PowerWall technical discussion

[liuping]

If look at traditional inverter manufacturers, you'll find the most they can warranty their inverters for is about 12 years.

I'm curious about the short warranties for string inverters (even the SolarEdge inverter only has a 12 year warranty )

I'm guessing it might be because traditional string inverters have to deal with a lot more power at once than each individual micro inverters, but I honestly don't know.

I personally have not had any issues with any of our 33 Enphase micro inverters (nor have any of the people I've recommended the installer to). Even even if one does fail under warranty, they are fairly easy to replace. On half the panels they are accessible directly and it would be a 5 minute job. On the other half, only one panel would need to be removed to access them. Labor costs should should not be all that much, probably much less than the cost of replacing a string inverter (plus that labor) which is very likely to be needed during that same 25 year period.

 

[schueppert]

I think the reliability issue with inverters is not so much power as frequency. Connecting to the grid requires 60Hz AC which means a big electrolytic capacitor, and electrolytic capacitors are unfortunately prone to failure. It's true, as others have said, that DC-DC conversion is done with an intermediary AC step, but this is done at at much higher frequency which allows for small ceramic capacitors which have a much lower failure rate.

 

[Cosmacelf]

I'm curious about the short warranties for string inverters (even the SolarEdge inverter only has a 12 year warranty ).

That SolarEdge reliability paper I linked to above explains why even their inverters have a shorter expected lifespan than the concentrators.

 

[mspohr]

So, have you read both whitepapers? Skimmed them at least? Enphase's is very, very unimpressive compared to SolarEdge's.



I did ASK in my post if anyone could find any complaints about SolarEdge reliability since I couldn't find ANY. If you can show me some complaints, I'm all ears.[/COLOR]

You can ask the Google for "SolarEdge problems" and come up with lots of anecdotes but these (and the Enphase anecdotes) don't mean anything in a statistical sense.
Yes, I read both white papers and they are both marketing materials so I take them with a grain of salt. They both claim good reliability and they both have 25 year warranties (except SolarEdge 12 years on the inverter). In my installation I can easily access any of the panels and modules for a 10 minute replacement.
I really don't want this to get sidetracked into a contest of SolarEdge vs Enphase products... I am interested in any design advantages for one approach over the other. My main concern about SolarEdge is that you have more failure points. A failure of any of the convertors or the inverter will take out the entire string whereas a failure of an Enphase module only takes out one panel. They each seem to have the about same efficiency.

- - - Updated - - -

So, have you read both whitepapers? Skimmed them at least? Enphase's is very, very unimpressive compared to SolarEdge's.



I did ASK in my post if anyone could find any complaints about SolarEdge reliability since I couldn't find ANY. If you can show me some complaints, I'm all ears.[/COLOR]

You can ask the Google for "SolarEdge problems" and come up with lots of anecdotes but these (and the Enphase anecdotes) don't mean anything in a statistical sense.
Yes, I read both white papers and they are both marketing materials so I take them with a grain of salt. They both claim good reliability and they both have 25 year warranties (except SolarEdge 12 years on the inverter). In my installation I can easily access any of the panels and modules for a 10 minute replacement.
I really don't want this to get sidetracked into a contest of SolarEdge vs Enphase products... I am interested in any design advantages for one approach over the other. My main concern about SolarEdge is that you have more failure points. A failure of any of the convertors or the inverter will take out the entire string whereas a failure of an Enphase module only takes out one panel. They each seem to have the about same efficiency.

 

[Cosmacelf]

You can ask the Google for "SolarEdge problems" and come up with lots of anecdotes but these (and the Enphase anecdotes) don't mean anything in a statistical sense.
Yes, I read both white papers and they are both marketing materials so I take them with a grain of salt. They both claim good reliability and they both have 25 year warranties (except SolarEdge 12 years on the inverter). In my installation I can easily access any of the panels and modules for a 10 minute replacement.
I really don't want this to get sidetracked into a contest of SolarEdge vs Enphase products... I am interested in any design advantages for one approach over the other. My main concern about SolarEdge is that you have more failure points. A failure of any of the convertors or the inverter will take out the entire string whereas a failure of an Enphase module only takes out one panel. They each seem to have the about same efficiency.

For the heck of it, I did do that Google search, didn't find any first hand complaints. Would still love to read of any SolarEdge problems.

The two white papers are substantially different.

I believe the failure of a SolarEdge converter only takes out that panel, not the string.

As far as design advantages, I don't know how you would efficiently integrate battery backup into an Enphase install, whereas it is easy to envision how you could do it with SolarEdge.

 

[nwdiver]

As far as design advantages, I don't know how you would efficiently integrate battery backup into an Enphase install, whereas it is easy to envision how you could do it with SolarEdge.

AC coupling

 

[Ampster]

Efficient is the oprative term in Cosmacelf's question. I have an Enphase battery backed system that uses a 48v pack and between the charger and additional inverter I am looking at 20% loss. Even with that loss I still benefit from peak shaving because of the $0.25 to 0.35 spread between my peak and super off peak rates.
I puchased my solar before I decided to add batteries so my decision process didn't result in the most efficient solution. I think higher DC voltages offer better efficiencies as well.

 

[Cosmacelf]

Thank you Ampster for picking up on what I was trying to convey! Yes, the key word there was efficiently.

 

[schueppert]

Yeah, I think that AC Coupling paper is just putting lipstick on a suboptimal technical solution. Enphase has even talked about launching an AC Battery (see AC Battery - Enphase Energy) which is clearly a technology that came out of the marketing department! I'm sure Enphase could adapt their product to output DC, probably without too much difficulty, but the SolarEdge Optimizers sell at half the price so I guess they would need to engineer out some cost too.

 

[liuping]

Yeah, I think that AC Coupling paper is just putting lipstick on a suboptimal technical solution. Enphase has even talked about launching an AC Battery (seeAC Battery - Enphase Energy) which is clearly a technology that came out of the marketing department! I'm sure Enphase could adapt their product to output DC, probably without too much difficulty, but the SolarEdge Optimizers sell at half the price so I guess they would need to engineer out some cost too.

The optimizers are 1/2 the price, but you still need an inverter, so the total system price for many people is very similar.

I use Enphase micro inverters in our system, but if I did it again, I would certainly consider SolarEdge. With Powerwall battery backup is getting much closer to main stream and cost effective, which was not the case when we installed the first part of our system

The Enphase AC battery are actually an interesting idea. Each pack has a bi-directional microinverter, so installation is super easy and very scalable. (Enphase Charges Into the Energy Storage Market : Greentech Media). Unfortunately, you'd need at least 6-8 of the Enphase batteries to have the storage of the 7kw Powerwall (the Enphase Batteries are 275W AC output each), and I just don't see them being cost competitive.

 

[arg]

That's a good technical note but it doesn't say how it performs the magic of telling the different Optimizers what voltage to output. There must be some communication along the string to get everyone to coordinate their output voltage to come up with 350v for the string with different amounts of shading of different modules.
I agree, this is magic.

It seems to me that no great magic is needed (though there's the usual devil in the detail to make it stable). In particular, there's no communication needed - the right voltage just comes out in the wash.

It only seems hard because we normally think of a buck or boost converter llike this with a control loop that regulates either the voltage or the current on the output. If I've understood it correctly, the control loop on these is driven by the INPUT only - the circuit delivers constant power to its output, rather than regulating the power to achieve a particular voltage or current. So if there's no load, the output voltage of each converter will just go up until it hits the maximum the converter can achieve. However, in this case there is a load - the inverter - which we are told regulates a constant-voltage at its input. So at startup, there is no current flowing and the output voltages of the converters rise until their sum is enough to start up the inverter. Now some current flows into the inverter and, because they are wired in series, the current through each module has to be the same. Since each module is regulating to deliver constant power (different per module, but constant for any one module), the voltage at the output is in proportion to that module's power. If the sun gets brighter, all modules increase their power and so their output voltages start to go up; the voltage at the inverter input starts to rise and so the inverter draws more current to keep the input voltage constant. If one panel suddenly becomes shaded, its power goes down and so, with the current remaining constant, its output voltage goes down; the input to the inverter goes down and so the inverter draws slightly less current; all the other modules now have to slightly increase their output voltages to keep delivering the same power at the newly reduced current, and we have a new stable state with the shaded module delivering less power.

So it's actually very simple indeed in concept, although I imagine it was somewhat tricky to design such that it is stable under all conditions. But it's the good kind of tricky: a very simple circuit that has to be designed 'just so' and once designed will be cheap and reliable, rather than something that gets huge and complex.

 

[DJ Frustration]

Do we know if the 2/3.3Kw limit is per pack or for the whole system. That is, if you buy the largest possible 90KWh system (9 powerwalls tied together), would you still get 2Kw power or 18kw? Also, the 3.3Kw peak--how long can the pack supply that? Long enough to use a toaster?

Has anyone addressed this question yet? I skimmed the entire thread and couldn't find anything. For those of us with, I don't know, air conditioning this is important since 2 kw won't power much.

 

[schueppert]

"It seems to me that no great magic is needed (though there's the usual devil in the detail to make it stable). In particular, there's no communication needed - the right voltage just comes out in the wash."

It's certainly possible to do MPPT at the module level in a string without communications. I think you basically want a DC-DC converter whose impedance is proportional to the power going through it. Nevertheless, the SolarEdge solution does indeed use communications between the inverter and the power optimizers so they clearly felt the extra complexity was worth it. Some product features require communications: DC disconnect at the panel level, panel level monitoring and reporting etc.. Communications also lets them keep the input voltage to inverter constant regardless of the power being generated by the string.

 

[Cosmacelf]

Has anyone addressed this question yet? I skimmed the entire thread and couldn't find anything. For those of us with, I don't know, air conditioning this is important since 2 kw won't power much.

Doesn't it HAVE to be 18 kW? This isn't voltage or amperage, it is both, really: power (VA). So, yeah, combining Powerwalls would give you up to 18 kW continuous power.

As far as what peak means, I don't think that has been addressed anywhere. However, since the battery chemistry is similar or the same as to what's in the Model S, I would suspect the peak power tolerances are similar and probably limited by how hot the pack gets. So I would expect longer peak power if the Powerwall were located in a cold basement than if it were outside in 100 degree temperatures. And, since the Model S only allows high discharge rates for a few minutes, that's what I would expect the Powerwall to do to. So maybe one slice of bread in a toaster

Actually, I would expect the peak power headroom to be used for things like motor starts (Eg. refrigerator compressor starting).

 

[bwa]

True, but that means a higher A/C load during most months, and it's not a small object either. It doesn't help that in my house none of the available walls are anywhere near the service panel.

PowerWall is already temperature controlled, so this basically shifts the air conditioning from the PW to your A/C installation. Is PW a passive temperature control system or active (compressor)? If active, then it's more of a wash. Maybe someone already posted this so if so I'll delete. Also, would the PW do "excess" temperature control if an A/C unit kept it in some constrained state? And what is the sweet spot for that? Multizone temperature might be prudent in that case. None of this has been put on Tesla's web page that I know of. Some speculators already said PW probably has no compressor, just a heater, so that means A/C could help (in over-heat climates), but if A/C were set lower than the PW is optimal at, they would fight each other.

- - - Updated - - -

I haven't studied the Powerwall as much as I might have (it wouldn't do anything for my situation). Do we absolutely know that it uses the same cells as the car, or is this just an assumption?

If so, I guess I'm confused -- I thought a major contributor to the low production for the Model S was a shortage in the supply of batteries. Why find another product that uses up the same cells? I can understand if the Gigafactory is operational and there are batteries in abundance, but why use up the current supply even further? I've heard that home storage could be a use for used batteries, but I don't think that is happening here...

The shortage is the reason for the push for a worse shortage (increased (unmet by supply, at first) demand): to meet the supply/demand mismatch, they're building a factory, so to be conservative, they are diversifying and strengthening the demand on the new factory as well (initializing the business market for that increased demand). Therefore, it is theoretically* the shortage that caused the business case for increasing the demand in spite of the shortage. Because that makes fixing the shortage a potentially more viable business, making this grand big factory to meet said shortage.


* As part of a bigger master plan, these force choices would almost seem like minutia, details, scheduling, rather than the ebb and flow of natural markets. In fact, both forces (markets and human needs) are simultaneously happening.

 

[bwa]

For some reason the 10 kWh can't be cycled regularly and can only be used as a backup. ... Oh, and solar city will only be installing the 10 kWh variant, not the 7.

Tesla's New Battery Doesn't Work That Well With Solar - Yahoo Finance

If true, I find Solar City annoyingly unuseful for the PowerWall product line.

My perfect situation is going off grid (solar + PowerWall), with the OPTION to be on-grid if I choose. This way, I don't have to worry about generating local pollution and the non-local pollution generated is a fixed capital cost, and I don't have to worry about other people complaining about how much of their electricity I'm using.

Edit: Aha, I was likely wrong. Elon explained it: Solar City is likely doing a distributed peak shaving bed using people's roofs. I still think it's kind of more constricted of a use than the more capable more expensive 7kWh one.

 

[Ampster]

......

Edit: Aha, I was likely wrong. Elon explained it: Solar City is likely doing a distributed peak shaving bed using people's roofs. I still think it's kind of more constricted of a use than the more capable more expensive 7kWh one.

Peak shaving is a daily use so I would think the 7kWh unit would be the one used for that. Or is that what you are saying?

 

[lolachampcar]

One data point for micro inverters (ABB's)

I put 31 on my roof. Eleven were under reporting AC by 10 volts and would not sync/start.
I replace them and went through two more rounds of replacing a few micros at a time until the whole system stabilized. It has been up and running without fault now for four months (knock on wood).

 

[dw_soothill]

Skimmed through this thread and I see lots of people talking about DC-DC for the batteries to the panels.

The financial problem with this is that where you are being paid generation payments they are all based on your AC power generated you want this number to be as high as possible so going Solar Panels - AC Inverter through AC Electric meter then back through DC inverter and into batteries whilst not being the most electrically efficient is the most financially beneficial to many with payments being made for power generated.

This also allows for connecting different sets of panels to the same system for charging. All you need is a meter that monitors to see if you are exporting electricity in which case you increase the rate of charging of the batteries when you start importing electricity you get the inverter that is fed by the batteries to ramp up its output until you are either not exporting or the inverter cannot produce any more power.

SMA already have a system for doing exactly that which works with existing installed solar panels without the need to change or replace the existing inverters all it needs to be aware of is if you are importing or exporting electric to decide if it should charge or draw power from the batteries. No need to disconnect from the grid to do that.


Supplying power when the grid loses power is a lot more of an issue as you have to disconnect the house from the supply grid to ensure that you can't accidentally send power back down the grid and cause an issue for someone working on the lines.

 

[GSP]

In the Q1 earnings call JB and Elon said (if I remember correctly):

The 10 kWh "backup" power wall uses NCA chemistry for 1000-1500 cycle life, same as the cars.

The 7 kWh "daily cycle" power wall uses NCM chemistry to get 5000+ cycle life

GSP