PowerWall Cold Start without Grid Power

[MorrisonHiker]

I think the first step should be to verify that the system works properly when the grid is down and the sun is shining. Then add the additional requirement of cold start, like if they shut down due to low SOC overnight.

That's the test that worked successfully when he came out yesterday. Since the panels were only producing at 50%, he's going to come back out tomorrow to try it when producing closer to 100% and ensure it works then too. I plan on testing the cold start again myself...before the next blizzard.

 

[reddy]

I don't see anyone mentioning that you can enable a cold start by applying a 12v signal to the 'JUMP' pins just below the AUXILIARY CONTROL pins.

 

[miimura]

I don't see anyone mentioning that you can enable a cold start by applying a 12v signal to the 'JUMP' pins just below the AUXILIARY CONTROL pins.

Do you have documentation to that effect? Which pin is positive?

 

[reddy]

Do you have documentation to that effect? Which pin is positive?

If you enlarge the photo, you can see the + and - on the diagram. The + is on the left.

 

[fresnoboy]

If you enlarge the photo, you can see the + and - on the diagram. The + is on the left.

This is interesting - what manual told you about this? It seems it could be a very useful capability, assuming it was actually implemented.

 

[reddy]

This is interesting - what manual told you about this? It seems it could be a very useful capability, assuming it was actually implemented.

That manual is only available to installers. I've looked and I don't see it online. You might try calling your installer to confirm, or , call Tesla tech support.

That being said, there ARE completely off grid Powerwalls running out there. And this is how they do it.

 

[Frankman60]

I want to run this test on my newly installed PWs. I understand that the PWs may raise the frequency (Hz) to force the inverters to shut down. But what variables need to exist to make this happen (e.g PW charge level, solar energy output from inverters in kWs)? If the inverters are shut down by the PWs, at what point are they started up again to allow the PWs to charge? I have only run one simulated grid outage so far and the solar was producing less than 1 kW and the PWs were at about 50 percent charged. I do not think that the inverters were shut down during this test.

I think the first step should be to verify that the system works properly when the grid is down and the sun is shining. Then add the additional requirement of cold start, like if they shut down due to low SOC overnight.

 

[Frankman60]

I ran the above test that I mentioned with the PWs at 100 percent charged and solar production about 5 kWs. After flipping the main breaker the PWs took over and immediately shut down my Sunny Boy inverters by raising the frequency to about 62.5 Hz. The PWs were powering the house and when they got down to 97 percent charge the inverters came back on line. I wasn’t expecting the inverters to come back on line so soon. I thought that the PWs might let the charge level to drop lower than that before turning the inverters back on. I’m wondering if this always happens at 97 percent or it also has to do with how big the load is on the PWs over X period of time. In my test, my home was drawing close to 5 kWs since the beginning of the test because I turned on the oven to speed the test up. Perhaps the rate of battery energy consumption affects when the inverters are turned back on?

 

[power.saver]

They always turn back on at 97% in my testing, and will shut down again if the % charge reaches 98%.

 

[reddy]

They always turn back on at 97% in my testing, and will shut down again if the % charge reaches 98%.

I had a grid outage that lasted over 30 hours.

Both days the powerwalls reached 100% and shut down my inverters, until the powerwalls reached 96-97% capacity and then turned on again. They’re did this all afternoon both days,

 

[Frankman60]

I’m glad to hear that the Powerwalls worked as they should during your long grid outage. I don’t often hear of actual long term grid outages and how the Powerwalls handled it. I’m sure that there are a lot of cases like this outside of the United States where the electrical grid is less stable.

View attachment 415038 View attachment 415039
I had a grid outage that lasted over 30 hours.

Both days the powerwalls reached 100% and shut down my inverters, until the powerwalls reached 96-97% capacity and then turned on again. They’re did this all afternoon both days,

 

[Frankman60]

Has anyone figured out the procedure to “jump start” the Gateway and PWs if the PWs are fully depleted and the grid is down? I would like to be prepared for this because it is possible to fully drain PWs during an extended grid outage. I see that the “jump start” terminals are 12 volts, but what amount of amps are needed to do this? I would not want to damage the Gateway if I had to do this and want to have a battery on hand to do it. I was thinking of a small lithium rechargeable battery with two wires to connect to the terminals.

I don't see anyone mentioning that you can enable a cold start by applying a 12v signal to the 'JUMP' pins just below the AUXILIARY CONTROL pins.
View attachment 390552

 

[gpez]

100% pure speculation but it is likely a physical normally open relay (switch) meaning only a few dozen milliamps would be needed to flip it. I would believe it to be unlikely that the power supplied to those contacts is used to energize any other component of the system and would either draw from whatever reserve is available in the Powerwall itself or external source (ie solar in a grid outage).

 

[woferry]

If the PWs are fully-depleted and the grid is down, simply connecting power to the jump header won't do anything AFAIK, as the solar inverters still won't start without seeing a 120V/60Hz signal, and where is that going to come from? The PWs would have to still have some energy in them to get their inverters up and running (for at least 5 minutes it seems) or the grid would need to be present (or you'd have to also have a generator, etc) before the solar inverters would see something they'd be happy synchronizing with to start supplying energy.

But the nice thing about DC voltage sources is there's no such thing as 'too much power' (provided you're not shorting something out). You could put 1000 12V batteries in parallel and connect them to this port, and it won't draw any more current than if you used a single battery. So not providing enough is an issue (wouldn't work), but providing too much is not. That style connector (and the gauge of wire you can connect to it) can't realistically deliver more than a few amps, so the draw should be pretty small, but I guess could be an amp or two perhaps. I'd suspect this is an alternate way of powering the electronics in the Gateway, which shouldn't draw more than a few watts since this is a constant load on your house, so hopefully Tesla didn't waste a bunch of power here. But unless you also have a generator tied-in it doesn't seem like it would really make any difference to me.

 

[gpez]

If the PWs are fully-depleted and the grid is down, simply connecting power to the jump header won't do anything AFAIK, as the solar inverters still won't start without seeing a 120V/60Hz signal, and where is that going to come from? The PWs would have to still have some energy in them to get their inverters up and running (for at least 5 minutes it seems) or the grid would need to be present (or you'd have to also have a generator, etc) before the solar inverters would see something they'd be happy synchronizing with to start supplying energy.

There has to be a way to have the inverters turn on without an external signal, right? I'm thinking for 100% off grid applications that would be a requirement.

 

[woferry]

I'm no expert here, but my understanding is that requires a non-grid-tied inverter. I.e. one that is designed to work without a grid, which is not what is normally used in a grid-tied system.

 

[gpez]

I'm no expert here, but my understanding is that requires a non-grid-tied inverter. I.e. one that is designed to work without a grid, which is not what is normally used in a grid-tied system.

That's fair. Now I wonder if some grid-tie inverters can be configured to operate off-grid temporarily to bootstrap...

 

[cwied]

The Powerwall inverters are exactly such inverters. They can both sync to the grid and run in "island" mode.

I believe that the Powerwalls shut down before they really are fully depleted so that there is enough energy to come back up when the solar could produce.

 

[Frankman60]

If that is true, what would be the purpose of the two terminals? For example, if the PWs get depleted overnight during a grid outage the house power shuts off. In the morning when there is enough sun shining on the solar array, the PWs should be able to begin charging and turn back on to power the house right? I”m just trying to figure out the steps in advance of something like this happening.

The Powerwall inverters are exactly such inverters. They can both sync to the grid and run in "island" mode.

I believe that the Powerwalls shut down before they really are fully depleted so that there is enough energy to come back up when the solar could produce.

 

[cwied]

If that is true, what would be the purpose of the two terminals? For example, if the PWs get depleted overnight during a grid outage the house power shuts off. In the morning when there is enough sun shining on the solar array, the PWs should be able to begin charging and turn back on to power the house right? I”m just trying to figure out the steps in advance of something like this happening.

I'm guessing the terminals on the gateway are for when the Powerwalls have no reserve left. According to this thread, the Powerwalls will shut down overnight when they reach 5% charge so they have enough to restart in the morning. According to @MorrisonHiker, the part that didn't work was having the Powerwalls start charging again after restarting, perhaps because of the cold temperature. They did start recharging from solar, though, so presumably if conditions were right it would have worked.