Bonding multiple Deye 12kW inverters in Off-Grid Mode

[silverstone]

Why so many ground (PE) wires?
Only one is required per piece of equipment.
It sounds like you have a lot of ground loops in your system.

It's not a ground loop unless neutral-earth is bonded in multiple points. As I said, normally PE is NOT supposed to nor conducting any current.

And it doesn't change much if I have 1, 2, 3 or 4 PE wires on the same equipment. All it does is reducing the resistance of the connection. And provide some kind of redundancy if e.g. one connection is not properly making contact.

The fact that the Neutral current is being (partially) diverted through the PE wire is due to MULTIPLE BONDING points (Neutral-Earth), NOT the multiple PE wires. Although multiple PE wires might increase the amount (in %/A) of diverted current since the resistance of the PE connection is lower compared to the Neutral, as said before.

Why multiple PE wires ?
- 1 on Grid Terminal Block (as according to Manual)
- 1 on Load Terminal Block (currently disconnected while trying to see if this was causing the multiple Neutral-Earth issue, which caused the RCD to trip)
- 1 on Generator Terminal Block (currently disconnected while trying to see if this was causing the multiple Neutral-Earth issue, which caused the RCD to trip)
- 1 on Frame/Chassis (as according to Manual)

In the latest version of the Manual they don't show this internal Neutral-Earth bonding contactor anymore in the Schematic, so not sure what is going on.


I just hope that the modifications regarding RCD and Breakers will fix the Issue. I'm not a fan on reworking 100 times on the same panel. Just removing the PERMANENT bonding jumper was an absolute nightmare.

The EMC filter might also be causing sporadic trips. Although, while working with the PERMANENT BOND a few days ago (between RCD and EMC filter), that seemed to work correctly.

 

[timselectric]

It's not a ground loop unless neutral-earth is bonded in multiple points.

A ground loop is multiple ground paths. (Creating a loop in the grounding system) which creates noise on the system.
It has nothing to do with the N/G bond. Although multiple N/G bonds can also create a ground loop.
We used to be very concerned about parallel inverters, each Creating a N/G bond. I've since changed my mind about it. These are multiple sources in parallel.
If everything is wired correctly. There shouldn't be any current flowing between the sources on the ground conductor. (Unless a neutral connection is lost)
For current to flow between them (on the ground) . It would be taking a longer, round about path. Instead of the direct path provided (by the neutral) .
With properly sized conductors, installed as they should be. This possible current should be very minimal. (Less than 1%, if any)

 

[silverstone]

A ground loop is multiple ground paths. (Creating a loop in the grounding system) which creates noise on the system.
It has nothing to do with the N/G bond. Although multiple N/G bonds can also create a ground loop.
We used to be very concerned about parallel inverters, each Creating a N/G bond. I've since changed my mind about it. These are multiple sources in parallel.
If everything is wired correctly. There shouldn't be any current flowing between the sources on the ground conductor. (Unless a neutral connection is lost)
For current to flow between them (on the ground) . It would be taking a longer, round about path. Instead of the direct path provided (by the neutral) .
With properly sized conductors, installed as they should be. This possible current should be very minimal. (Less than 1%, if any)

Well I guess only one way to find out . I got to rewire this part of the panel and put a clamp meter on the earth wires ...

 

[timselectric]

One properly sized earth wire per inverter.
Ran to the main ground bar.

 

[timselectric]

Another question about your system.
If I understand correctly. You have the grid connected, but turned off?
Is the grid neutral also switched off?
Depending on how the rest of the wiring is done. This neutral can also create a ground loop path. Through the grids N/G bond.
If this is also a 4p breaker that includes the neutral. Then this is not a problem.

 

[silverstone]

One properly sized earth wire per inverter.
Ran to the main ground bar.

Well that's against the manual though... At least 2 if not 3.


But then again you could argue that 3-4 PE wires of 6mm2 each is equivalent to approx 18-24 mm2 (actually a bit better since also the inductance gets lower with multiple wires in parallel).

I would be much more afraid of ground loops with regards to control cables, measurement cables, etc, as these are typically more sensitive to noise. And sometimes the shield could carry the fault current, so you also have to watch out for that (sometimes only ground shield on one side to avoid ground loops).

 

[silverstone]

Another question about your system.
If I understand correctly. You have the grid connected, but turned off?
Is the grid neutral also switched off?
Depending on how the rest of the wiring is done. This neutral can also create a ground loop path. Through the grids N/G bond.
If this is also a 4p breaker that includes the neutral. Then this is not a problem.

I connected the grid cable on the inverter side, but since I couldn't find an electrician to finish the job and do the paperwork, the grid connection is floating / "parked" on some 4p breakers or contactors. Thus no potential (phase or neutral) is present on the grid terminals of the inverters. Only the earth wire of the grid cable is connected to the inverter and the ground bar.

I have 4p breakers. And the ac combiner box also includes 4p contactors to do an automatic switch over from inverter to grid, should the inverter fail.

Thus the grid neutral and the inverter neutral are fully separate at all times.

If it wasn't, since the inverter neutral is bonded to earth, then the grid rcd (main panel) would trip, due to (yet another) neutral diverted current.

 

[silverstone]

So, quick update on the situation: I reconfigured the Inverters
- Without a PERMANENT BOND
- With "Signal ISLAND MODE" ENABLED (CHECKED)
- With a Single RCD

Now the system seems to work, at least at low load (~ 1-1.5 kW). Didn't measure any diverted Neutral / Earth current since right now the load is very small for that ...

 

[silverstone]

And thank you all for your help , in particular @timselectric and @houseofancients

 

[liberumyu]

Hi all.

I hope someone reads this thread still.

I’m getting my two Deye 12kW 3P hybrids soon to run in parallel and I have one question.

Can I connect my entire apartment as backup load? I don’t have a dedicated wiring for backup loads so what I would do is run everything as backup load.
If I do that what will happen while I’m having power from the grid?

My current Inverters don’t have a sell to grid option so everything runs trough them. Is it same with Deye inverters for backup load?

 

[timselectric]

Hi all.

I hope someone reads this thread still.

I’m getting my two Deye 12kW 3P hybrids soon to run in parallel and I have one question.

Can I connect my entire apartment as backup load? I don’t have a dedicated wiring for backup loads so what I would do is run everything as backup load.
If I do that what will happen while I’m having power from the grid?

My current Inverters don’t have a sell to grid option so everything runs trough them. Is it same with Deye inverters for backup load?

Yes, assuming that the backup load is large enough for it. (More than likely it will be with two)
The difference between backup load and the regular output. Is that the backup will remain powered at all times.

 

[liberumyu]

Yes, assuming that the backup load is large enough for it. (More than likely it will be with two)
The difference between backup load and the regular output. Is that the backup will remain powered at all times.

Is there a limit of backup power?

Doesn’t say anything about it in manual or specs either? I will run 2 inverters in parallel and so far I have never got more than 4000W per phase on present system, although car charging was suffering from time to time to make room for other appliances.

 

[houseofancients]

Hi all.

I hope someone reads this thread still.

I’m getting my two Deye 12kW 3P hybrids soon to run in parallel and I have one question.

Can I connect my entire apartment as backup load? I don’t have a dedicated wiring for backup loads so what I would do is run everything as backup load.
If I do that what will happen while I’m having power from the grid?

My current Inverters don’t have a sell to grid option so everything runs trough them. Is it same with Deye inverters for backup load?

Yes, this is exactly how I have wired my 12k 3p's.
Do not forget to tick the "anti islanding" , so you will have an automatic ground - neutral binding when your grid fails.
Also , make sure you have enough batteries to sustain 2* 12kw..
Each inverter can pull 240A from the batteries, so make sure you have 500a of BMS/battery amps available

 

[houseofancients]

Is there a limit of backup power?

Doesn’t say anything about it in manual or specs either? I will run 2 inverters in parallel and so far I have never got more than 4000W per phase on present system, although car charging was suffering from time to time to make room for other appliances.

12k

Is there a limit of backup power?

Doesn’t say anything about it in manual or specs either? I will run 2 inverters in parallel and so far I have never got more than 4000W per phase on present system, although car charging was suffering from time to time to make room for other appliances.

If you're not pulling more than 4kw per phase, you could consider leaving 1 inverter as a cold backup..
These inverters will pull between 60 and 90 watts per inverter, so leaving 1 off will leave you with a considerable amount of kwh's per 24h not being wasted

 

[liberumyu]

Yes, this is exactly how I have wired my 12k 3p's.
Do not forget to tick the "anti islanding" , so you will have an automatic ground - neutral binding when your grid fails.
Also , make sure you have enough batteries to sustain 2* 12kw..
Each inverter can pull 240A from the batteries, so make sure you have 500a of BMS/battery amps available

I have 4x 200A, so I guess it should work.

I’ll have some time to prepare everything until my “Peter boards” arrive so my batteries can be connected to inverter.

That’s another question, can inverters run without communication to batteries?

Now I’m controlling charge, float and absorbing with HA so wouldn’t mind doing it in future…

 

[timselectric]

Is there a limit of backup power?

I thought there was. But maybe not, if the manual doesn't specify it.

 

[liberumyu]

12k

If you're not pulling more than 4kw per phase, you could consider leaving 1 inverter as a cold backup..
These inverters will pull between 60 and 90 watts per inverter, so leaving 1 off will leave you with a considerable amount of kwh's per 24h not being wasted

Cold backup? How to do that?

60 watts is better what I have now. Ones I have now pull on idle 120w each and I have 3 of them, so 60 is saving

 

[silverstone]

Hi all.

I hope someone reads this thread still.

I’m getting my two Deye 12kW 3P hybrids soon to run in parallel and I have one question.

Can I connect my entire apartment as backup load? I don’t have a dedicated wiring for backup loads so what I would do is run everything as backup load.
If I do that what will happen while I’m having power from the grid?

My current Inverters don’t have a sell to grid option so everything runs trough them. Is it same with Deye inverters for backup load?

That's what I am doing with 3 of them in parallel . Off grid only, everything is connected to the Load/Backup Port (a big UPS basically).

I had some mishaps yesterday not sure why though ... I was getting lots of F18_AC_over_current_fault_of_hardware, F41_Parallel_system_stop as well as F58_BMS_communication_fault.

I later identified the cause of the F58_BMS_communication_fault (problem sitting between the screen and the chair - I was doing a BMS Interface Board Software Update, so of course that triggered the F58 due to BMS_Err-Stop being Enabled on the Inverter, which lost communication with the BMS and stopped correctly).

And I could imagine F41_Parallel_system_stop being caused by the Inverters Stopping not in the right "Sequence" (all the the same time), maybe one stopped before the other 2.

But F18_AC_over_current_fault_of_hardware is something completely new. Some search online might suggest this is due to load shedding or AC undervoltage voltage cutout being set too low, or even a fault in the AC relay (either the grounding one or the Static Transfer Switch, the latter which I do not use).

During one F18_AC_over_current_fault_of_hardware I swear I heard a spark. What the hell ??

The only possible cause I can imagine are:
- The output filter of the Inverter + the EMC filter that I have installed do not discharge sufficiently in the 30-60 seconds changeover that I have with a set of contactors, therefore when Inverter reconnects, effectively it might be having the wrong phase, thus causing an Inrush Current (doubtful for several reasons: the 30-60s and the fact that such Inrush would have VERY low energy (we might be talking about 20-100 uF, not more).
- There is not enough impedance between the 3 Inverters (only approx. 5m of cables from each to the distribution board where they are connected in parallel)

The F18_AC_over_current_fault_of_hardware error never appeared until now.

I am running my 12kW Heatpump and the whole house on these 3 Inverters. The Heatpump being an Inverter type Heatpump, it's particularly smooth at ramping up the Power (although I'd say that the harmonic content of the current it's a whole other story ... Uhm ... Probably a stupid cheap 3ph Diode Bridge Rectifier ).

THD Values of <2.5% of the Voltage (THD_U) are OK-ish.

But THD Values of ~ 60% at around 2kW on the Current (THD_I) is quite hight ...

 

[timselectric]

If you're not pulling more than 4kw per phase, you could consider leaving 1 inverter as a cold backup..

Would you need an external bypass?
Or will it pass through power, even in the off position?

 

[silverstone]

I have 4x 200A, so I guess it should work.

I’ll have some time to prepare everything until my “Peter boards” arrive so my batteries can be connected to inverter.

That’s another question, can inverters run without communication to batteries?

Now I’m controlling charge, float and absorbing with HA so wouldn’t mind doing it in future…

I suggest using BMS Communication.

I also thought that it wasn't needed but:
- You might hit OVP all the time (I had 5 times per day quite regularly unless I monitored constantly on HomeAssistant)
- You might never go High enough for Long enough and therefore you could suffer from an Unbalance

I also thought of doing the same, but given those caveats (which sometimes BOTH occur almost at the same time, i.e. you reduce float/absorbtion to avoid OVP, but then cause unbalance), and I am somewhat skeptical of doing continuous write cycles on the Inverter Flash Memory (I don't want to wear it out and therefore cause the Inverter to Fail), this Project definitively came in handy and it's working quite well: https://diysolarforum.com/threads/jk-bms-can-with-new-cut-off-charging-logic-open-source.79325/. Although I'd caution about lowering a bit the OVP anyway, there is a bit of delay when things can get rought. BMS communication = Inverter goes "Full Voltage & Full Current" for a few seconds (?) until the CANbus heartbeat missing causes it to trip. It was enough to get 1 of my cells to 3.69 V a couple of times yesterday ...

 

[liberumyu]

That's what I am doing with 3 of them in parallel . Off grid only, everything is connected to the Load/Backup Port (a big UPS basically).

I had some mishaps yesterday not sure why though ... I was getting lots of F18_AC_over_current_fault_of_hardware, F41_Parallel_system_stop as well as F58_BMS_communication_fault.

I later identified the cause of the F58_BMS_communication_fault (problem sitting between the screen and the chair - I was doing a BMS Interface Board Software Update, so of course that triggered the F58 due to BMS_Err-Stop being Enabled on the Inverter, which lost communication with the BMS and stopped correctly).

And I could imagine F41_Parallel_system_stop being caused by the Inverters Stopping not in the right "Sequence" (all the the same time), maybe one stopped before the other 2.

But F18_AC_over_current_fault_of_hardware is something completely new. Some search online might suggest this is due to load shedding or AC undervoltage voltage cutout being set too low, or even a fault in the AC relay (either the grounding one or the Static Transfer Switch, the latter which I do not use).

During one F18_AC_over_current_fault_of_hardware I swear I heard a spark. What the hell ??

The only possible cause I can imagine are:
- The output filter of the Inverter + the EMC filter that I have installed do not discharge sufficiently in the 30-60 seconds changeover that I have with a set of contactors, therefore when Inverter reconnects, effectively it might be having the wrong phase, thus causing an Inrush Current (doubtful for several reasons: the 30-60s and the fact that such Inrush would have VERY low energy (we might be talking about 20-100 uF, not more).
- There is not enough impedance between the 3 Inverters (only approx. 5m of cables from each to the distribution board where they are connected in parallel)

The F18_AC_over_current_fault_of_hardware error never appeared until now.

I am running my 12kW Heatpump and the whole house on these 3 Inverters. The Heatpump being an Inverter type Heatpump, it's particularly smooth at ramping up the Power (although I'd say that the harmonic content of the current it's a whole other story ... Uhm ... Probably a stupid cheap 3ph Diode Bridge Rectifier ).

THD Values of <2.5% of the Voltage (THD_U) are OK-ish.

But THD Values of ~ 60% at around 2kW on the Current (THD_I) is quite hight ...

Amazing. This really made my day

Are you using HA maybe to control inverters? Is that even possible or just you can view data?

 

[houseofancients]

I have 4x 200A, so I guess it should work.

I’ll have some time to prepare everything until my “Peter boards” arrive so my batteries can be connected to inverter.

That’s another question, can inverters run without communication to batteries?

Now I’m controlling charge, float and absorbing with HA so wouldn’t mind doing it in future…

They can, on voltage, but these class of inverters run best with comms, as it will allow you to use the full feature set

 

[houseofancients]

Cold backup? How to do that?

60 watts is better what I have now. Ones I have now pull on idle 120w each and I have 3 of them, so 60 is saving

Simply, keep one off, and don't parallel till you need it

 

[silverstone]

I thought there was. But maybe not, if the manual doesn't specify it.

I think you are confusing with the AUX/Generator/MicroInverter Port. There I think it's only rated for 50% of Rated Power.

But I think there is definitively a lower Rating in terms of Peak (Inrush) Current on the Load Port, at least from what I could understand by looking at the label, although it's always a bit sketchy what is meant by "nominal", "rated", "maximum".

They could make it much more clearer by simply stating "ARMS" or "Apeak". Then there are no more doubts. Is it a "maximum" peak or a "maximum" short-term RMS ?

 

[houseofancients]

I suggest using BMS Communication.

I also thought that it wasn't needed but:
- You might hit OVP all the time (I had 5 times per day quite regularly unless I monitored constantly on HomeAssistant)
- You might never go High enough for Long enough and therefore you could suffer from an Unbalance

I also thought of doing the same, but given those caveats (which sometimes BOTH occur almost at the same time, i.e. you reduce float/absorbtion to avoid OVP, but then cause unbalance), and I am somewhat skeptical of doing continuous write cycles on the Inverter Flash Memory (I don't want to wear it out and therefore cause the Inverter to Fail), this Project definitively came in handy and it's working quite well: https://diysolarforum.com/threads/jk-bms-can-with-new-cut-off-charging-logic-open-source.79325/. Although I'd caution about lowering a bit the OVP anyway, there is a bit of delay when things can get rought. BMS communication = Inverter goes "Full Voltage & Full Current" for a few seconds (?) until the CANbus heartbeat missing causes it to trip. It was enough to get 1 of my cells to 3.69 V a couple of times yesterday ...

That could also be caused by outdated firmware..

Tip to OP, make sure both are running the latest fw ( 1150 main , c042 for lcd) , and both are running the same exact version of it