diy solar

diy solar

Deye Parallel 8kW not working the same

No way this is normal behavior.
So you feed into the grid (or to house with limiter out in AC in port ) and then grid is down. At that moment it should stop feeding into the grid.
And this unit has all the certificates from around the world, so it does, tested.
Something is not right on your setup or with your unit.
Can you make a pic and diagram of your system ? (Deye will ask for it also)
Also if inverter overloads it has a timer. After that it restarts itself.

We use 230V 50Hz here. But single phase 250A ... nope. Maybe they allow 60A single phase, and then they recommend to use 3 phase :) (and it is like 3x one phase)
thanks for all the responses and suggestions.
No way this is normal behavior. Maybe but I am not really surprised from my experience.
The set up is exactly in accordance with the Deye recommendation for parallel units as attached. The CT is placed as this diagram and connected to the master unit. The system operates with 'no export to CT". Total average load is 4-6 kW with max of 12 and min of 2. There is no import or export to the grid until the battery gets to 50% (between 1 and 6 am depending on how much sun ) and then the load comes from the grid until the solar can take over again. The UPS (Backup Load) is limited to 2-3 kW so that it can run for 4-6 hours when the battery is already at 50% and there is no grid power.
Might have certificates but these standards I believe only cover voltage and frequency and have long trip time delay requirements and the testing methods are basic and not simulating the type of conditions that we are experiencing.
They definitely do not reset until the grid is restored. I would also expect a timer before reset but this is not the case.
Basic 3 phase here is 3x230V in delta configuration (only benefit is that it saves on cable cost). Not possible to earth and hence no earth fault protection because this requires very sophisticated protection. We do also have 400V 3 phase star connected with neutral which is safer.
 
thanks for all the responses and suggestions.
No way this is normal behavior. Maybe but I am not really surprised from my experience.
The set up is exactly in accordance with the Deye recommendation for parallel units as attached. The CT is placed as this diagram and connected to the master unit. The system operates with 'no export to CT". Total average load is 4-6 kW with max of 12 and min of 2. There is no import or export to the grid until the battery gets to 50% (between 1 and 6 am depending on how much sun ) and then the load comes from the grid until the solar can take over again. The UPS (Backup Load) is limited to 2-3 kW so that it can run for 4-6 hours when the battery is already at 50% and there is no grid power.
Might have certificates but these standards I believe only cover voltage and frequency and have long trip time delay requirements and the testing methods are basic and not simulating the type of conditions that we are experiencing.
They definitely do not reset until the grid is restored. I would also expect a timer before reset but this is not the case.
Basic 3 phase here is 3x230V in delta configuration (only benefit is that it saves on cable cost). Not possible to earth and hence no earth fault protection because this requires very sophisticated protection. We do also have 400V 3 phase star connected with neutral which is safer.
 

Attachments

  • IMG_3940.jpg
    IMG_3940.jpg
    68.9 KB · Views: 67
Worst ... the phase imbalance must be HUGE :)
Maybe. Depends on what you do with the other two phases. Typical generation specs require continuous operation with 8% phase imbalance but usually much less than this.
 
thanks for all the responses and suggestions.
No way this is normal behavior. Maybe but I am not really surprised from my experience.
The set up is exactly in accordance with the Deye recommendation for parallel units as attached. The CT is placed as this diagram and connected to the master unit. The system operates with 'no export to CT". Total average load is 4-6 kW with max of 12 and min of 2. There is no import or export to the grid until the battery gets to 50% (between 1 and 6 am depending on how much sun ) and then the load comes from the grid until the solar can take over again. The UPS (Backup Load) is limited to 2-3 kW so that it can run for 4-6 hours when the battery is already at 50% and there is no grid power.
Might have certificates but these standards I believe only cover voltage and frequency and have long trip time delay requirements and the testing methods are basic and not simulating the type of conditions that we are experiencing.
They definitely do not reset until the grid is restored. I would also expect a timer before reset but this is not the case.
Basic 3 phase here is 3x230V in delta configuration (only benefit is that it saves on cable cost). Not possible to earth and hence no earth fault protection because this requires very sophisticated protection. We do also have 400V 3 phase star connected with neutral which is safer.


Ben tested the 1 minute restart when AC over current fault, and it worked (Sol-Ark, but they are almost the same)

The DC bus unbalanced ... that error does not seem right:
F26 - The DC busbar is unbalanced - When the hybrid in split phase mode, and the load of L1 and load of L2 is big different, it will report the F26
But you are not using it in split phase mode ....



What type of battery do you use? LFP or lead ?
 
Ben tested the 1 minute restart when AC over current fault, and it worked (Sol-Ark, but they are almost the same)

The DC bus unbalanced ... that error does not seem right:
F26 - The DC busbar is unbalanced - When the hybrid in split phase mode, and the load of L1 and load of L2 is big different, it will report the F26
But you are not using it in split phase mode ....



What type of battery do you use? LFP or lead ?
Yes it resets after a minute and goes to ON and NORMAL but after 2 seconds it trips again with the same fault. it only resets and stays reset after the grid power is restored.
Yes i do not understand the DC bus unbalanced error F26 but we dont really know what it is looking at to determin an unbalance. No it is not split phase.
Valve regulated lead acid.
 
Yes it resets after a minute and goes to ON and NORMAL but after 2 seconds it trips again with the same fault. it only resets and stays reset after the grid power is restored.
Write it all down to Deye. All scenario, step-by-step what happens.
It is faster this way. They will ask all the questions I did.
I am very interested to see they response.

I can not find right now the Deye inner switching pic, but here is the Victron ... and it is almost the same (the same in the interesting part)
Victron3.jpg

So left side AC in port. When a blackout occurs the relay (not a contact relay I think) on the left side opens.
That separates the inverter from the grid, it is the islanding solution.
So if this relay is not working, then inverter generated power goes out to unprotected load and grid.
And it can be an AC over current (you want to give power to all your neighbours on the same line)

Did you test that what happens if you make a blackout? Shut down your main breaker?
The other possibility is that in the right side, your protected load is bigger than what the inverter can produce.

Did you try to set
- from Zero export to CT (outer current sensor that you put into your line)
- to Zero export to Load (inner current sensor in the inverter )

Yes i do not understand the DC bus unbalanced error F26 but we dont really know what it is looking at to determin an unbalance. No it is not split phase.
That is what bothers me. Both inverters connected to the same battery? Same length of cable ?

Valve regulated lead acid.

How big is it? You know, lead-acid can not "jump" as fast as LFP.
So if there is a sudden big load on them then can happen they do not give enough DC power.
 
Write it all down to Deye. All scenario, step-by-step what happens.
It is faster this way. They will ask all the questions I did.
I am very interested to see they response.

I can not find right now the Deye inner switching pic, but here is the Victron ... and it is almost the same (the same in the interesting part)
View attachment 68539

So left side AC in port. When a blackout occurs the relay (not a contact relay I think) on the left side opens.
That separates the inverter from the grid, it is the islanding solution.
So if this relay is not working, then inverter generated power goes out to unprotected load and grid.
And it can be an AC over current (you want to give power to all your neighbours on the same line)

Did you test that what happens if you make a blackout? Shut down your main breaker?
The other possibility is that in the right side, your protected load is bigger than what the inverter can produce.

Did you try to set
- from Zero export to CT (outer current sensor that you put into your line)
- to Zero export to Load (inner current sensor in the inverter )


That is what bothers me. Both inverters connected to the same battery? Same length of cable ?



How big is it? You know, lead-acid can not "jump" as fast as LFP.
So if there is a sudden big load on them then can happen they do not give enough DC power.
Yes I did test what happens opening the two grid tie breakers simultaneously and it appeared to work ok but I have now made a detailed test with the attached results. Clearly the master inverter is not completely disconnecting from the grid and has some back voltage. Not good at all and I have already installed an independent grid disconnect system. Opening the grid breakers is seamless like the UPS function should be. Reclosing the grid breakers immediately restores grid voltage to the load as expected but the two inverters stop working and show F18 alarm. After about 60 seconds they try to start again and go back to normal operation. I am not sure how the re-synchronising to the grid happens but the F18 alarms may not be normal. It seems that there is a mechanical isolator for the grid since I could hear it on the slave but not on the master. They both have a mechanical isolator for reconnecting the inverter to the grid which you can hear.

Battery system is fine with 96 kWh and balanced cables and installation with balancers on all parallel stings.

Looks like I may have two faulty inverters !!!! I bought these because I thought the design and QA was better than the average China supply. It is impossible to maintain here unless you do it yourself.
 

Attachments

  • Inverter Voltages.pdf
    28.9 KB · Views: 31
Yes I did test what happens opening the two grid tie breakers simultaneously and it appeared to work ok but I have now made a detailed test with the attached results. Clearly the master inverter is not completely disconnecting from the grid and has some back voltage. Not good at all and I have already installed an independent grid disconnect system. Opening the grid breakers is seamless like the UPS function should be. Reclosing the grid breakers immediately restores grid voltage to the load as expected but the two inverters stop working and show F18 alarm. After about 60 seconds they try to start again and go back to normal operation. I am not sure how the re-synchronising to the grid happens but the F18 alarms may not be normal. It seems that there is a mechanical isolator for the grid since I could hear it on the slave but not on the master. They both have a mechanical isolator for reconnecting the inverter to the grid which you can hear.

Battery system is fine with 96 kWh and balanced cables and installation with balancers on all parallel stings.

Looks like I may have two faulty inverters !!!! I bought these because I thought the design and QA was better than the average China supply. It is impossible to maintain here unless you do it yourself.
I see other problems in this sheet.
Tell me about your neutral and ground lines :)
Do you have a TN-C-S system ? (sorry I do not know what system do you use there)
 
Incoming supply is 13.2 kV single phase with earth wire. We have a dedicated 13.2 kV/ 230 V 70 kVA pole mounted transformer. One side of the primary and secondary windings are strapped together and connected to the earth wire and also an earth stake at the transformer. There is then a separate earth system for the solar and the building with two earth stakes. Following Deye recommendation the grid incoming neutral is connected to the UPS load neutral so they are the same potential. The earth for the load at the inverter is not connected but the earth for the grid is connected to the building earth. Again following Deye diagram. The voltages on the sheet are measured to the building earth so there is a difference of 9-10 V between the transformer earth and the building earth. But of course this is just measuring with a Fluke. I have not checked earth impedances because I don't have a tester here but I suspect that the earth at the transformer may not be that good and certainly the incoming earth wire will have a significant potential to true earth. I don't believe that this setup is causing any issues. Our electrical systems here are not like Western Countries even though the regulations are similar. Most domestic 230 V supplies will be either floating (3 x 230 V transformers in delta configuration) or not properly grounded on one side. This is why all domestic systems use 2 pole breakers. Industrial installations are more normal.
 
Incoming supply is 13.2 kV single phase with earth wire. We have a dedicated 13.2 kV/ 230 V 70 kVA pole mounted transformer. One side of the primary and secondary windings are strapped together and connected to the earth wire and also an earth stake at the transformer. There is then a separate earth system for the solar and the building with two earth stakes. Following Deye recommendation the grid incoming neutral is connected to the UPS load neutral so they are the same potential. The earth for the load at the inverter is not connected but the earth for the grid is connected to the building earth. Again following Deye diagram. The voltages on the sheet are measured to the building earth so there is a difference of 9-10 V between the transformer earth and the building earth. But of course this is just measuring with a Fluke. I have not checked earth impedances because I don't have a tester here but I suspect that the earth at the transformer may not be that good and certainly the incoming earth wire will have a significant potential to true earth. I don't believe that this setup is causing any issues. Our electrical systems here are not like Western Countries even though the regulations are similar. Most domestic 230 V supplies will be either floating (3 x 230 V transformers in delta configuration) or not properly grounded on one side. This is why all domestic systems use 2 pole breakers. Industrial installations are more normal.


The same old ... common neutral VS separated neutral problem.
On page -18- (pdf page 20) 3.9 Wiring System for Inverter there is 2 diagram.
1, separated neutral
2, common neutral
So you made the 2, common neutral version ?
Both works but I like the separated neutral version better. The protected load is more ... separated :)
Also if you check the parallel connection on page -20- (on pdf 22) ... there is also the separated neutral version (UPS neutral not connected to Grid neutral).

So the transformer gives you a Line and a PEN. Transformer has ground rod, your main panel has ground rod, solar panels have a ground rod (last two how far from each other ? Less than 20m here it is recommended to connect them together directly)
Then L+PEN comes to main panel where you connect PEN to house ground rod and separate it as earth, and neutral.
Earth is common, all connected here
Lgrid + Ngrid goes into inverters AC in (maybe to unprotected load panel if you have too)
From inverter AC out Lout and Nout goes to your house (protected load panel), and Nout is also connected to your Ngrid (makes it not separated).

But what I do not like is
  • grid connected: there is 9-10V difference between L-N and L-E. That is strange. N and E are connected together on your main panel.
    There should be 0V.
  • off-grid, grid side: 112V between L-N ? L-E 213V ? That seems to be wrong.
    That should be 0V.
  • off-grid, load side: N-E 9V? If inverter goes into off-grid mode a relay connects N-E together in it. But in common neutral not, since Eout is not even connected ? Hmmmm, strange.
Isolated means that the transformer is shut down or the main incoming L+N in your house side ?
You know others maybe do not care on the line and feedback power even on a blackout. And that sine wave if V reaches the inverter set 180-200V limit is enough for the inverter to think it is grid connected, can sell to grid.
And that can cause an AC overload (you are not a power plant :) )
There are simple relays that if the incoming V is low then auto disconnects you from grid. (if the Grid Low Volt on your inverter is not enough ... first try to raise it)
 
E to N connection is at the transformer but I don't know how good the E is and may have a voltage to true earth. With this system Deye system diagram connects the grid and load N together and does not earth the load earth terminal. There is one ground system for the main panel and the solar and all other equipment. This is not connected to the neutral in the main panel because the main EN connection is at the transformer. Depends on which electrical standards you follow whether multiple E to N connections are allowed but multiple E to N connections can result in circulating currents if the Es are not all good that effect protection particularly if you have residual current devices between two. I decided to have only one N to E connection at the transformer hence the voltage difference between the building earth and the building neutral. The N of the UPS load is connected directly to the N to E connection at the transformer so whatever you switch on or off at the main panel including the main breaker the UPS load always has the N to E connection and this is the same whether the inverter is grid tied or islanded.

The 9-10 V difference (remember this is just a value on the Fluke with no current) is the difference between the two earths.

I think that the spurious V measurements to earth may be because the disconnection in the Deye is single pole on the L only. You need the detailed wiring diagram to confirm. Interestingly when I first got them and was engineering the connections I checked whether the neutrals were connected together and it appeared that they were not which at the time lead me to believe that the disconnections were two pole which would be contradictory.

In any case the 112 V between L and N on the Master grid when it is not connected to anything is a big problem and can only be a back feed because it is not isolating correctly (actually the AC on light is illuminated). The slave correctly has 0 V. It's a big problem. I did check for DC as well but did not measure anything.

The moment that I found out yesterday that the master inverter was not disconnecting properly from the grid I installed a separate islanding protection using an under voltage relay and contactors that I had. To do it properly and not have spurious disconnections from voltage dips (one of which happened just now) you need a more sophisticated relay like the ComAp Mains-Pro and DC contactors. Spurious disconnections however are not really a problem because they are very short term voltage dips and the UPS load does not see them and the non UPS load connected to the grid may just roll through them.
 
Just one clarification. The Deye does not connect the N to E internally in off grid but gives an output for you to do it. I don't like that because it relies on a relay operating correctly to stop the system floating and causes transients. Hence I permanently connected the UPS N to E by using the incoming N from the transformer which is hard wired to earth.
 
E to N connection is at the transformer but I don't know how good the E is and may have a voltage to true earth.
You get from transformer:
- 2 wire : L + PEN
- 3 wire : L + N + PE
?

If only 2 wire then the PEN you need to ground (connect to local ground rod) on your house panel too !
If 3 wire then the PE should be connected to local ground rod.

With this system Deye system diagram connects the grid and load N together and does not earth the load earth terminal.
Because that is a TS (separated) system ... there 3 wires come in: L, N, PE
In this case you do not allowed to connect N and PE together ... since they are already separated.

There is one ground system for the main panel and the solar and all other equipment. This is not connected to the neutral in the main panel because the main EN connection is at the transformer. Depends on which electrical standards you follow whether multiple E to N connections are allowed but multiple E to N connections can result in circulating currents if the Es are not all good that effect protection particularly if you have residual current devices between two. I decided to have only one N to E connection at the transformer hence the voltage difference between the building earth and the building neutral. The N of the UPS load is connected directly to the N to E connection at the transformer so whatever you switch on or off at the main panel including the main breaker the UPS load always has the N to E connection and this is the same whether the inverter is grid tied or islanded.

The 9-10 V difference (remember this is just a value on the Fluke with no current) is the difference between the two earths.
If you ground your Earth on house side too ... there should be none. They are connected like it should be ... like with Equipotential bonding.

I think that the spurious V measurements to earth may be because the disconnection in the Deye is single pole on the L only. You need the detailed wiring diagram to confirm. Interestingly when I first got them and was engineering the connections I checked whether the neutrals were connected together and it appeared that they were not which at the time lead me to believe that the disconnections were two pole which would be contradictory.

In any case the 112 V between L and N on the Master grid when it is not connected to anything is a big problem and can only be a back feed because it is not isolating correctly (actually the AC on light is illuminated). The slave correctly has 0 V. It's a big problem. I did check for DC as well but did not measure anything.
Yeah. Also L-N not the same V as L-E ! That is strange too !
 
Just one clarification. The Deye does not connect the N to E internally in off grid but gives an output for you to do it. I don't like that because it relies on a relay operating correctly to stop the system floating and causes transients. Hence I permanently connected the UPS N to E by using the incoming N from the transformer which is hard wired to earth.
Right. But believe me, external relays are better than inner :)
You can use a good one.
 
Two wires from transformer. Like a TT system. Just don't know how good the earths are. Using 2 pole breakers ensures correct protection on a TT system. I need an earth impedance tester to check which if either earth is good.
 
Two wires from transformer. Like a TT system. Just don't know how good the earths are. Using 2 pole breakers ensures correct protection on a TT system. I need an earth impedance tester to check which if either earth is good.
In a TT system you got a Line, Neutral and Earth from the transformer.
If you got only 2 cables that is Line + PEN (Earth and Neutral together)
That you have to ground locally, and separate to PE + N.
 
I think that the attached are the EU types of earth system.
 

Attachments

  • Earth Systems.JPG
    Earth Systems.JPG
    54 KB · Views: 19
I checked what would happen if I connected the local E to the incoming N and there would be a 4 A circulating current. Obviously one or both of the E connections are not that good most likely the one at the transformer since it is connected to the earth wire of the HT system. It needs this for the fuse protection on the HV line to the transformer to operate.
 
I have two real problems. First the failure of the Master to correctly disconnect from the grid when the grid is lost. I have a temporary fix with another islanding system. Second is the incorrect operation of both MPPT controllers on the Slave which is reducing the solar generation to about 60% because it is holding the voltage down at 130 V. Now if I stop and start the DC to the slave early in the morning it will come back on correctly with 270+ Volts and stay there if the solar radiation is good but if not it drops back to 130 V and will not come up again when the solar radiation increases.
 
Back
Top