Tesla PwrWall/Gateway shutdowns

[jaggedben]

So where is this frequency signal generated? The gateway or the battery?

The battery inverter provides the reference frequency when the grid is not connected. The gateway just monitors the grid and tells the battery inverter whether there's a stable grid or not. If the grid is connected and stable then the gateway closes the circuit to the grid and the battery and solar inverters sync to the grid. If the grid is disconnected then the solar inverters sync to the battery inverter.

And where's the other 'receiving' end?

If the reference frequency is not within spec the solar inverters will shut down. This is how the battery unit protects itself from overcharging, it shifts the AC frequency (perhaps from 60.0Hz to 58.7Hz ) and this results in the solar inverters going offline. This 'bang, bang' method of controlling the solar output isn't very sophisticated but it has the advantage of working with any grid-tied solar inverter without requiring other compatible technologies. Tesla isn't the only ones to use this method.

And if this is how it signals to the panels, how can that result in a battery shutdown if the battery has charge?

It shouldn't (except in the scenario I explained in post #4). I suppose it might be the trigger for some internal failure, but if all equipment is working as designed it should not result in a power interruption.

 

[GoldDigger]

FWIW, I believe that the frequency-based control interface between the SMA Sunny Island battery inverter and their Sunny Boy grid tie inverter is proportional rather than on-off.

Sent from my XT1585 using Tapatalk

 

[caribconsult]

The battery inverter provides the reference frequency when the grid is not connected. The gateway just monitors the grid and tells the battery inverter whether there's a stable grid or not. If the grid is connected and stable then the gateway closes the circuit to the grid and the battery and solar inverters sync to the grid. If the grid is disconnected then the solar inverters sync to the battery inverter. This is our setup. The grid only goes on when (1) it's after sundown and (2) the battery is too low to last though the night and into the morning when the sun is up. And it is under this "not connected" condition that the shutdowns occur. I can't say that connecting the grid prevents it, because or grid usage so infrequent and usually not more than 3 hours, but I don't believe we've ever seen it happen while the grid is on. And according to Tesla, we can use the grid this way without violating any terms of system usage. It's OK with them.



If the reference frequency is not within spec the solar inverters will shut down. This is how the battery unit protects itself from overcharging, it shifts the AC frequency (perhaps from 60.0Hz to 58.7Hz ) and this results in the solar inverters going offline. This 'bang, bang' method of controlling the solar output isn't very sophisticated but it has the advantage of working with any grid-tied solar inverter without requiring other compatible technologies. Tesla isn't the only ones to use this method. We've observed the panels shutting down almost every day, as the battery reaches full charge. If I look a the graph in the app for solar energy, you can see a big 'notch' in the output right around 12:30 pm - 1pm, then,usually within an hour, the panels will begin to cycle on and off as the battery is discharged by the house. Kind of a 'topping off' mode. That's all happening nicely and normally. Sometimes the house load is low (>500W) that the panels supply all the energy to the house - nothing from the battery. This cycle continues until sundown, at which point the battery takes over the entire load. . You are saying there's no way this frequency shift should shut down the battery, just the panels. Is that correct?. So we are left with the question, "what is triggering the battery shutdown?" It can't be the grid because it's not connected. The frequency shift is to control output from the panels, not the battery. By process of elimination, what's left?



It shouldn't (except in the scenario I explained in post #4). I suppose it might be the trigger for some internal failure, but if all equipment is working as designed it should not result in a power interruption. Something internally is intermittent. Do you suspect it's in the gateway, or the battery, or just that the connection between them is lost and that shuts it down? No one has explained just exactly how this thing controls itself beyond what you said here about the frequency drop shutting the panels down to prevent overcharge. And that seems to be working. Just how many other things in there could initiate a shutdown AND a restart at the time? I did say the system almost always restarts itself, sometimes with seconds, sometimes minutes. It's some sort of electronic reset button. Tesla is ordering the installers to come back and get on the phone with them and run some specific tests. We'll see.


Thanks as always for your input.

 

[ggunn]

FWIW, I believe that the frequency-based control interface between the SMA Sunny Island battery inverter and their Sunny Boy grid tie inverter is proportional rather than on-off.

Sent from my XT1585 using Tapatalk

That is correct. As the SI shifts the frequency up or down from 60Hz, the SB ramps down its output until the frequency hits its max or min (I don't remember the exact values) where the inverter totally shuts down. IIRC the SI alternates between shifting the frequency up and down so that synchronous electric clocks do not keep drifting in the same direction.

 

[caribconsult]

Interesting discussion about the frequency shift as a signal. Seems like a heavy-handed way to shut down the panels, but that's how it works. Can any of this cause the battery to shutdown? The on/off action of the panels is easily observed in the app and the pattern is virtually identical each day, unless we have lots of clouds and erratic sun. Around 12:30pm the battery is fully charged and the panels shut down, as ordered by this frequency shift, and within an hour+ apparently something similar causes the panels to start putting out again when the battery percentage drops below about 94-5%. All that works repeatedly.

The shutdown can occur any time, day or night, loaded heavy/loaded light, grid NOT connected. What do you think?

 

[ggunn]

Interesting discussion about the frequency shift as a signal. Seems like a heavy-handed way to shut down the panels, but that's how it works. Can any of this cause the battery to shutdown? The on/off action of the panels is easily observed in the app and the pattern is virtually identical each day, unless we have lots of clouds and erratic sun. Around 12:30pm the battery is fully charged and the panels shut down, as ordered by this frequency shift, and within an hour+ apparently something similar causes the panels to start putting out again when the battery percentage drops below about 94-5%. All that works repeatedly.

The shutdown can occur any time, day or night, loaded heavy/loaded light, grid NOT connected. What do you think?

Your PV system shuts down at night? Horrors!

 

[caribconsult]

Tesla Powerwall issues

Tesla Powerwall issues

Well, sure the PV shuts down at night since there's no sun. Whether it's connected or not is up to the gateway and I don't know if you could backfeed the array without damaging something. It appears the gateway is handling the PV shutoffs during the day perfectly well. Once the battery is charged, the PV output goes to zero for an hour or so, the house runs on the battery for a while, and once the battery level sinks to around 94% the panels will start generating power according to how much light is present until the battery is ~98%, then the PV's turn off. This 'top it off' mode cycles every few minutes or so, and continues until sundown. That is all happening repeatedly, every day and seemingly without interruption. Even if the panels didn't shut down, wouldn't the excess attempt to 'backfeed' into the grid, if it were present? And if no grid present, as is generally the case, obviously the excess goes nowhere, except it will produce an error message on the meter if we leave it connected after the battery is full, which I don't thing is a wise move. We don't have nor will we get a net metering system where you sell your excess back to the power company. No dice.

The shutdowns my posts refer to are complete power off battery shutdowns and total loss of power to the house, followed usually withing seconds or minutes by an automatic reset and restoration of power. Since the battery is the only source of power at night, where else could an AC failure come from? The grid remains disconnected as is normal unless the battery is drained because of not much sunlight. In that case, and ONLY in that case, after sundown, the grid goes on and into the gateway, the house runs on some combination of the grid, or the battery, or both, depending on the situation, the battery charges,and once charged, at that point we shut off the grid (via breaker) and by the next day, if the sun is reasonably out, the battery will be charged fully by around noon. Wash, rinse, repeat.

This works well nearly all the time. I've observed this behavior in the app and it can be seen how the gateway controls the flow of power. I can also access the PowerWalls User Interface directly via the wifi in the power wall. The reaction time is close to real time, but the app gives you way more information about how your system is running, albeit it less responsive. Our grid usage is minuscule. But the shutdowns can occur anytime of the day or night, under any load conditions, and under any level of battery charge, so long as it is charged, and we have never let it get down to zero. Whether the shutdowns would stop if the grid were fully connected is something I can not answer, since we don't (and Tesla says this is OK) us it like that, connected 100% of the time. Only when needed.

That's the situation. What could cause the battery to either shut itself down, or receive a command from some other part of the system, to shut down? what? It has to be somewhere between the battery, and its inverters and control circuits, to the gateway, with its control circuits. It's all done electronically, there's no big switches to throw.

So what do you think? Have you seen this behavior before in a Tesla system? Thanks for your input.

 

[ggunn]

Well, sure the PV shuts down at night since there's no sun. Whether it's connected or not is up to the gateway and I don't know if you could backfeed the array without damaging something. It appears the gateway is handling the PV shutoffs during the day perfectly well. Once the battery is charged, the PV output goes to zero for an hour or so, the house runs on the battery for a while, and once the battery level sinks to around 94% the panels will start generating power according to how much light is present until the battery is ~98%, then the PV's turn off. This 'top it off' mode cycles every few minutes or so, and continues until sundown. That is all happening repeatedly, every day and seemingly without interruption. Even if the panels didn't shut down, wouldn't the excess attempt to 'backfeed' into the grid, if it were present? And if no grid present, as is generally the case, obviously the excess goes nowhere, except it will produce an error message on the meter if we leave it connected after the battery is full, which I don't thing is a wise move. We don't have nor will we get a net metering system where you sell your excess back to the power company. No dice.

The shutdowns my posts refer to are complete power off battery shutdowns and total loss of power to the house, followed usually withing seconds or minutes by an automatic reset and restoration of power. Since the battery is the only source of power at night, where else could an AC failure come from? The grid remains disconnected as is normal unless the battery is drained because of not much sunlight. In that case, and ONLY in that case, after sundown, the grid goes on and into the gateway, the house runs on some combination of the grid, or the battery, or both, depending on the situation, the battery charges,and once charged, at that point we shut off the grid (via breaker) and by the next day, if the sun is reasonably out, the battery will be charged fully by around noon. Wash, rinse, repeat.

This works well nearly all the time. I've observed this behavior in the app and it can be seen how the gateway controls the flow of power. I can also access the PowerWalls User Interface directly via the wifi in the power wall. The reaction time is close to real time, but the app gives you way more information about how your system is running, albeit it less responsive. Our grid usage is minuscule. But the shutdowns can occur anytime of the day or night, under any load conditions, and under any level of battery charge, so long as it is charged, and we have never let it get down to zero. Whether the shutdowns would stop if the grid were fully connected is something I can not answer, since we don't (and Tesla says this is OK) us it like that, connected 100% of the time. Only when needed.

That's the situation. What could cause the battery to either shut itself down, or receive a command from some other part of the system, to shut down? what? It has to be somewhere between the battery, and its inverters and control circuits, to the gateway, with its control circuits. It's all done electronically, there's no big switches to throw.

So what do you think? Have you seen this behavior before in a Tesla system? Thanks for your input.

I was under the impression that the PowerWall systems have not yet been blessed by Tesla for off-grid operation. My company deals with them a lot, but only for backup apps.

 

[caribconsult]

We are using it in backup mode, it's just ALWAYS in backup mode unless we turn on the grid. Tesla say this is ok. We get more than enough sun almost every day to fully recharge the battery by midday, and keep it charged well into the 93+% range as the sun sets. It thinks the grid has failed and just goes into backup mode and charges and discharges, generally without a burp, except as I've noted...randomly, and seemingly without pattern.

 

[jaggedben]

Yeah, I would say it doesn't make all that much sense if Tesla's says that your system can operate indefinitely in backup mode in case of a power outage, but that you're not allowed to operate it that way on purpose. Either it can do it, or it can't.

 

[caribconsult]

PowerWall Issues

PowerWall Issues

That's what Tesla says; backup mode is ok, but I've already caught them twice giving out conflicting info. When this shutdown issue first appeared I called into tech support and was told by at least 2 different reps that others were having the same issue and Tesla had this under study. In my most recent calls with both a rep and a supervisor at tech support I was told I'm the only one with this issue, so Tesla's veracity here is somewhat questionable. But they have consistently said I could run it as a full-time backup system and by cutting off the grid, that fools the gateway into thinking there's an outage, and so it shifts the load to the battery and/or the panels, as conditions warrant. This seems to be working repeatedly, every day. What in this system could trigger a random total AC output failure, battery off, pilot light off? It's not being overloaded.

Thanks again for your help.

 

[caribconsult]

Recent developments

Recent developments

It finally got told to the technician who showed up here to call Tesla. Unfortunately he arrived around 9am, at which time there was no one to speak to at Tesla because they are 3 hour earlier in the day than us! So he waited, finally got the Tesla people on the phone and after a number of tests, they discovered that my software in the gateway (I guess that's where it lives) was not the latest version, so Tesla upgraded it via their internet link or possibly his phone app (not sure) to the gateway. It is now on the most recent software version and possibly that will cure the shutdown issue.

We are in a "wait and see" mode. If it doesn't shutdown within the next 30 days, I will declare it a solved issue and pay them their final payment. If if shuts down even once, we're going to have a very different scenario. More to follow, but thanks again to everyone who contributed here...we certainly covered all the bases, thanks to you folks.

 

[Sierrasparky]

So you still owe the installer money?

Who specified the install Tesla or the installer?

Did they supply the gateway and power wall?

If they just acted as a installer and not specifier I don't see how the installer could be responsible if installed as per manufacture specifications. That would be like installing electrical for a huge chiller system and the equipment delivered is oversized and not per plan and the result is a less than adequate electrical system. Then you hold the Electrical contractors money hostage because you don't want to pay for the upgrade. Something's are just not the fault of the electrical contractor and thus they need to be paid for their work. This may be one case.

 

[caribconsult]

The Tesla system was purchased from the installer and they specified the equipment, not me. Tell me what leverage I would have to get this fixed correctly if I didn't owe them any money? Despite the fact that I owe them 5 figures their customer service is terrible. Do you think it would be better if I didn't owe them anything? I'm funny that way. I don't pay for stuff that doesn't work as advertised. This issue has been going on since Feb 1. What would you do?

 

[Sierrasparky]

The Tesla system was purchased from the installer and they specified the equipment, not me. Tell me what leverage I would have to get this fixed correctly if I didn't owe them any money? Despite the fact that I owe them 5 figures their customer service is terrible. Do you think it would be better if I didn't owe them anything? I'm funny that way. I don't pay for stuff that doesn't work as advertised. This issue has been going on since Feb 1. What would you do?

I think that since the installer specified and provided the equipment then you should stand your ground. The key here is specified and installed. If it were specified by others and the installers just followed manufacture instructions then the performance is all the specifier. In your case it is both. So double wammy.

 

[caribconsult]

The ongoing Tesla issues

The ongoing Tesla issues

Standing my ground is exactly what I'm doing, and I believe you understand the situation correctly. The system does everything it's supposed to do - the battery charges from the panels, the gateway distributes everything correctly, etc - it's just also doing something it shouldn't, which is the shutdown/reset business. As I mentioned, a technician was here this week and spoke with Tesla and they upgraded the software to the most recent version. Hopefully that will fix it. If it does, they get paid. I really regret buying this system from this company - their customer service is almost non-existent.

If it continues to shutdown, I'm going to give them three options:

(1) take all your stuff out of here and refund all my money, but not in that order, or (2) I want a new battery and new gateway, and the installers can send the old one back to Tesla as defective and this won't cost them a dime, or (3) I'm going to take the money I owe them and use it to replace the battery and gateway with some other distributor. I don't want to blow my horn here, but I'm a New Yorker by birth and we don't take kindly to being jerked around like this. We've learned from living in PR for almost 16 years that you NEVER pay all the money before the job is done and done right.

Thanks very much for your advice and help with this issue...it's only going on about 4 months now and I'm up to my neck with it and want to finish it and move on with my life.

Jeff

 

[ramsy]

.. I'm a New Yorker by birth and we don't take kindly to being jerked around like this. We've learned from living in PR for almost 16 years that you NEVER pay all the money before the job is done and done right..

If I recall correctly, Tesla donated those power walls to Puerto Rico thru disaster-relief channels, best described as donation conversion disbursed by lottery, not to any intended party. If I understand this process correctly, a qualified local installer may have acquired the Power Walls arriving in Puerto Rico at substantial discount. If so, the installers material cost may be minimal, and volunteer laborers may be directed to install some of these systems.

It was my understanding, the only way to ship larger items to an intended individual or address in Puerto Rico after the disaster was to hire private shipping companies.

 

[caribconsult]

The ongoing Tesla issues

The ongoing Tesla issues

You may be right about Tesla donating PW's to some relief organizations here but that's not how we got ours, and we don't know anyone who got one like that. We bought it retail from an authorized Tesla dealer that does only solar installations, with a variety of battery options, depending on what you can afford. Since we were part of the rush to acquire solar in the aftermath of María, we wound up paying top dollar and had to wait almost 3 months before the batteries arrived and they could come install. But I'm not complaining about the price, or the quality of the installation, or the delay...they did a very nice, neat job putting in all the panels, the tubing for the wires, switch boxes, breakers, etc. Very nicely done. I'd gladly pay them the balance and be done with this.

But what we have here is a case of who am I going to take the problem to, and with what leverage? If you've read the prior posts then you ought to know what is going on with my system, and it's not right. Tesla says it's not right. The dealer admits it's not right. So they're going to fix it, either my way, or my way, or my way, at this point. The first my way is they find the problem and fix it. The second my way is they give us a new battery and new gateway and hopefully that will fix it. The third my way is I use the money I owe them to buy a new battery and gateway elsewhere and they can stuff it. They were here a few days ago and updated the gateway or battery software (not sure which) and perhaps that will cure the problem. We're in a 30 day "wait and see" mode right now.

If you have any better ideas, I'd love to hear them. At this point I am so fed up with dealing with this issue for 4 months I'd try almost anything within reason. Paying them before it's fixed is not one those things.

 

[jaggedben]

Without seeing the contract I of course am ultimately speculating, but I would bet that you are probably already obligated by contract to have paid your installer. If so, not paying may give you some 'leverage' short term, but longer term, depending on exactly what the contract says, it may give the contractor more leverage to say 'you never paid us, you're on your own, and no, we're not obligated to remove it for you for free or to service Tesla's warranty for no compensation'. If you drive them to a point where they feel you're such a pain in the ass that's it's not worth dealing with you even to collect a significant remaining balance, then they won't, especially if they would hold the advantage in a court of law because they really fulfilled their contract obligations (installing the system) and you didn't (pay them). i.e. they are not worried about being sued, they probably could sue you but won't because it's not worth their time, but they could wash their hands of you. Contractors are usually pretty busy people. If you think it makes sense to punish them financially for something that's hardly their fault, good luck with that. Paying them but threatening a bad public review might give you better leverage than not paying them. Teaming up with them to put pressure on Tesla might be another strategy to consider, too.

 

[caribconsult]

The ongoing Tesla issues

The ongoing Tesla issues

All valid points, JaggedBen, but the installer was the one who sold us this, not Tesla. Their name is on the invoice. They said we don't have to pay until it's working right, so I don't feel like I am violating anything here, just holding them to their word. So let's call them what they are: the seller.

I have been extremely patient with them for 4 months, but they don't talk to each other at the seller company, so messages die on the desk of whomever you talked to, and their hi-tech phone system STILL doesn't work right. It is now some 7 months since the storms. and it is very frustrating to get them to straighten this out, so I don't know exactly how much service we might lose over this. The guarantee is by Tesla, not the seller company, and Tesla is well aware of the issues I've been having with them. I'm willing to take my chances, since signing off on a system that is exhibiting problems doesn't sound like the way to go.

Getting them to team up with Tesla would probably be a good idea, but as I said, the seller company is very blasé, uncommunicative and seemingly doesn't give a hoot about payment....they aren't barraging us with calls to pay, and in fact I gave them an additional 40% of the price a month ago as a show of good faith even though we were still having issues. My biggest regret here is buying it from them. The one reference we got about them was good, so we committed. As you'll note, power was (and still is, to some extent) a major issue here in Puerto Rico after Maria. So I think your suggestions are all good ideas but the reality of who the seller is makes that difficult. Let's hope the recent upgrade to system software resolves this. It's only been about 5 days, but there have been no interruptions SO FAR.

Thanks for your input.