I am currently working on a Solarplant project and I have an issue about inverter's MODBUS rtu comms.
When the inverters on low production of energy there is no problems with reading the meas values.
Therefore when the sun's up and our inverters at the high production my comms goes to error state.
I checked my master device config and the slaves. I am pretty sure they are correct.
My wiring is okay as it should be.(At least I am aware of.)
My question is if the AC cables causing an interference on my data line or something else ?
If it is like that what is the solution?
Is there anyone who faced the same situation?
Thanks for response!!

 

RS-485 is quite robust and has excellent noise immunity, but only if you use the proper RS-485 cabling and make the proper connections. AC noise can certainly cause issues if all wiring recommendations are not properly followed. This is especially true around sources of large electrical interference, such as variable frequency drives (VFD's).

Confirm that you are using proper RS-485 cabling and connections, meaning the following:

• The cabling should be 1.5 or 2 twisted pair, shielded cabling (1.5 means 1 twisted pair plus 1 single conductor).
• 1 twisted pair should be used for the RS-485 data signals and 1 conductor (or 1 twisted pair) for the RS-485 ground reference.
• The shield drain wire should be connected to earth ground at one end of the cabling only.
• The cabling should have a characteristic impedance of 100 to 140 ohm (the typical is 120 ohms).
• The wire gauge should be 24 AWG to 16 AWG.
• The capacitance (wire to wire) should be less than 30pF/ft (~100pF/m).
• The capacitance (wire to shield) should be less than 60pF/ft (~200pF/m).

To review the wiring connections:

• RS-485 + and RS-485 - signals should be wired using a single twisted pair in the cabling.
• All devices' RS-485 ground reference should be connected together using a single conductor (or a single twisted pair) in the cabling.
• The shield drain wire should be earth grounded at one end only, typically done at the front end (i.e. at or near the Modbus master device).

 

RS-485 is quite robust and has excellent noise immunity, but only if you use the proper RS-485 cabling and make the proper connections. AC noise can certainly cause issues if all wiring recommendations are not properly followed. This is especially true around sources of large electrical interference, such as variable frequency drives (VFD's).

Confirm that you are using proper RS-485 cabling and connections, meaning the following:

• The cabling should be 1.5 or 2 twisted pair, shielded cabling (1.5 means 1 twisted pair plus 1 single conductor).
• 1 twisted pair should be used for the RS-485 data signals and 1 conductor (or 1 twisted pair) for the RS-485 ground reference.
• The shield drain wire should be connected to earth ground at one end of the cabling only.
• The cabling should have a characteristic impedance of 100 to 140 ohm (the typical is 120 ohms).
• The wire gauge should be 24 AWG to 16 AWG.
• The capacitance (wire to wire) should be less than 30pF/ft (~100pF/m).
• The capacitance (wire to shield) should be less than 60pF/ft (~200pF/m).

To review the wiring connections:

• RS-485 + and RS-485 - signals should be wired using a single twisted pair in the cabling.
• All devices' RS-485 ground reference should be connected together using a single conductor (or a single twisted pair) in the cabling.
• The shield drain wire should be earth grounded at one end only, typically done at the front end (i.e. at or near the Modbus master device).

First of all thanks for answer.I really appreciated.


The image i sent you below refers the situation. I quote this from user manual of the ınverter.
As I said before there is no comm problem when the all 53 inverters producing at some amount of energy. Only they are producing more than 30% of their capacity i lost all my comms.(The RS cable and the AC cables are in same box)

Sincerely

 

The statement that RS-485 communication wires should be routed separately from input wires, output wires, power wires, etc. is a common disclaimer in documentation for devices that generate electromagnetic interference (EMI). To add to this, the worst scenario is when the RS-485 cabling is run right next to the other wiring. If at all possible, separate the RS-485 cabling from other wiring as much as possible, and if it must cross other wiring, try to have it cross at a 90 degree angle to reduce interference.

That being said, even if the RS-485 cabling is run right next to AC cabling or other EMI producing cabling, following my previous post's recommendations for proper wiring and shielding of the cabling will help significantly in reducing the effects of EMI induced onto the RS-485 communication signals.

 

The statement that RS-485 communication wires should be routed separately from input wires, output wires, power wires, etc. is a common disclaimer in documentation for devices that generate electromagnetic interference (EMI). To add to this, the worst scenario is when the RS-485 cabling is run right next to the other wiring. If at all possible, separate the RS-485 cabling from other wiring as much as possible, and if it must cross other wiring, try to have it cross at a 90 degree angle to reduce interference.

That being said, even if the RS-485 cabling is run right next to AC cabling or other EMI producing cabling, following my previous post's recommendations for proper wiring and shielding of the cabling will help significantly in reducing the effects of EMI induced onto the RS-485 communication signals.

Thanks for all the knowledge.
I will be applying all the instruction.
Have a nice day !

 

One other thing.

I know you're currently focusing on noise, but another, more common, cause of communication issues, such as the issue you're seeing, is not connecting the devices' RS-485 ground references properly. In other words, if all of the RS-485 devices do not share a common 0V ground for their RS-485 interfaces, there can be communication issues. Local ground potentials can fluctuate significantly, especially as equipment draws or produces more and more power. The fluctuations in local grounds can cause RS-485 communication issues, and in extreme cases, can actually damage RS-485 devices.

If you would like further recommendations, please provide a wiring diagram and the manufacturers and models of the equipment involved.

 

One other thing.

I know you're currently focusing on noise, but another, more common, cause of communication issues, such as the issue you're seeing, is not connecting the devices' RS-485 ground references properly. In other words, if all of the RS-485 devices do not share a common 0V ground for their RS-485 interfaces, there can be communication issues. Local ground potentials can fluctuate significantly, especially as equipment draws or produces more and more power. The fluctuations in local grounds can cause RS-485 communication issues, and in extreme cases, can actually damage RS-485 devices.

If you would like further recommendations, please provide a wiring diagram and the manufacturers and models of the equipment involved.

I am currently working on Kehua SPI 100~125K-B inverters and as I searched on it's manual,inverter doesn't have a common ground for RS-485 comms. It has only 2 RS ports for jumping from one to another. The image below shows us the wiring diagram and the RS ports that I was connecting for the RS comms. First I thought I may need a termination resistor to clear comm. But then I realized my connection with the inverters doesn't goes error state when they are working on low state. They are only goes errors state when they are producing a lot. As you can see the RS ports and the AC wiring too close to each other.That made me think about noise.

 

I always use Lapp unitronic bus LD cable. 120ohm impedance corrected cable, specially suited for RS-485.
Can be ordered by the meter at for example Conrad
https://www.conrad.nl/nl/p/lapp-217...-1-x-2-x-0-22-mm-violet-per-meter-600095.html

 

I always use Lapp unitronic bus LD cable. 120ohm impedance corrected cable, specially suited for RS-485.
Can be ordered by the meter at for example Conrad
https://www.conrad.nl/nl/p/lapp-217...-1-x-2-x-0-22-mm-violet-per-meter-600095.html

Thanks for the advice. I will be remembering this for further jobs.

 

From reviewing the manual for the inverter, and the fact that it only exposes the RS-485 + and - signals, I would suspect it is a non-isolated RS-485 device. This probably means the inverter uses it's "internal grounding terminal" as the RS-485 0V reference. Therefore, how you have grounded the inverter plays an important role in the RS-485 communications.

The other important aspect is the RS-485 electrical interface of the Modbus master device and whether it exposes an RS-485 reference terminal or if it's also non-isolated device and uses its power supply or earth ground connection as the RS-485 0V reference.

The goal is to make all devices' RS-485 0V references common with one another, whether that means directly connecting them together or indirectly connecting them by locally earth grounding each device to the building's ground network.

In order to provide more specific recommendations, please respond with the manufacturer and model of the Modbus master device and details on the cabling you're using and how you currently have it connected to the devices (specifically how many conductors, whether it is twisted pair, whether it is shielded, and where all the wires are landed).

Regarding termination, the inverter's manual implies that it has built-in termination resistors. These are enabled by SW2 on the communication board. The manual recommends turning SW2 ON for the first and last inverter on the communication link. However, this really depends on the physical topology of the network, as termination should only be applied to the very first device and very last device in the RS-485 daisy-chain. If inverters are at those locations, then as the manual states, you would simply need to turn SW2 to the ON position on those two inverters. But if another device, such as the Modbus master device is at one of the far ends, you would need to apply termination on that device - whether that be installing a resistor (typically 120 ohms) or enabling built-in termination on the device.

As a disclaimer, although many manufacturers unconditionally state that termination should be used, it is only really necessary when using long cable runs and/or high baud rates. The point of termination is to absorb signal reflections that bounce back and forth on the cabling. For short cables and low baud rates, these reflections dampen out very quickly on their own and will not affect the sampling of the signal. For example, at 9600 baud (the default baud rate for the inverter) termination is not necessary, as the maximum possible length supported by RS-485 is shorter than the length where termination would be required. Additionally, in some instances, applying termination can actually cause more issues than it solves. Not only does it apply additional load to the network (and so voltage swings are smaller, reducing the distance the signal can travel and making it more susceptible to noise), but it also may reduce the differential idle voltage (i.e. the voltage difference between the + and - signals when no device is transmitting) into the undefined region between -200mV and +200mV and may cause some RS-485 devices to be unable to communicate. For this reason, it is important to consider whether biasing (i.e. adding a resistor from the + signal to a positive voltage, such as 5V, and adding a resistor from the - signal to ground) is also required when adding or enabling termination.