diy solar

diy solar

EG4 18kPV Q+A general thread

100% zero export is not a simple task.
Surges happen. And responding quickly enough to never export a single electron is very hard.
You are always taking a risk. If you install a grid-tied system without having an agreement with the utility company.
No it should be very simple. It does not blow up if I have my main grid turned off and charges from PV and powers all my loads no problem.
Solark manage it just fine.
With my usage scenario I just want it to be a battery charger from time to time. I don't want it to power loads from the grid or feedback to the grid, just charge the battery.
 
No it should be very simple. It does not blow up if I have my main grid turned off and charges from PV and powers all my loads no problem.
Solark manage it just fine.
Off-Grid is easy - grid connect with Zero export is hard.
Why? Measuring amps is really hard, just think about the sensor on the grid side which is measuring how much power is exported - it's a stupid coil which is wound around the conductors. Those coils are not perfect, maybe not even calibrated. Zero on one coil might be not zero on another coil.

Usually a good amp sensor has a tolerance of +-0.5% bad ones have 2.5% The sensor needs to capable of 200A.

200A * 0.5% = 1A - so you can have 1A - per leg - flowing without the inverter even knowing about it.

In a off-grid application - the voltage just goes up an down by half a volt to compensate for load differences - something you can not do in grid connect.
 
No it should be very simple.
But, it's not.
It does not blow up if I have my main grid turned off and charges from PV and powers all my loads no problem.
Yes, disconnecting works perfectly.
Solark manage it just fine.
I don't think so. I'm pretty sure that people have complained about it drawing from the grid, when they didn't want it to. (Which is the opposite side of the same issue)
With my usage scenario I just want it to be a battery charger from time to time. I don't want it to power loads from the grid or feedback to the grid, just charge the battery.
That would be nice.
But requires more dedicated hardware.
Would need two separate connections to the battery. Instead of a bidirectional one.
 
I don't think that any grid-tied system can guarantee 100% zero export. (It's a grid-tied system)
Unless it can completely (physically) disconnect from the grid.
Trying to control individual electrons in a free flowing connection. Is a reactive response. Some have to be seen getting through, before it even sees a need to react.
Good point. This may be the crux of the issue.
 
I don't think that any grid-tied system can guarantee 100% zero export. (It's a grid-tied system)
Unless it can completely (physically) disconnect from the grid.
Trying to control individual electrons in a free flowing connection. Is a reactive response. Some have to be seen getting through, before it even sees a need to react.
Maybe we are using different terminology here but I thought the issue is if you are trying to zero out grid usage by using current transformers to detect how much power the inverter needs to produce in order to only produce enough to prevent it from having to use any grid supplied power. If you are grid connected and only switch to it when batteries get too low or if you are using inverter in "UPS" mode such that you immediately switch to battery/solar when grid goes down then in these cases it should work flawlessly.

My 15 year old Outback "grid tied" inverter had no issues with these two cases and never sent any power back to the grid.

It's only when dealing with current transformers to detect usage that there would be potential issues with grid-tied.
 
Frankly, if it can not handle those two cases I mentioned above properly then I don't know how something could be UL1741 certified.
 
I don't think so. I'm pretty sure that people have complained about it drawing from the grid, when they didn't want it to. (Which is the opposite side of the same issue)
That would be an interesting question to pose, maybe in a new thread. Have any of the SolArk owners had this problem and if so how did they address it. Or any other grid tied inverter that claims to offer zero export.

I'm certainly no EE but it seems to me it should be relatively easy to design the inverter so that it is physically disconnected from the grid until and unless the batteries get down to a user defined level.
 
I don't think so. I'm pretty sure that people have complained about it drawing from the grid, when they didn't want it to. (Which is the opposite side of the same issue)
probably Sol Ark went on the safe side and configured/calibrated it that it always draws a little bit - instead of back-feeding.

I'm certainly no EE but it seems to me it should be relatively easy to design the inverter so that it is physically disconnected from the grid until and unless the batteries get down to a user defined level.
I'm a EE and just from my understanding of measurement technology and sensors - I can tell you that true "close to" Zero Export is not possible with current means of measuring energy flow.

You can setup "Zero Export" in a way that it's essentially building a one way - which means like a online UPS - so you are always going through a charger - battery Inverter. = The Inverter is never connected to the Grid.

1688567289097.png

From what I read on the 18kPV documentation and from what I see in this thread:

"grid peak shaving" = Conservative true No Export, While "Zero Export" = almost Zero Export
 
That would be an interesting question to pose, maybe in a new thread. Have any of the SolArk owners had this problem and if so how did they address it. Or any other grid tied inverter that claims to offer zero export.

I'm certainly no EE but it seems to me it should be relatively easy to design the inverter so that it is physically disconnected from the grid until and unless the batteries get down to a user defined level.
This comes from the value of an AIO.
Different parts pulling double duty.
Bidirectional connections to battery and grid.
You can't physically have input, without having output. (Or vice-versa)
 
I don't think that any grid-tied system can guarantee 100% zero export. (It's a grid-tied system)
Unless it can completely (physically) disconnect from the grid.
Trying to control individual electrons in a free flowing connection. Is a reactive response. Some have to be seen getting through, before it even sees a need to react.
thinking about what you said, perhaps there is a way to set it to only import during the early hours of the morning like 4am to 8am as the rest of the time I can run on solar and battery. I cannot disconnect completely or I have no power once the batteries dies around 4-5am (yes I am planning to buy one more and that will allow a total disconnect). but for now I still cannot see how that would stop the export if the no export option is on and still its sending .4-.7kw a day back into the grid its just a matter of time until I get a fine from utility or worse.
 
thinking about what you said, perhaps there is a way to set it to only import during the early hours of the morning like 4am to 8am as the rest of the time I can run on solar and battery. I cannot disconnect completely or I have no power once the batteries dies around 4-5am (yes I am planning to buy one more and that will allow a total disconnect). but for now I still cannot see how that would stop the export if the no export option is on and still its sending .4-.7kw a day back into the grid its just a matter of time until I get a fine from utility or worse.
As long as you are importing, you won't be exporting.
You just have to set it up to always import some power.
 
probably Sol Ark went on the safe side and configured/calibrated it that it always draws a little bit - instead of back-feeding.


I'm a EE and just from my understanding of measurement technology and sensors - I can tell you that true "close to" Zero Export is not possible with current means of measuring energy flow.

You can setup "Zero Export" in a way that it's essentially building a one way - which means like a online UPS - so you are always going through a charger - battery Inverter. = The Inverter is never connected to the Grid.

View attachment 155888

From what I read on the 18kPV documentation and from what I see in this thread:

"grid peak shaving" = Conservative true No Export, While "Zero Export" = almost Zero Export
not sure I follow this but since I am trying to run with as little as possible grid input and zero grid export you saying to use the peak shaving option to allow only limited power say .1kw (the smallest allowed) as input from grid during say 7am to 5pm then the other time since solar is down to almost nothing it has nothing to export except battery and export is turned off so it should stop the constant in/out fluctuations I am seeing,
is that correct?
Screenshot 2023-07-05 at 9.05.32 AM.png
 
The manual has this.. Are we saying then the Fast zero export is not working as described? Do you have that set @rmacgurn ?

Feed-in grid: This selection is for users to set a zero-export function. If exporting solar power is not allowed, users need to disable the ‘Feed-in Grid’ option. If a user’s utility meter is tripped with minimal solar export, ‘Fast zero export’ can be enabled; thus, the export detection and adjustment will take place every 20ms, which will effectively avoid any solar power being exported. If export is allowed, users can enable ‘Feed-in Grid’ and set a maximum allowable export limit in ‘Feed-in Power (kW).’
 
The manual has this.. Are we saying then the Fast zero export is not working as described? Do you have that set @rmacgurn ?

Feed-in grid: This selection is for users to set a zero-export function. If exporting solar power is not allowed, users need to disable the ‘Feed-in Grid’ option. If a user’s utility meter is tripped with minimal solar export, ‘Fast zero export’ can be enabled; thus, the export detection and adjustment will take place every 20ms, which will effectively avoid any solar power being exported. If export is allowed, users can enable ‘Feed-in Grid’ and set a maximum allowable export limit in ‘Feed-in Power (kW).’
no fast zero export has only one functions to read the CT's faster then every 5 min. to limit the amount of export or control it better. the export to grid option is off and yet system still exports and imports when unnecessary or required to do so.
it is importing when I am putting out 6+kw of solar and house is only using under 1K but its still pulling in power from grid. also sending power back. its a like a cycle going back and forth. so I am going to try to stop the import using peak shaving options and try to limit the export using CT offset of -25W. below you can see the ping pong of import and export. just before 7am my solar goes above 2kw and no need for the grid but I am not interested in going to street to disconnect every day and reconnect at sundown. Inverter should handle this correctly.
Screenshot 2023-07-05 at 9.46.54 AM.pngScreenshot 2023-07-05 at 9.50.35 AM.png
 
no fast zero export has only one functions to read the CT's faster then every 5 min. to limit the amount of export or control it better. the export to grid option is off and yet system still exports and imports when unnecessary or required to do so.
it is importing when I am putting out 6+kw of solar and house is only using under 1K but its still pulling in power from grid. also sending power back. its a like a cycle going back and forth. so I am going to try to stop the import using peak shaving options and try to limit the export using CT offset of -25W. below you can see the ping pong of import and export. just before 7am my solar goes above 2kw and no need for the grid but I am not interested in going to street to disconnect every day and reconnect at sundown. Inverter should handle this correctly.
View attachment 155894View attachment 155895
I can help you out send me a personal message with your login info. I have had to fix this with other accounts already.
 
I can help you out send me a personal message with your login info. I have had to fix this with other accounts already.
Can you explain what exactly is the fix and why @rmacgurn isn't able to fix it himself? From my perspective it should either be a firmware update or a setting the end user can change and not something hidden that is magically fixed by someone remote.
 
Back
Top