diy solar

diy solar

Some zero export inverter questions

Texas-Mark

Solar Addict
Joined
Aug 4, 2021
Messages
1,284
1. If you have a grid-tie inverter that has zero export capability, I assume it has to have a CT to manage the zero export, correct?
2. Do all of these zero export inverter have a way to set a threshold for when to stop generating?
3. If you can set it for let's say 300 watts (coming from the grid) as the threshold, how quickly can it throttle down? In other words can it "accidentally" back-feed if the load suddenly drops (i.e. HVAC turns off)?
4. Related to question 2, if lets say a 120V grid-tie inverter is tied to L1, but L2 has a 500 watt draw, and the inverter did not throttle down fast enough when that 300 watt threshold was met, would you still have a 500 watt cushion because the electric meter is reading the total of both L1&L2?
5. Any particular inverter in the 6-8kW range that would do this?

The reason I ask, is I live rural (where you can do pretty much whatever you want) and for many years I was running a small grid-tie setup. I never generated more than I was using, but one day all of the planets lined up and I back-fed a little. The electric company (that had put in smart meters a few months earlier) called me up and asked about it. It should be noted that the two previous times my meter was changed out, they saw my solar setup and couldn't care less. But around the time of the smart meters, they had teamed up with some solar company, and that is when the flood of red tape came about. It was too much hassle to change things the way they wanted, so I switched to a solar-battery-inverter setup to power just certain loads like my pool pump and sheds. But now I want to go back to a grid-tie system but do not want another call from them.
 
Plenty of threads on this. I will summarize my position here.

"Basically, what you people are looking for is a stealth zero export. A lot of companies intended for grid tie probably do not target that, they target zero export that is good enough to satisfy Interconnect Agreement with zero export, which is more lax."

1) Yes
2) Not sure what you mean. Some of them will let you set a target level for import or export, that is not zero
3) Sounds like you're thinking of playing the game where, if you consume some from grid that will cover up glitches. You would need to find a level that is bigger than the lag / glitch from the zero export function
4) The meter could have independent channel for L1/L2. The meter could have high resolution rollup of export. I did a deep dive a while back into some spec sheets / cut sheets for these meters and both such capabilities are readily available. And if you game out what kind of hardware is needed, it's very cheap to them to add. If they really want to catch you, it's probably easier for them to write this software / buy it from the Smart Meter vendor, than it is for you to improve your zero export tightness.

Interconnect agreement is not a big deal where I am, but I also have a high tolerance for red tape. Like $800 + time (plans/permits/interconnect application + 20 hours of paper shuffling and emails)
 
Plenty of threads on this.

Yes, but too many too try and find the specifics I was looking for.

2) Not sure what you mean. Some of them will let you set a target level for import or export, that is not zero

What I meant was having a setting that as long when it detected that power from the grid was below some threshold (I used 300 watts), it would throttle back. How else can it be zero export work, unless it can monitor grid flow in?

3) Sounds like you're thinking of playing the game where, if you consume some from grid that will cover up glitches. You would need to find a level that is bigger than the lag / glitch from the zero export function

Yes, that is the general idea. That is why I was asking how much of a "lag" there is, and what is a safe level to guarantee never to back-feed. If that guarantee is even possible.

4) The meter could have independent channel for L1/L2.

I had not thought of that. All I know is I was back-feeding probably less that 100 watts and they were calling me.

I could just expand my existing system to include more circuits, but that means more batteries.

It just really cheeses me off that there is so much red tape for a grid-tie system. By design they are fool proof. Heck, I was a beta tester for one of the original Xantrex grid-tie inverters like 20 years ago. I lived in CA at the time and just "plugged it in". Nobody cared.
 
Yes, but too many too try and find the specifics I was looking for.
OK, no problem. You've thought about this a decent amount so I'm happy to give my $.02 again, for whatever it's worth.

What I meant was having a setting that as long when it detected that power from the grid was below some threshold (I used 300 watts), it would throttle back. How else can it be zero export work, unless it can monitor grid flow in?
There are CTs that go into a power or energy calculation, and it seeks to adjust the inverter output to reach the set point, as averaged out over time. That set point can be 0W and it will still work, however it will have to see a deviation to be able to act.

Oh did you mean you wanted 300W so that the deviation is still net consuming so not a spurious export that the meter can see?

Yes, that is the general idea. That is why I was asking how much of a "lag" there is, and what is a safe level to guarantee never to back-feed. If that guarantee is even possible.
With grid tie equipment on interconnection agreement the lag only has to be as good as the zero export terms stipulate, and there is a good amount of slop. That's probably the majority of sales volume in the US.

I wouldn't be surprised if there's equipment for places like South Africa where the grid is terrible and the unsanctioned grid tie enforcement is heavy, that will seek to achieve a lower lag. I don't know how you would shop for it in the US though.

I have never seen a zero export bounds / lag tightness spec on a datasheet, and I doubt anyone else here has either, that has been on those threads.

I could just expand my existing system to include more circuits, but that means more batteries.
Another thing you can consider is double conversion system.

You go

AC -> Charger (like chargeverter) -> battery bus -> AIO

There's zero chance they can detect this.

However you need enough inverter capacity to service the additional circuits & your whole surge load on the house. And you would need to be smart about turning off the extra inverters if the grid goes out, because otherwise the BMS will disconnect. Or add load shed hardware. There is also a 10-15% loss from the double conversion and step-down/up wrt 48V.
 
Oh did you mean you wanted 300W so that the deviation is still net consuming so not a spurious export that the meter can see?

Yes, I wanted a "fudge factor" to never accidentally back-feed. Mainly I just want to offset some of my HVAC usage, so that number could be higher if need be.
Another thing you can consider is double conversion system.

You go

AC -> Charger (like chargeverter) -> battery bus -> AIO
There is also a 10-15% loss from the double conversion and step-down/up wrt 48V.

Yes, that loss is a reason I was trying to avoid that route. I was considering doing this just for the A/C condenser unit, but have to crunch some numbers to see if it is feasible. And find an inverter that can handle the 4 ton unit.
 
Yes, that loss is a reason I was trying to avoid that route. I was considering doing this just for the A/C condenser unit, but have to crunch some numbers to see if it is feasible. And find an inverter that can handle the 4 ton unit.

Is it a single speed or inverter?

With a single speed there is the trick of picking a grid tie inverter that is always smaller than the single power level. And then naughtily connect the inverter to the load side of the start contactor (or some other interlocked way to start). Microinverters or small grid tie string.

Two speed probably works ok too

Obviously with the grid support surge is no worry.

With an inverter AC with way more levels of power the matching is tricky. But you can get away with a smaller inverter. With a single speed you need to be able to eat the surge. With inverter AC the surge is lower, so you only have to upsize the inverter to handle the crest power from not fully corrected power converter. That’s probably 20% (albeit continuous).

For a non stackable unit a lot of people have been playing with SRNE 10K. It is 120/240 in one box (as pretty much every modern one is)
 
Is it a single speed or inverter?

Single.

With a single speed there is the trick of picking a grid tie inverter that is always smaller than the single power level. And then naughtily connect the inverter to the load side of the start contactor (or some other interlocked way to start).

This was actually another thought I had. I was thinking of using the contractor to energize a relay that would turn on a grid-tie inverter when running. Of course the down-side is lot of solar going to waste when it's not running.
 
This was actually another thought I had. I was thinking of using the contractor to energize a relay that would turn on a grid-tie inverter when running. Of course the down-side is lot of solar going to waste when it's not running.
You could attempt to do the same with a hybrid battery inverter. It’s more complicated though. You would need to string CTs out to the AC and zero export to CT.

Some people here have rigged up microinverters to batteries, that would work here too and save some complexity. At the expense of needing to eBay a big stack of microinverters if you want to go cheap and use old Enphase inverters. If you are willing to buy new there are some good ones that output 2kW from a single unit.

Maybe you can find a grid tie string inverter that takes 48V in (tough due to the silly voltage for a string).
 
Lost me on this one. Where does 48V come into play??
Oops.

So this addresses the stranded solar that is only usable for the AC as follows.

1) All solar is collected into a 48V battery. It is available directly (IE along DC bus, for real-time loads) or stored
2) You do: 48V battery -> microinverter -> line side of AC contactor
3) The power is delivered to AC contactor via a single 240VAC circuit (which is cheaper and safer than 48VDC)
4) There is no need for a CT here, you size the microinverter stack to be smaller than AC. The HMS-2000 microinverter I alluded to can also have a custom wattage limit set (you need to buy $200-300 management box). If you buy a ton of used M250 or whatever microinverters you just the right number enough in parallel to get the power you need. They're probably like $30 each at this point
 
Sol-ark can do what you wrote. It can use internal CT or an external CT to keep from selling back. You can set a minimum draw from 1-100 watts from grid to minimize risk of sell back. Keep your batteries under 100% to give buffer to dump power.

Might want to consider off-grid inverter. It can't export.
 
How do you
2) You do: 48V battery -> microinverter -> line side of AC contactor

4) There is no need for a CT here, you size the microinverter stack to be smaller than AC.

Lost again on this. Are these microinverters grid-tie? If so, how do you turn them on only when the HVAC is running?

If they are not grid-tie, how can they be smaller than the what the HVAC needs?
 
Yes, but it needs to be big enough for the load, and again would need way more batteries.
Eg4 6000 can passthrough 50 amps. It can only produce 25 amps. Consider the 3000 because it can supplement grid power with pv. The 6000 will switch to grid if it can't produce enough.
 
Lost again on this. Are these microinverters grid-tie? If so, how do you turn them on only when the HVAC is running?

If they are not grid-tie, how can they be smaller than the what the HVAC needs?
Yes, that is grid tie.

The connection to AC is

Grid -> contactor -> AC

The microinverter connects to load side of the contactor. When the AC turns on the

If you can’t access the contactor you can use a relay to keep the microinverter disconnected normally. Connect output side of micro to the AC circuit. Then have a current sensor on AC circuit to detect when it goes on, and connect the inverter into it via relay. This may be easier to deploy too since you could maybe put the micro, battery, sensor, and relay in the same room as subpanel feeding the AC.

IOW the microinverter turns on using its anti islanding/grid detection logic. It will only turn on after detecting grid for a long enough time.

(There’s probably better writeups out there for this, it’s a known trick).
 
You are assuming things happen instantly. How long does it take to disconnect from the grid when the ac turns off?

In the scenario described, the microinverter would be connected to the load side of the contactor. So as soon as the contactor opens, the microinverters are instantly disconnected from the grid. Even if they took a few milliseconds to turn off, it would go to the condenser unit, not the grid.
 
Back
Top