Powerwall 2.0 Backup Runtime Extender

[miimura]

Just curious if you've tested the scenario where the small runtime extended inverter (even if really small) provides more watts than the house load? Does the small inverter just ramp down in that scenario because it's aware of the house load?

When the Powerwalls are in Backup mode, they will balance the micro-grid by charging the batteries inside the Powerwall as necessary. If you are going to use an inverter larger than your typical house load, you should probably run it through the Powerwall system solar CTs so that it properly registers it as generation. That was my original plan, but it's not necessary with a small inverter. This scheme is intended to prevent the Powerwalls from running out of energy, so the high SOC frequency shifting should never come into the picture.

The inverter I chose to use is just blindly pushing battery energy into the AC outlet it's connected to. If you connected a solar panel, it would modulate more as the voltage varied.

I suspect the voltage would increase beyond the Powerwall's limit, and the Powerwall would turn itself off (turning the grid-tie inverter with it).

In the specific test that I presented in the first post, I am only feeding power in at 120VAC. That means that the Powerwall is likely absorbing energy on one side of the neutral and pushing out energy on the other side. This is one of the features of the Powerwall inverter - it can handle 100% phase imbalance on split-phase. Whether they are doing it through a transformer or natively in the inverters doesn't really matter to us.

So, "turning off" never really applies because it's always doing something as the micro-grid master in Backup mode.

It doesn't make any sense to use this Runtime Extender technique when the grid is up because of the compound losses incurred. This is intended to keep the Powerwall system running during an extended outage when you don't have enough solar generation to make it through the night.

 

[JayClark]

When the Powerwalls are in Backup mode, they will balance the micro-grid by charging the batteries inside the Powerwall as necessary. If you are going to use an inverter larger than your typical house load, you should probably run it through the Powerwall system solar CTs so that it properly registers it as generation. That was my original plan, but it's not necessary with a small inverter. This scheme is intended to prevent the Powerwalls from running out of energy, so the high SOC frequency shifting should never come into the picture.

The inverter I chose to use is just blindly pushing battery energy into the AC outlet it's connected to. If you connected a solar panel, it would modulate more as the voltage varied.


In the specific test that I presented in the first post, I am only feeding power in at 120VAC. That means that the Powerwall is likely absorbing energy on one side of the neutral and pushing out energy on the other side. This is one of the features of the Powerwall inverter - it can handle 100% phase imbalance on split-phase. Whether they are doing it through a transformer or natively in the inverters doesn't really matter to us.

So, "turning off" never really applies because it's always doing something as the micro-grid master in Backup mode.

It doesn't make any sense to use this Runtime Extender technique when the grid is up because of the compound losses incurred. This is intended to keep the Powerwall system running during an extended outage when you don't have enough solar generation to make it through the night.

Makes sense. This is all very good information.

 

[gpez]

When the Powerwalls are in Backup mode, they will balance the micro-grid by charging the batteries inside the Powerwall as necessary. If you are going to use an inverter larger than your typical house load, you should probably run it through the Powerwall system solar CTs so that it properly registers it as generation. That was my original plan, but it's not necessary with a small inverter. This scheme is intended to prevent the Powerwalls from running out of energy, so the high SOC frequency shifting should never come into the picture.

The inverter I chose to use is just blindly pushing battery energy into the AC outlet it's connected to. If you connected a solar panel, it would modulate more as the voltage varied.


In the specific test that I presented in the first post, I am only feeding power in at 120VAC. That means that the Powerwall is likely absorbing energy on one side of the neutral and pushing out energy on the other side. This is one of the features of the Powerwall inverter - it can handle 100% phase imbalance on split-phase. Whether they are doing it through a transformer or natively in the inverters doesn't really matter to us.

So, "turning off" never really applies because it's always doing something as the micro-grid master in Backup mode.

It doesn't make any sense to use this Runtime Extender technique when the grid is up because of the compound losses incurred. This is intended to keep the Powerwall system running during an extended outage when you don't have enough solar generation to make it through the night.

So furthering this scenario: what happens if your extender is plugged in when the grid comes back online?

 

[miimura]

So furthering this scenario: what happens if your extender is plugged in when the grid comes back online?

I refer you to my prior statement:

The inverter I chose to use is just blindly pushing battery energy into the AC outlet it's connected to.

Since the Powerwalls will re-synchronize the AC waveform to the grid before closing the Gateway switch, the small battery inverter won't know the difference when the grid is back online.

I should make a video some day with an oscilloscope showing the off-grid waveform in Backup mode and the grid waveform and what happens when you put the Powerwalls back on the grid.

 

[mckemie]

Where do you get the M215s?

eBay. Prices are best when you buy a box of 12.
I envision about 4-5 M215s on a single EV to give a boost of around 1kw.

 

[mckemie]

I suspect the voltage would increase beyond the Powerwall's limit, and the Powerwall would turn itself off (turning the grid-tie inverter with it).

For M125s, the inverters turn themselves off when voltage reaches around 260.
For off grid, where the PW is supplying the grid signal, the PW will turn M215s off by frequency shifting when there is more power than load.

 

[mckemie]

Looking ahead to the time when this scheme, with M215s, might be proven to work:
I don't hope to automate operation to the PowerWall level but I will be looking to switching generation and car charging on and off based on time of day. Probably charge car from 10am until 4pm and generate from 7pm unit 8am. Can anyone suggest timer switches that will do 3+kw 240vac? For generation, it seems nicer to switch the 12vdc but I imagine it will suffice to switch the output of the M215s.

A very handy item would be a switch that would be controlled by ac voltage. One use would be to switch on loads as the micro grid voltage rises. It would need to be ajustable with the expected range 250-260. Anyone seen anything like that?

 

[miimura]

Looking ahead to the time when this scheme, with M215s, might be proven to work:
I don't hope to automate operation to the PowerWall level but I will be looking to switching generation and car charging on and off based on time of day. Probably charge car from 10am until 4pm and generate from 7pm unit 8am. Can anyone suggest timer switches that will do 3+kw 240vac? For generation, it seems nicer to switch the 12vdc but I imagine it will suffice to switch the output of the M215s.

A very handy item would be a switch that would be controlled by ac voltage. One use would be to switch on loads as the micro grid voltage rises. It would need to be ajustable with the expected range 250-260. Anyone seen anything like that?

The micro-grid voltage should not be changing significantly. You should be controlling things based on other parameters like SOC which is available by API.

You can use an industrial contactor to switch the 240VAC circuit. They are available with AC and DC coils for actuation.
This one for example. 12VDC coil, 3 power contacts up to 16A. ~$20
IEC Miniature Contactor: 16A, 12 VDC coil voltage, low consumption (PN# CWC016-10-30L02) | AutomationDirect

The other way to go would be to use a 120VAC coil and use an off the shelf smart plug to do any sort of cloud connected actuation you want. This one is 120VAC coil, 3 power contacts up to 18A. ~$26
IEC Contactor: 18A, 120 VAC (60Hz)/110 VAC (50Hz) coil voltage (PN# SC-E04-110VAC) | AutomationDirect

 

[gpez]

I refer you to my prior statement:

The inverter I chose to use is just blindly pushing battery energy into the AC outlet it's connected to.

Since the Powerwalls will re-synchronize the AC waveform to the grid before closing the Gateway switch, the small battery inverter won't know the difference when the grid is back online.

I should make a video some day with an oscilloscope showing the off-grid waveform in Backup mode and the grid waveform and what happens when you put the Powerwalls back on the grid.

Sorry, my question was about how power flow will work in that scenario. If the inverter is blindly pushing battery energy in to the AC outlet and the grid comes back up and solar is being generated is there a risk to any of the equipment at that point? Can the Powerwall support production from the home (ie negative consumption)?

Totally agree that lining the extender on the PV side of the Powerwall is the safest but very curious what happens in different scenarios where an extender is just dumping power on the load side.

 

[power.saver]

Yes, the PW can support negative consumption. I have tried this. Although the screen shows 0kW for the house, it was actually negative.

Also the inverter is no different than your solar inverter, it pushes current into the grid (or PW). As long as the amount it is pushing is less than what the PW can accept, it will work (during a grid outage).

 

[miimura]

Sorry, my question was about how power flow will work in that scenario. If the inverter is blindly pushing battery energy in to the AC outlet and the grid comes back up and solar is being generated is there a risk to any of the equipment at that point? Can the Powerwall support production from the home (ie negative consumption)?

Totally agree that lining the extender on the PV side of the Powerwall is the safest but very curious what happens in different scenarios where an extender is just dumping power on the load side.

Everything in this scheme is designed for grid-tied operation, so everything is synchronized, whether the grid is up or down. In Backup Mode (grid disconnected), the Powerwall is doing everything it can to maintain the voltage and frequency and will produce or absorb AC power up to its limits to do that. Solar inverters are only distinguishable from this kind of added inverter because they are measured by the Powerwall system CT clamps. They are connected through the electrical panel in the same way as any other load that is on the backup side of the Gateway switch.

Just don't try something like this with a normal mobile battery inverter that freewheels at whatever frequency it wants.

 

[Ulmo]

Yes, the PW can support negative consumption. I have tried this. Although the screen shows 0kW for the house, it was actually negative.

Also the inverter is no different than your solar inverter, it pushes current into the grid (or PW). As long as the amount it is pushing is less than what the PW can accept, it will work (during a grid outage).

Yes. That implies that a tiny bit of logic to stop inverting into the batteries when they are full would be useful, such as a sneaker poll system (you manually check state of charger of PowerWalls, and when it's high, go disconnect the backup inverter injection). Even still, @miimura said the main problem is a restart at low state of charge, but a restart at high state of charge doesn't sound like a peach either.

 

[mckemie]

Yes. That implies that a tiny bit of logic to stop inverting into the batteries when they are full would be useful, such as a sneaker poll system (you manually check state of charger of PowerWalls, and when it's high, go disconnect the backup inverter injection). Even still, @miimura said the main problem is a restart at low state of charge, but a restart at high state of charge doesn't sound like a peach either.

For off grid operation, I believe the PW will attempt to turn off PV generation by shifting frequency when it finds more power supply than demand. The capactiy to charge the PW battery is part of the demand; you typically run into oversupply problems only after the battery is fully charged. I imagine all "grid tie" inverters mentioned here will respond to the frequency shift by ceasing operation until they again see a "good" grid signal.

 

[mckemie]

The micro-grid voltage should not be changing significantly. You should be controlling things based on other parameters like SOC which is available by API.

You can use an industrial contactor to switch the 240VAC circuit. They are available with AC and DC coils for actuation.
This one for example. 12VDC coil, 3 power contacts up to 16A. ~$20
IEC Miniature Contactor: 16A, 12 VDC coil voltage, low consumption (PN# CWC016-10-30L02) | AutomationDirect

The other way to go would be to use a 120VAC coil and use an off the shelf smart plug to do any sort of cloud connected actuation you want. This one is 120VAC coil, 3 power contacts up to 18A. ~$26
IEC Contactor: 18A, 120 VAC (60Hz)/110 VAC (50Hz) coil voltage (PN# SC-E04-110VAC) | AutomationDirect

Can you explain the statement about "API"?

I have a stiuation where my peak PV production may not be accepted by my utility supplied tranformer. At such times, my voltage does rise to the point where PV inverters are shut down. That, typically, is around 260vac. For instance, my 15kva transformer never takes more than about 15kw. I believe I could make good use of a device that turns on non critical demand, such as vehicle chargin, above about 255vac.

I intend to use mechanical water heater timers to control both the generation from my EV attached M215s as well as the charging of the EV if I wish to operate several days. I intend to turn on generation from about 8pm until about 8am and the vehicle charging from about 10 am until about 4pm.

 

[Odiemac]

Can you explain the statement about "API"?

I have a stiuation where my peak PV production may not be accepted by my utility supplied tranformer. At such times, my voltage does rise to the point where PV inverters are shut down. That, typically, is around 260vac. For instance, my 15kva transformer never takes more than about 15kw. I believe I could make good use of a device that turns on non critical demand, such as vehicle chargin, above about 255vac.

I intend to use mechanical water heater timers to control both the generation from my EV attached M215s as well as the charging of the EV if I wish to operate several days. I intend to turn on generation from about 8pm until about 8am and the vehicle charging from about 10 am until about 4pm.

Someone on GitHub is working on documenting the PowerWall API. You can see their progress here: vloschiavo/powerwall2

And here is the part on powerwall charge status:


State of Charge / State of Energy GET /api/system_status/soe

This returns the aggregate charge state in percent of the powerwall(s).

request: curl https://192.168.xxx.xxx/api/system_status/soe

response: {"percentage":69.1675560298826}

When site master or the Powerwalls are off, the response is: HTTP Status 502

 

[mckemie]

Someone on GitHub is working on documenting the PowerWall API. You can see their progress here: vloschiavo/powerwall2

And here is the part on powerwall charge status:


State of Charge / State of Energy GET /api/system_status/soe

This returns the aggregate charge state in percent of the powerwall(s).

request: curl https://192.168.xxx.xxx/api/system_status/soe

response: {"percentage":69.1675560298826}

When site master or the Powerwalls are off, the response is: HTTP Status 502

Neat! Super neat!

Worked for me though I monitor my PW's web page of the time. I have it "cast" to an otherwise unused TV.
I've go read the GitHub reference.

 

[pgrovetom1]

I've been reading through these posts and was trying to understand the principle behind what is being described? Is the purpose to add additional AC power injected into one AC leg from an additional battery via an inverter while the PW2 system is in backup to extend its time to shutdown?

is it plugged in and turned on manually while the PW2 is providing backup power? Is the idea that another battery has been charged and a small grid-tie inverter such as an Enphase M215 ( I bought one for $39 on Ebay) sees the PW2 60Hz power and locks on and injects an additional 25W or so onto the home grid. Then when the PW2 runs out of battery and shuts down, the Enphase sees the loss of 60Hz on its output and shuts down as a grid-tie inverter.

Is the hope that the additional 250W will supplement the PW2 by the 250W which will extend its run-time by whatever that 250W will allow. Is the idea that this small battery/inverter system will "look" like a small solar system that is providing 250W?

The Enphase or any inverter intended for solar is used, it has an MPPT tracker on its input which would cause over current problems for a direct battery connection. One can buy a 36V at 20-30AH lithium battery and charger on Amazon or Ebay for about $350 which is normally intended to be sued with a E-Scooter. If a current limiter is put on its output at about 6A, then that limiter makes the battery output "appear" as similar to a solar panel. When the MPPT circuit tries to increase the voltage at which it draws power form the battery, a maximum power point will be found at the current limit knee of 36V and 6A = 36x6=216W not unlike a solar panel at its IV curve knee. This is also within the the M215 power rating and it will just produce the 215W until the 36V battery does a low voltage discharge disconnect or the PW2 shuts down and the inverter does a UL1741 shutoff. The 36V lithium battery operates between about 30V and 38V. This is within the input range of the Enphase M215 M215-60-2LL-S22.

Does any of this make sense?

 

[Dan123]

I've been reading through these posts and was trying to understand the principle behind what is being described? Is the purpose to add additional AC power injected into one AC leg from an additional battery via an inverter while the PW2 system is in backup to extend its time to shutdown?

is it plugged in and turned on manually while the PW2 is providing backup power? Is the idea that another battery has been charged and a small grid-tie inverter such as an Enphase M215 ( I bought one for $39 on Ebay) sees the PW2 60Hz power and locks on and injects an additional 25W or so onto the home grid. Then when the PW2 runs out of battery and shuts down, the Enphase sees the loss of 60Hz on its output and shuts down as a grid-tie inverter.

Is the hope that the additional 250W will supplement the PW2 by the 250W which will extend its run-time by whatever that 250W will allow. Is the idea that this small battery/inverter system will "look" like a small solar system that is providing 250W?

The Enphase or any inverter intended for solar is used, it has an MPPT tracker on its input which would cause over current problems for a direct battery connection. One can buy a 36V at 20-30AH lithium battery and charger on Amazon or Ebay for about $350 which is normally intended to be sued with a E-Scooter. If a current limiter is put on its output at about 6A, then that limiter makes the battery output "appear" as similar to a solar panel. When the MPPT circuit tries to increase the voltage at which it draws power form the battery, a maximum power point will be found at the current limit knee of 36V and 6A = 36x6=216W not unlike a solar panel at its IV curve knee. This is also within the the M215 power rating and it will just produce the 215W until the 36V battery does a low voltage discharge disconnect or the PW2 shuts down and the inverter does a UL1741 shutoff. The 36V lithium battery operates between about 30V and 38V. This is within the input range of the Enphase M215 M215-60-2LL-S22.

Does any of this make sense?

Yes, the idea is to inject 5KWH per day. You an add a second M215 inverter and inject a total of 10KWH per day. That will increase the runtime of the powerwall.

Why would a M215 cause over current problem for a direct battery connection? The inverter is limited to 215W output. So I don't think it can draw more than 10 amps from the battery.

 

[gpez]

So my proposed setup doesn't have a 240v outlet available to use something like an M215 from an EV.

@Dan123 would something like this (600W 600 Watt Grid Tie Inverter MPPT Accept 10.8v-30 V DC/120v) grid tie inverter work without a step up 12v->24v? Thinking EV 12v battery -> fuse -> inverter -> 120v AC outlet.

 

[miimura]

So my proposed setup doesn't have a 240v outlet available to use something like an M215 from an EV.

@Dan123 would something like this (600W 600 Watt Grid Tie Inverter MPPT Accept 10.8v-30 V DC/120v) grid tie inverter work without a step up 12v->24v? Thinking EV 12v battery -> fuse -> inverter -> 120v AC outlet.

The grid tie inverter you linked is basically the same unit that I used in my test shown in the first post of this thread. The only problem is that is has a poor thermal design and it is not able to maintain its rated output continuously. It runs steady state at about 480W. I did not try it without the 24V step up. It may just have lower output at the lower voltage.

@pgrovetom1
The purpose of this scheme is to keep your house running overnight until your solar starts generating again. It makes sense to use the large battery in an EV for this kind of buffer. It is conceivable that if you had a large solar system relative to your Powerwall system, that you could charge the EV in the afternoon after the Powerwall is full and use this extender overnight to allow you to use more lights or appliances. It doesn't make sense to me to use a small battery pack for this purpose when many of us have large battery packs that we drive around each day.