn't easy for me, I'd likely be going with a SunnyBoy Fronius rather than SolarEdge to much more substantially mitigate the risk of roof visits. Electrolytic capacitors are by no means the only reason (or necessarily even primary reason) that electronics fail.

[HWD]

I'd recommend trying to line these up to their actual financials. They're losing a ton of share still, and their costs are actually up y-o-y (COGS per watt) versus SolarEdge which are down 20%+. i.e. they were forced into cutting their prices to not be completely forced out of the market, and they're only on track with their cost targets if they were targeting flat to increasing cost.

[ltbighorn]

These line up with their actual financials -- cost improvements unfortunately won't really start benefiting their balance sheets significantly until later this year, as sales shift towards their 5th gen products, they revise their builds to benefit from their improved cost designs, etc. SolarEdge is certainly ahead of the game here.