EG4 18KPV - What not to do in parallel

[ksmithaz1]

I am. He used an RJ45 cable in accordance with the installation manual published at the time. Had he been working to the Lux manual or the current EG4 manual, I would say it's his fault, but it would still bring into question the suitability of this unit for its intended use with this failure mode and place some blame on EG4/Lux.

So . . . Looking at the installation manual below we see. . .

"Please contact your inverter supplier for more detailed guidance on paralleling a system."

*I* did exactly this, because frankly the instructions were not clear. I even bitched about this experience in another thread! That is where I found out the recommended switch settings to use for a dual unit setup, and was told you don't need a full loop (two cables) with only two inverters, and verified it needed to be a straight thru cable. I made notes in my manual to that effect.

So your trying to convince me the "professional" installer followed the manual instructions, called support and they told him to grab some random RJ-45 cable, because the pinout is irrelevant, and hook it up to the inverters? This was not my experience. I stand by my statement the damage was 100% the responsibility of the "professional installer" that clearly did not use an RJ-45 cable in accordance with the installation manual as you state, since the manual clearly states to contact someone for guidance before you parallel a system.

I am not defending the engineering of the signaling on the connector, it's a problem that should be addressed in hardware/firmware moving forward, but the installer clearly did not follow the instructions in the manual as you state. I'm really struggling with people saying you ought to be able to hook something up any way you feel like / improperly and it should work as advertised anyway. I do not think this is realistic.

 

[42OhmsPA]

So . . . Looking at the installation manual below we see. . .

"Please contact your inverter supplier for more detailed guidance on paralleling a system."

*I* did exactly this, because frankly the instructions were not clear. I even bitched about this experience in another thread! That is where I found out the recommended switch settings to use for a dual unit setup, and was told you don't need a full loop (two cables) with only two inverters, and verified it needed to be a straight thru cable. I made notes in my manual to that effect.

So your trying to convince me the "professional" installer followed the manual instructions, called support and they told him to grab some random RJ-45 cable, because the pinout is irrelevant, and hook it up to the inverters? This was not my experience. I stand by my statement the damage was 100% the responsibility of the "professional installer" that clearly did not use an RJ-45 cable in accordance with the installation manual as you state, since the manual clearly states to contact someone for guidance before you parallel a system.

I am not defending the engineering of the signaling on the connector, it's a problem that should be addressed in hardware/firmware moving forward, but the installer clearly did not follow the instructions in the manual as you state. I'm really struggling with people saying you ought to be able to hook something up any way you feel like / improperly and it should work as advertised anyway. I do not think this is realistic.



View attachment 219698

Why was the line circled below in the Lux manual omitted from the old version of the EG4 manual?

That's a very important question, in my opinion.

 

[automatikdonn]

So . . . Looking at the installation manual below we see. . .

"Please contact your inverter supplier for more detailed guidance on paralleling a system."

*I* did exactly this, because frankly the instructions were not clear. I even bitched about this experience in another thread! That is where I found out the recommended switch settings to use for a dual unit setup, and was told you don't need a full loop (two cables) with only two inverters, and verified it needed to be a straight thru cable. I made notes in my manual to that effect.

So your trying to convince me the "professional" installer followed the manual instructions, called support and they told him to grab some random RJ-45 cable, because the pinout is irrelevant, and hook it up to the inverters? This was not my experience. I stand by my statement the damage was 100% the responsibility of the "professional installer" that clearly did not use an RJ-45 cable in accordance with the installation manual as you state, since the manual clearly states to contact someone for guidance before you parallel a system.

I am not defending the engineering of the signaling on the connector, it's a problem that should be addressed in hardware/firmware moving forward, but the installer clearly did not follow the instructions in the manual as you state. I'm really struggling with people saying you ought to be able to hook something up any way you feel like / improperly and it should work as advertised anyway. I do not think this is realistic.



View attachment 219698

Then I guess we can at this point agree to disagree. I think it's irrelevant.

UL1741 and IEEE1547 say what they say. One of those items on that list is the inverter must disconnect (shutdown) when voltage or phase is out of range for the US grid. It doesn't say anything about comms cables or wired correctly. Nowhere in that spec does it say it's ok if the installer made a mistake that the inverter can send 400v to the grid. The unintentional islanding section specifically calls out human error.

I don't know who the installer is, so I honestly couldn't tell you where they are. That would be a question for @canadianintruder

 

[Dave911]

CATx cables get used for Ethernet, with transformer-isolating magnetics in the jack and layers of error detection, correction, retry in software.

They get used with lower level hardware and maybe firmware/software but lacking the higher levels, e.g. for Philips ProfiNet, which can be used for keep-alive signal and shutdown on fault. I dealt with EMI/EMC failures where kV noise injected on power lines didn't affect other communications but triggered shutdown. Higher quality shielded cables were required for the longer runs.

I've also seen these CATx RJ-45 cables used between boxes where there is no transformer isolation. That was the case for a KVM extender, which showed image distortion and shut down during bulk current injection tests, a clamp transformer which injects modulated 10V common mode noise.

SMA uses RJ-45 cables for both their bus between Sunny Island battery inverters (similar communications and sync signals), also for communications from SI to older Sunny Boy battery inverters, switching between UL-1741 on-grid and "island" off-grid modes, and to monitoring hardware. Those are not transformer isolated, and both (I think) run 1200 baud. Newer SMA models use "SpeedWire" which is Ethernet so isolated.

I use Profinet a lot with Siemens controllers and drives. They can run safety signals across standard Ethernet. Estops, machine coordination, light curtains, and safety switch status (redundant, dual channel), Safe Torque off and safe shutdown for drives can all be run on Ethernet. So Ethernet can be very robust. But if comms arent perfect everything will shutdown. Perhaps Luxpower should have spent a few bucks and done it right the first time and put it all on Ethernet instead of just using cheap Ethernet cables.
Its not like the inverters are cheap.

 

[Tinmanbt]

I don't know if this has been asked, but is there an external device for overvoltage protection that could be installed next to the inverter and used to prevent the 400+ volts from being sent down the line? This wouldn't negate the need for the inverters to detect and handle this serious issue, but could be at least an additional safety item, kind of like fuses? I'm not an expert so I don't know if something like this is possible.

 

[Dave911]

If an inverter is prone to that issue, it should simply be avoided. I havent heard if anything was done about this problem, but now that its known, its pretty easy to avoid.
There are voltage sensitive devices which can be used to trip a breaker. But that doesnt mean you wont have damaged appliances and electronic devices.

 

[k490]

I use Profinet a lot with Siemens controllers and drives. They can run safety signals across standard Ethernet. Estops, machine coordination, light curtains, and safety switch status (redundant, dual channel), Safe Torque off and safe shutdown for drives can all be run on Ethernet. So Ethernet can be very robust. But if comms arent perfect everything will shutdown. Perhaps Luxpower should have spent a few bucks and done it right the first time and put it all on Ethernet instead of just using cheap Ethernet cables.
Its not like the inverters are cheap.

Isn't CANBUS the industry standard with inverters I don't think they will change over to ethernet which is 802.3 IEEE standard communication protocol. Seems to me that any communication errors should cause a fault shut down both inverters I suppose this cross cable got past their checks. Good reason to always check voltages before flipping breakers.