Communication Interference in Optimizers

[TheElectrician]

Hey everyone, In one of our projects are using Tigo optimizers with solectria inverters. These optimizers communicate with the inverters via the power lines, so if we put all the dc conductors from all the arrays in a single cable tray, it seems that the communication signals tend to interfere with each other. So the solution provided by the manufacturer was to put all the DC conductors respective to their inverters in separate conduits with good amount of space between each conduit. The problem is we don't have enough space on the roof to use many conductors. The solutions we came up with are:

1. Twist pair the conductors (+ & -) in a flexible conduit for each strings and put them all in a cable tray (More Labor)
2. Use wireless communication enabled optimizers instead of wired (Expensive)

Did anyone face such an issue here? If so, how did you solve it?

Thanks

 

[synchro]

Hey everyone, In one of our projects are using Tigo optimizers with solectria inverters. These optimizers communicate with the inverters via the power lines, so if we put all the dc conductors from all the arrays in a single cable tray, it seems that the communication signals tend to interfere with each other. So the solution provided by the manufacturer was to put all the DC conductors respective to their inverters in separate conduits with good amount of space between each conduit. The problem is we don't have enough space on the roof to use many conductors.

I've no personal experience with these but here are a few comments for what it's worth.
Are these metallic conduits that the manufacturer is proposing?
Metallic conduit will provide significant attenuation by itself even when more closely spaced, although increased spacing certainly helps also.

The solutions we came up with are:

1. Twist pair the conductors (+ & -) in a flexible conduit for each strings and put them all in a cable tray

Does Tigo couple the PLC into only one side of the DC bus with a transformer or coupling network, as is typical? If so then some of the PLC signal current injected into the one DC conductor may go into the self capacitance of the equipment on the DC side instead of returning on the other DC conductor back to the inverter. This would make the magnitudes of the PLC signal currents through your + & - conductors going to the inverters unequal, and therefore not be a purely differential signal but instead have some common mode component. Twisting pairs of conductors will reduce the coupling of differential-mode signals (+-, -+) from one pair to another, but not of any common mode (++,--) signal that may be present. Still, twisting can only help. Just be aware that it may not be a cure-all for your situation.

 

[BandGap1.1eV]

First I've heard of the cross talk issue. Interesting. What kinds of interference are you seeing? It's only a "heartbeat" sensor from what I understand and when the inverter loses AC power, the heartbeat goes away and the tigo units shutdown. What is happening in your system? Are you getting false arc fault trips due to the interference?

Possible solution would be to switch to the Cloud Connect version of Tigo that uses a wireless mesh network independent of any signal injectors built into the interters.

 

[SceneryDriver]

Can you use a shielded cable? My first thought would be to use VFD cable (no brake pair), but you'd end up with an extra conductor.

Shielded tray cable comes to mind as a possibility:

#12-2C THHN-PVC Shielded Tray Cable

THHN-PVC Shielded Tray Cable. Cut to length - sold by the Foot.

www.wireandcableyourway.com

If you use shielded cable, ground the shield at both ends. In this case, you want to treat it like a run of VFD cable between the drive an motor - you want to keep the interfering noise IN the cable. Grounding the shield at only at one end, like you would for a cable carrying an analog signal, is much less effective at keeping noise IN.


SceneryDriver

 

[TheElectrician]

First I've heard of the cross talk issue. Interesting. What kinds of interference are you seeing? It's only a "heartbeat" sensor from what I understand and when the inverter loses AC power, the heartbeat goes away and the tigo units shutdown. What is happening in your system? Are you getting false arc fault trips due to the interference?

Possible solution would be to switch to the Cloud Connect version of Tigo that uses a wireless mesh network independent of any signal injectors built into the interters.

It is kind of a cross talk issue that makes the units shutdown like you mentioned.

 

[TheElectrician]

I've no personal experience with these but here are a few comments for what it's worth.
Are these metallic conduits that the manufacturer is proposing?
Metallic conduit will provide significant attenuation by itself even when more closely spaced, although increased spacing certainly helps also.



Does Tigo couple the PLC into only one side of the DC bus with a transformer or coupling network, as is typical? If so then some of the PLC signal current injected into the one DC conductor may go into the self capacitance of the equipment on the DC side instead of returning on the other DC conductor back to the inverter. This would make the magnitudes of the PLC signal currents through your + & - conductors going to the inverters unequal, and therefore not be a purely differential signal but instead have some common mode component. Twisting pairs of conductors will reduce the coupling of differential-mode signals (+-, -+) from one pair to another, but not of any common mode (++,--) signal that may be present. Still, twisting can only help. Just be aware that it may not be a cure-all for your situation.

Yes, They are proposing a metal conduit placed apart. Twisting the pair is one of the option we are considering. We are alto thinking of using ferrite cores.

 

[TheElectrician]

Can you use a shielded cable? My first thought would be to use VFD cable (no brake pair), but you'd end up with an extra conductor.

Shielded tray cable comes to mind as a possibility:

#12-2C THHN-PVC Shielded Tray Cable

THHN-PVC Shielded Tray Cable. Cut to length - sold by the Foot.

www.wireandcableyourway.com

If you use shielded cable, ground the shield at both ends. In this case, you want to treat it like a run of VFD cable between the drive an motor - you want to keep the interfering noise IN the cable. Grounding the shield at only at one end, like you would for a cable carrying an analog signal, is much less effective at keeping noise IN.


SceneryDriver

Interesting, I will look into this!

 

[tortuga]

Hey everyone, In one of our projects are using Tigo optimizers with solectria inverters. These optimizers communicate with the inverters via the power lines, so if we put all the dc conductors from all the arrays in a single cable tray, it seems that the communication signals tend to interfere with each other. So the solution provided by the manufacturer was to put all the DC conductors respective to their inverters in separate conduits with good amount of space between each conduit. The problem is we don't have enough space on the roof to use many conductors. The solutions we came up with are:

1. Twist pair the conductors (+ & -) in a flexible conduit for each strings and put them all in a cable tray (More Labor)
2. Use wireless communication enabled optimizers instead of wired (Expensive)

Did anyone face such an issue here? If so, how did you solve it?

Thanks

This comes up often on here now, and I'll just add that the concept of mixing line voltage and data has been tried in all kinds of products since the 70's and there is a long list of dead and abandoned products trying that.
Its basically very error prone and usually abandoned in favor of RS-485/modbus or tcp/ip networking.
Its obviously too late now, but I would use a manufacturer that has a option that allows hard wired data cable to be used.