Powerwall 2 Vampire Drain

[avirtuos]

I shouldn't be surprised but here I am, surprised. I've had my 3 powerwall 2 units up and running for 5 days. I finally got around to writing some scripts to query the backup gateway APIs and log the state of charge, load, solar, and grid power figures periodically and noticed a couple things that I'm not thrilled about and curious if others have seen.

1. My powerwalls are using ~ 7 watts of power continuously from the grid when in backup mode. This is no biggie until you see #2
2. My powerwalls are losing about 0.1% of charge every hour or 336Wh per powerwall every 24hrs. So my 3 powerwalls consume 1kWh per day sitting around doing nada.

Wtf are they doing with 1kWh/day? The powerwalls are in a temperature controlled area that is a steady 64 F. Can't get more ideal than that at idle. When you combine the 10% round trip efficiency loss on the inverter plus this, you are loosing a nice chunk of power.

I'm going to try shutting off the powerwalls and see what their state of charge is after a full day. I can't imagine their electronics are actually using that much power. That more power than running a mackbook all day watching YouTube (25watt/hr)

Anyone else tracking their idle drain?

 

[miimura]

Yes, this has been observed by many people. I don't have the figures since I don't use my Powerwalls this way, but your numbers seem like the ballpark I recall hearing. It also becomes a problem when you have a power outage when your panels are covered in snow. After the Powerwalls shut down, they keep losing energy until the grid comes back.

 

[avirtuos]

Good to know it's the norm but damn that sucks. Really can't depend on these for 'long term' storage then. At best it's days or a couple weeks before those losses really cut into efficiency.

As an aside, we failed our electrical inspection today. Yay Tesla for violating basic building codes.

 

[CWNE88]

I worked out my average over a year to be about 1kWh per day of loss. That was the average of energy in to energy out over the year. It didn't seem too bad to me.
There will always be some trickle to/from the grid, as if the voltages aren't exactly the same (which they never will be) then current will flow to/from. Where I am, it works out to be about 250Wh per day.

 

[gpez]

I've noticed the same. About 300wh/day charge as soon as the sun comes up from the PV.

As you noted these devices are effectively small computers that monitor and control everything, plus the fans and pumps to control the liquid cooling system. 20w draw per Powerwall (7w + 300wh/day) is not bad at all and is about half what a laptop consumes during normal operation.

 

[avirtuos]

A 2018 MacBook pro uses about 112wh when watching YouTube videos at full brightness for 24hrs. So a powerwall seems to be nearly 3x that in vampire Drain when in a perfectly climate controlled space. Tesla can do way better unless some of this loss is just chemistry. I really wish Tesla published more info about this stuff.

 

[miimura]

In the scope of operating an American single family home, the vampire drain of a Powerwall is not financially significant. Think about it in terms of Wh/day times $/kWh and you come up with a pretty small monthly figure. Just my opinion.

 

[NinjaVece]

A 2018 MacBook pro uses about 112wh when watching YouTube videos at full brightness for 24hrs. So a powerwall seems to be nearly 3x that in vampire Drain when in a perfectly climate controlled space. Tesla can do way better unless some of this loss is just chemistry. I really wish Tesla published more info about this stuff.

avirtuous, since the PW2 must run pumps and fans daily, it's not quite the same comparison as a Macbook running Youtube on its screen. The first requires physical movement of parts while the second is purely solid-state.

 

[avirtuos]

avirtuous, since the PW2 must run pumps and fans daily, it's not quite the same comparison as a Macbook running Youtube on its screen. The first requires physical movement of parts while the second is purely solid-state.

Sorry but this statement makes false assumptions about sold state vs moving parts it also assumes that the powerwalls, at idle, run their cooling loop (they do not, I've tested with a decibal meter you can easily tell when the pumps are running with any urgency).

As for solid state vs moving parts... Again false logic. Take for example a resistive heater... It's cable of consuming many kwha of power or as an example a solid state graphics card... Some are capable of 1kw draw all on their own. Conversely and peristolic pump or a ceiling fan can both run on as little as 30watts.

After reading, I suspect at least some of the drain is the BMS system constantly monitor cell balance and temperature as well as running the inverter for being prepared for load shifts at a moment's notice. More I research it the more I'm convinced it's largely the inverter.

 

[NinjaVece]

Sorry but this statement makes false assumptions about sold state vs moving parts it also assumes that the powerwalls, at idle, run their cooling loop (they do not, I've tested with a decibal meter you can easily tell when the pumps are running with any urgency).

As for solid state vs moving parts... Again false logic. Take for example a resistive heater... It's cable of consuming many kwha of power or as an example a solid state graphics card... Some are capable of 1kw draw all on their own. Conversely and peristolic pump or a ceiling fan can both run on as little as 30watts.

After reading, I suspect at least some of the drain is the BMS system constantly monitor cell balance and temperature as well as running the inverter for being prepared for load shifts at a moment's notice. More I research it the more I'm convinced it's largely the inverter.

Actually its spot on. Powerwalls do all sorts of monitoring and adjusting through all hours of the day. Unless you spent 24 hours one day with your decibel meter next to the battery, I'll take your statement with a grain of salt.

Next, again, solid-state versus moving parts IS lower usage. Take your resistive heater example. Use of a resistive heater is MUCH lower electricity usage versus heat created by a moving part. And you're moving a bit far field here. Neither a Macbook nor a PW2 are looking to create heat either via resistance or other method. However, running fans and pumps - basically any moving part - is more consumptive of electricity than a solid-state alternative.

Again, it is the comparison you are making that is odd. Comparing a Macbook that is displaying graphics based on solid-state electronics to a battery management system that is solid-state AND has moving parts doesn't make any sense. Any comparison of the PW2 BMS would need to ALSO include moving parts.

Now, your final assumption is closer to the mark. Running an inverter for internal purposes yields 90-95% efficiency. There will be some losses. And inverter use leads to fan use and, at times, pump use...

 

[avirtuos]

I powerwalls drain doest vary it's an even slope so either it's always running it's cooling (moving parts) ors it's never running them. I'm pretty convinced it's the inverter (a solid state part btw )

 

[JohnRatsey]

I treat the PW's internal consumption as part of the battery inefficiency. The Tesla app tells me 2125 kWh in and 1803 kWh out over the past year which is about 85% efficiency. Those numbers show that the battery has a relatively easy life so the internal consumption becomes a bigger proportion of the power in than if the battery was subjected to daily charge / discharge nearer to its full capacity.

Assuming Tesla's 90% conversion efficiency then the 2125 kWh in would provide 1913 kWh out so the difference (1913 - 1803) is the battery power consumption and this averages at 0.3 kWh/day - similar to the numbers in the starting post in this discussion. The battery is in the garage adjoining the house and stays cool in the summer but might want to heat itself during the winter. Nonetheless I don't recall Tesla's literature saying anything about this self power consumption which has a modest impact on the cost-benefit calculations.

 

[holstein13]

I'm finding the same issue on my two Powerwall IIs. I'm really glad to see this is normal. Although it breaks my heart to lose this potential energy, I'm slightly comforted by the fact that I've installed the batteries in my utility room right next to a heat pump water heater so I can potentially retrieve part of the energy lost into heat for my water.

 

[Dan123]

I shouldn't be surprised but here I am, surprised. I've had my 3 powerwall 2 units up and running for 5 days. I finally got around to writing some scripts to query the backup gateway APIs and log the state of charge, load, solar, and grid power figures periodically and noticed a couple things that I'm not thrilled about and curious if others have seen.

1. My powerwalls are using ~ 7 watts of power continuously from the grid when in backup mode. This is no biggie until you see #2
2. My powerwalls are losing about 0.1% of charge every hour or 336Wh per powerwall every 24hrs. So my 3 powerwalls consume 1kWh per day sitting around doing nada.

Wtf are they doing with 1kWh/day? The powerwalls are in a temperature controlled area that is a steady 64 F. Can't get more ideal than that at idle. When you combine the 10% round trip efficiency loss on the inverter plus this, you are loosing a nice chunk of power.

I'm going to try shutting off the powerwalls and see what their state of charge is after a full day. I can't imagine their electronics are actually using that much power. That more power than running a mackbook all day watching YouTube (25watt/hr)

Anyone else tracking their idle drain?

I am curious how you measured 7 watts of power from the grid? Is it the powerwalls, or the gateway?

I can see 600 watts used daily by 2 PWs in the Tesla app. There is a power loss of about 2% per day, and then they get recharched from the sun daily.

 

[NuShrike]

I'm not tracking the idle drain, but I'm tracking the discrepancy between my utility's daily usage numbers (double-correlated) vs Tesla's, and I see an average delta of 696 watts over the last 7 days.

So that's ~700 watts of phantom drain over 2 PWS over 24 hours not accounted by Tesla's numbers, but is by my utility (and RainForest Eagle-200).

About 14.5 watts per PW per hour in daily, normal, self-consumption/TBC cycling.

 

[miimura]

I'm not tracking the idle drain, but I'm tracking the discrepancy between my utility's daily usage numbers (double-correlated) vs Tesla's, and I see an average delta of 696 watts over the last 7 days.

So that's ~700 watts of phantom drain over 2 PWS over 24 hours not accounted by Tesla's numbers, but is by my utility (and RainForest Eagle-200).

About 14.5 watts per PW per hour in daily, normal, self-consumption/TBC cycling.

There's a difference between measurement error and totals accumulation versus vampire drain. Also, be careful about the use of Watt-hour units and Watts.

 

[bemyax]

New user here. A new set of 8 panels up on the second story here in northeast Iowa. One Powerwall in the basement initialized in December. The basement temperature is set to range from 57 to 59 degrees. The panels are in pairs, each pair with its own microinverter. Two pair are completely covered in frozen snow. One pair has half a panel showing. And one pair has a whole panel showing. Here's the topper: we are in day 8 of completely overcast weather; it's the dominant topic on the TV weather.

Result: At the beginning of this gloomy stretch the battery level was at 56%. At the moment, it's at 37%. (My reserve level is 65%) According to the Tesla app, my solar production on each of the days - 0.5, 0.1, 0.1, 0.2, 0.2, 0.1, 0.1, 0.1 kWh. Also according to the app 0.1 and 0.1 kWh has gone to the Powerwall in those 8 days.

Someone who is good with numbers could do the figuring. My ballpark estimate is that a little more than 300 Watt-hours is evaporating into the ether each day.

 

[charlesj]

New user here. A new set of 8 panels up on the second story here in northeast Iowa. One Powerwall in the basement initialized in December. The basement temperature is set to range from 57 to 59 degrees. The panels are in pairs, each pair with its own microinverter. Two pair are completely covered in frozen snow. One pair has half a panel showing. And one pair has a whole panel showing. Here's the topper: we are in day 8 of completely overcast weather; it's the dominant topic on the TV weather.

Result: At the beginning of this gloomy stretch the battery level was at 56%. At the moment, it's at 37%. (My reserve level is 65%) According to the Tesla app, my solar production on each of the days - 0.5, 0.1, 0.1, 0.2, 0.2, 0.1, 0.1, 0.1 kWh. Also according to the app 0.1 and 0.1 kWh has gone to the Powerwall in those 8 days.

Someone who is good with numbers could do the figuring. My ballpark estimate is that a little more than 300 Watt-hours is evaporating into the ether each day.

Yep. There are mentions on this forum that there is about that used by the battery for maintenance, internal heaters, etc to keep it operating.

 

[gpez]

Yep. There are mentions on this forum that there is about that used by the battery for maintenance, internal heaters, etc to keep it operating.

Which is wild to me that here in the US for people looking to take advantage of the ITC that Tesla can't configure the system to simply draw that 300wh/day from the grid instead of draining the battery. Seems like that'd be the solution to stay within the ITC parameters and not do this slow drain, lowering the utility of the Powerwall as a solar attached backup solution in low winter production climates.

 

[JayClark]

New user here. A new set of 8 panels up on the second story here in northeast Iowa. One Powerwall in the basement initialized in December. The basement temperature is set to range from 57 to 59 degrees. The panels are in pairs, each pair with its own microinverter. Two pair are completely covered in frozen snow. One pair has half a panel showing. And one pair has a whole panel showing. Here's the topper: we are in day 8 of completely overcast weather; it's the dominant topic on the TV weather.

Result: At the beginning of this gloomy stretch the battery level was at 56%. At the moment, it's at 37%. (My reserve level is 65%) According to the Tesla app, my solar production on each of the days - 0.5, 0.1, 0.1, 0.2, 0.2, 0.1, 0.1, 0.1 kWh. Also according to the app 0.1 and 0.1 kWh has gone to the Powerwall in those 8 days.

Someone who is good with numbers could do the figuring. My ballpark estimate is that a little more than 300 Watt-hours is evaporating into the ether each day.

Yes, my calculations are close to yours. I have 4 Powerwalls, and have run some tests with one, and all four... it seems each Powerwall uses about on average 20watts per hour, and like once every day or so they take a big "gulp" of juice to jump back to my reserve when I'm not otherwise pulling juice from the PWs (like on the weekends) - oddly this big gulp to replenish seems to even occur by pulling from the grid from what I can tell when it happens in the middle of the night. Anyway, my loss calculation came out around 450wHrs per day, per PW.... so for me it's about 2000kWhrs per day in just "maintenance' loses when not cycling the Powerwalls otherwise (charging/draining). Maybe four power walls have more losses staying balanced and synced with each other. Not sure.

I know how batteries and inverters work, and the losses involved, so I kind of expected this, but oddly, after some testing, I feel like my utility must be erring (or rounding my hourly usage, and solar offsets) in their own favor when I'm drawing from the grid, but otherwise letting all extra solar back-feed to the grid as an offset, the utility gives me a theoretical 1:1 offset for every kWh's I send to them (i.e. with the Powerwalls off).

But it doesn't seem to quite calculate out that way on my bills. Because, with the PWs on, I seemed to more than make up for the PW losses (as compared to my PW-off testing) by somehow "magicially" using less overall power for the house than on nearly identical days when I turn the PWs off and power the house from the solar & grid only, and send all extra solar to the grid. In theory, with the calculated PW losses, on identical solar & house usage days, I should use more when the Powerwalls are active but also incurring the expected losses.

After multiple tests, on identical (or nearly so) solar days and house usage days, I should see the PWs cause the overall house usage to be an extra 2kWhrs or so because of the PW losses. But what actually happens, is I use about 2-4kwhrs less when I avoid sending solar to the grid or puling house power from the grid - utilizing the PWs as much as I can. It doesn't make sense unless the utility is really skewing my house-grid usage, and my solar production sent to the grid in their favor. And the less my house uses, the less their is to skew. So this first became noticeable when I tested here in the winter when my usage is low and I can almost zero out my usage with Solar and the PW. The first time I tested this, I thought maybe the Tesla App was skewing the numbers to make itself look better, but the Tesla App and Utility numbers are agreeing with each other as much as they ever do in both scenarios, and I have a separate whole house energy Monitor on the mains leaving the house, and it's numbers sync well with the App and the Utility numbers.

Now I'm trying to figure out if I'm doing something dumb in my testing, and missing something obvious that would cause these results - but not coming up with anything I've done wrong in my testing at this point.