XW Pro Grid Support limitations?

[400bird]

In my case my contactor is a 3 pole, so I disconnect L1,L2 , and N so I have nothing but ground connecting my off grid system to the grid. Everything I have shares the same ground and a single ground rod. So since the N is disconnected from the grid panel when in off grid mode I made the N-G bond in the off grid panel like I said. Slightly more complex I know but this way the inverter without a doubt is not capable of either drawing or pushing power to the grid.

I think you're creating unneeded problems for yourself.
Your house already has the one neutral ground bond, in the main panel.

Is your off grid panel ever connected to the grid pabel? For example, when the transfer relays inside the XW close? In that case, I believe you have both of your neutral ground bonds connected together.

You can't push power to the grid on neutral and ground. That's just a single (unmetered) wire coming from the utility. If the hot lines are open, it can't push power to the grid. No need to disconnect the neutral.

 

[GXMnow]

I just dug into my XW manual and as I remember it explicitly states that there is no control over the N, switching or disconnecting, furthermore it states it does not connect it to ground and an external bond must be made.

The way the relays in the XW+ and XW-Pro appear to be the same. There is no setting to change how it controls the relays, other than some timing settings when using multiple units and/or and external contactor. The contacts that are open or closed in each operating mode are always the same.

The output neutral is always tied to the AC1 and AC2 input neutral. There is no time they are ever disconnected. The reason it breaks the neutral to the transformer is to eliminate any issue with the voltage of the 2 poles being out of balance and causing excessive current flow. When you are on grid in a split phase system, the utility transformer has it's center tap tied to the neutral. The L1 and L2 voltages should be identical, but they can vary by a small amount due to different loading, or even a winding being off a turn or two. If the transformer in the XW was also connected, and the voltage balance was not the same as the turns ratio (it should be exactly 1:1) then the XW transformer would fight the utility grid to try and balance the two legs. Even a tiny voltage difference could cause a huge amount of current to flow. That is why the neutral center tap of the XW transformer needs to be disconnected.

Many of the cheap split phase inverter/charger units on the market don't even have a connection point for the neutral on the input. I don't like that solution for a few reasons. If there is a balance error between the input L1, L2, and the transformer turns ratio, then the neutral will not be at the grid neutral potential. And ground bonding the output neutral will cause current to flow in the ground. It's not a good design.

 

[hogback]

To make sure I'm getting what you're saying @GXMnow , if you had a G N bond upstream (in a taking energy from the grid sense) of the AC1 terminals, this bond would not exist for the system when powering from the grid? I don't have any grid connection, but instead rely on the N G bond internal to the generator weather the generator is running or not. The generator is the only place in my system where the bond exists, and this usage seems to be suggested in the manual. Does AC2 work differently? I think I'm not understanding what you are saying is disconnected when grid is powering.... let me take a guess then... all Ns (load, ac1, ac2) are always bonded to each other internally in the xw, so a single N G bond on any of these will always suffice, but there is an additional internal relay to the xw that opens when taking power from ac1 or ac2, but this just applies to the xw's transformer, yes?

 

[Chris30559]

The way the relays in the XW+ and XW-Pro appear to be the same. There is no setting to change how it controls the relays, other than some timing settings when using multiple units and/or and external contactor. The contacts that are open or closed in each operating mode are always the same.

The output neutral is always tied to the AC1 and AC2 input neutral. There is no time they are ever disconnected. The reason it breaks the neutral to the transformer is to eliminate any issue with the voltage of the 2 poles being out of balance and causing excessive current flow. When you are on grid in a split phase system, the utility transformer has it's center tap tied to the neutral. The L1 and L2 voltages should be identical, but they can vary by a small amount due to different loading, or even a winding being off a turn or two. If the transformer in the XW was also connected, and the voltage balance was not the same as the turns ratio (it should be exactly 1:1) then the XW transformer would fight the utility grid to try and balance the two legs. Even a tiny voltage difference could cause a huge amount of current to flow. That is why the neutral center tap of the XW transformer needs to be disconnected.

Many of the cheap split phase inverter/charger units on the market don't even have a connection point for the neutral on the input. I don't like that solution for a few reasons. If there is a balance error between the input L1, L2, and the transformer turns ratio, then the neutral will not be at the grid neutral potential. And ground bonding the output neutral will cause current to flow in the ground. It's not a good design.

I think I understand what you're saying now, and it makes complete sense, the neutral break you referred to is between the actual transformer in the inverter and the input either AC1 or AC2 (it doesn't matter since all the N are tied together "I agree"). This bond is crucial to balance and sync up with the grid AC with the transformer inside the inverter and has nothing to do with the N-G bond. I see now you only brought that up to indicate the N taps in the inverter always stay connected to each other and never break, which will allow me to have the single NG bond in the grid panel. I do follow and agree. I understand that would certainly work for my case and to also answer 400bird's comment about the N being unmetered and unable to push power through to the grid, I Agree 100%. I guess the reason I did it the way I did was that I wanted true total isolation from the grid when in off grid mode, in which I am in 98% of the time. I don't truly have an argument as to why. to me using the 3pole contactor and adding the NG break relay was a simple enough solution to achieve this.

 

[GXMnow]

all Ns (load, ac1, ac2) are always bonded to each other internally in the xw, so a single N G bond on any of these will always suffice, but there is an additional internal relay to the xw that opens when taking power from ac1 or ac2, but this just applies to the xw's transformer, yes?

I think I understand what you're saying now, and it makes complete sense, the neutral break you referred to is between the actual transformer in the inverter and the input either AC1 or AC2 (it doesn't matter since all the N are tied together "I agree"). This bond is crucial to balance and sync up with the grid AC with the transformer inside the inverter and has nothing to do with the N-G bond. I see now you only brought that up to indicate the N taps in the inverter always stay connected to each other and never break, which will allow me to have the single NG bond in the grid panel. I do follow and agree.

Yes to both.

This function is required in the way the XW series operates to allow both on and off grid operation without fighting the neutral balance. No matter how the system is operating, only a single Neutral / Ground bond is required and there should never be a second one and there is no need for the neutral to be switched.

This is different to how some other inverters work. Several inverters in the Victron line do have a relay that ties the neutral and ground together when it is running in the off grid mode. But this is for mobile applications like an RV or boat. When they unplug from the grid connection, they lose their N/G bond and the relay then creates it. I have also seen RV transfer switch units that duplicate that function. This is only needed when the system fully disconnects from the grid, including breaking the neutral tie. Unless this is a mobile system, I would just always leave the neutral connected back to the grid and have the one constant bond back at the point of entry.

 

[BentleyJ]

This issue has come up recently and I was also under the impression Schneider uses dynamic neutral bonding via the 5th relay (smaller) attached to the PCB just above the AC connection terminals. On April 10th @hyw17 posted a paragraph from the Conext XW installation manual specifically stating that the inverter does not bond neutral to ground under any operating conditions. Pg69 attached for review.

So who whats to pull the front cover off their Schneider, disconnect the Neutral wires and do the continuity test both with qualified AC input and without? I would volunteer but our entire main panel is powered through the XW+ and really don't want to do all that at this moment.

 

[hogback]

So, if you were charging with just L1 and L2, and never even connected the N, it would charge fine, but you'd lose all 120V circuits on pass through during the charge, or would the internal wizards throw a fit?

 

[400bird]

This issue has come up recently and I was also under the impression Schneider uses dynamic neutral bonding via the 5th relay (smaller) attached to the PCB just above the AC connection terminals. On April 10th @hyw17 posted a paragraph from the Conext XW installation manual specifically stating that the inverter does not bond neutral to ground under any operating conditions. Pg69 attached for review.

Seem pretty cut and dry with these two lines:
"Conext XW+ does not connect the neutral to ground."

"In a system without a generator, or with a generator that does not provide a
grounded neutral, you must make the connection from neutral to ground..."

So who whats to pull the front cover off their Schneider, disconnect the Neutral wires and do the continuity test both with qualified AC input and without? I would volunteer but our entire main panel is powered through the XW+ and really don't want to do all that at this moment.

I don't need to pull the cover. I have AC in and AC out (two hot wires each) and a single Neutral connection that goes back to my main (grid side panel) if the XW opened the connection to the neutral, there's no way my system would function off grid.

 

[400bird]

So, if you were charging with just L1 and L2, and never even connected the N, it would charge fine, but you'd lose all 120V circuits on pass through during the charge, or would the internal wizards throw a fit?

I haven't tested it, probably the second option. It should function fine, but I suspect it is testing for the neutral connection.

 

[BentleyJ]

I have AC in and AC out (two hot wires each) and a single Neutral connection that goes back to my main (grid side panel) if the XW opened the connection to the neutral, there's no way my system would function off grid.

All the neutrals for AC1, AC2 & Load are hard wired together at the terminal block, that never changes. We are only talking about there potentially being a relay that makes or breaks a connection between N-G so unless there was a ground fault somewhere, your system would work normally with or without a N-G connection.

 

[400bird]

All the neutrals for AC1, AC2 & Load are hard wired together at the terminal block, that never changes. We are only talking about there potentially being a relay that makes or breaks a connection between N-G so unless there was a ground fault somewhere, your system would work normally with or without a N-G connection.

Right, my mistake. I think the other portion of my post is still valid. The XW doesn't have the capability to and never will bond ground to neutral.

 

[GXMnow]

So, if you were charging with just L1 and L2, and never even connected the N, it would charge fine, but you'd lose all 120V circuits on pass through during the charge, or would the internal wizards throw a fit?

If you only connected L1 and L2 with no neutral, I would expect the XW-Pro would never qualify the input as good power and it would leave the input contactor open, the neutral connected to the transformer center tap, and it would stay running off grid. And in that case, it would not be able to charge from the grid power.

If the neutral ground bond did hold it stable enough to read as the AC input as "qualified", it may try to connect, but as soon as there is any unbalanced load on the system, it would fault and drop off grid again.

 

[kevinich]

Bringing up an old thread again but any one have any updates on this issue? I never had any issues in the past until they installed a new smartmeter recently. What I found was that no matter what mode I'm in, Grid Support, Grid Sell or Load Shave in all modes when it encounters an imbalance when inverting it 'sells' or 'exports' the energy through one leg and imports energy through the other leg to balance itself out. With the new smartmeter whenever I export this energy to the grid this kwh isn't registered!

Over the course of the month I realized that my energy bill went up 150kwh more than what I was registering as using from the grid. After reviewing all the logs I realized all the imbalance loads that it 'sold' or 'exported' wasn't counted and I end up paying for it.

So this is now a big issue since I literally am throwing away kwh of battery power generated.

Did anyone find a way for the XW to be true grid support (ie. only use the grid when battery is low) and zero export?

Otherwise I'm narrowing it down to:
1. Using an autotransformer (like the Victron Autotransformer) - anyone have any issues? Maybe neutral bonding since it's no longer going through the XW?
2. Have a contactor and only enable grid when battery is running low (but this is risky as there's an added device that needs to monitor battery voltage)
3. Maybe have the above and also have the AC2 (generator) grid hooked up to the actual grid as a backup so if #2 doesn't failover correctly it will automatically use the AC2?

I swear the XW has the worst firmware in the world. Made by engineers that don't even know how it work. Trace Engineering should be ashamed of themselves.

 

[GXMnow]

This sounds more like an electric company issue. You do have a net metering agreement, right? If you don't, that could be a whole different issue. The meter could read the current on each leg separately. If it sees back feed, and you don't have a net metering contract, that could throw up a red flag.

Assuming you have a net metering deal, it should sum the two and import and export between the legs should cancel out. Both my XW-Pro and my Enphase microinverters only push current at 240 volts with a floating neutral. They depend on the grid transformer to produce the neutral while on grid. Right now I am seeing about 3 amps going out on one leg and 2 amps coming in on the other leg. The net result is just 1 amp of export at 120 volts, or about 120 watts going to the grid. My net metering deal allows me to export up to 16 amps at 240 volts. But it could be 8 amps on one leg and 24 amps on the other if a had a massive imbalance of loads in my house. That still totals up to 3,840 watts.

There are a few different smart meters, and even the same model could be running different firmware. Watch the display for a few minutes and you should be able to figure out what each screen is doing. On mine, it has total energy taken from the grid, then a separate display shows energy exported back into the grid. Then it says something like Pid232, not sure what that is. But then it shows the live reading of power. Right now it shows -0.12 which is kilowatts, or -120 watts exporting to the grid. The rest of my solar power is charging the battery bank or running the house. Another screen is a display test, it just turns on every segment on the LCD screen. Watching the meter do a few rounds of all it's functions, and then turning on and off some loads should show you if it is reading properly.

At one time, I did have an idea of getting a second XW-Pro. Wiring both of them as 120 volt only systems. Have one on each leg of my split phase grid. Then I could have each one perfectly zero the current on that leg and there would then be no current going either way on either leg and no neutral current issue with the grid transformer. It does look like that is one of the possible setups in the multi unit configuration. But my power needs have not maxed out a single unit, so it is hard to justify the cost to do it. But some dealers are closing out the XW-Pro for under $1,500 and that makes it tempting. But it would also double the standby power draw.

 

[BKY2003]

Some of the EG4 users have discussed using a chargverter on their battery to eliminate back feed to the grid. https://signaturesolar.com/eg4-chargeverter-gc-48v-100a-battery-charger-5120w-output-240-120v-input/
You can program it to start charging at certain SOC.

 

[GXMnow]

Some of the EG4 users have discussed using a chargverter on their battery to eliminate back feed to the grid.

If the power imbalance is very consistent, yes, you could certainly add a Chargeverter as a load to just one leg. They do work on 120 volts as well. Adjust the charge current to balance the load between the two legs. At 85% efficient (they are over 90% when run at 240 volts), it is a bit wasteful, but it could still be better than being hit with the charges if your local utility frows on back feed on a single leg.

 

[BKY2003]

Also they were discussing connecting the grid straight to the chargverter. That way the grid won”t get back feed. I have a chargverter but I have only tried mine with a generator.

 

[kevinich]

Yes my issue is the fact that I do not have net-metering. That's why this is an issue as the selling just goes nowhere as the meter just discards those exports.

Isn't the chargeverter just a 48v battery charger? So using the chargeverter to consume the power sent by the backfeeding and match the output and send that back to the battery. That will require two chargeverters, one for each leg. Then more programming. Sounds like a roundabout way to fix it. Also don't forget when the energy backfeeds through L1, it acutually balances it by consuming the same amount of energy on L2. A chargeverter will not fix that. I end up still paying for energy that I don't need to consume from the grid.

 

[BKY2003]

I was referring to connecting the grid to the 240v chargverter, and disconnecting AC1 on the XW6848.

 

[GXMnow]

Yes my issue is the fact that I do not have net-metering. That's why this is an issue as the selling just goes nowhere as the meter just discards those exports.

If you don't have net metering, why are you pushing current to the main panel? How much load is in the main panel you want to cover? Can you move loads to the output side of the XW-Pro? It handles loads on the output side with no issues. To cover loads on the grid AC1 side, you need to add the Watt-Node power meter and program the Schneider system for "Zero Sell". You may still have some imbalance with one leg importing while the other is exporting.

The idea with the Chargeverter was to only connect one of them to the leg that is exporting power. Adjust the charge current to consume the export and push it back to the batteries. Yes, you lose about 15% of the energy, but that is better than being billed for energy you send out. In my case, I would need to dial it for just about 800 watts of charge power to balance the legs when the house is at light base loads. The 5,000 watt chargeverter is a bit overkill for that function, but then I would still have it if I need to use on my generator.

 

[BKY2003]

I am thinking of kicking my net metering agreement to the curb. I was under the impression that the wattnode could eliminate back feeding?

 

[GXMnow]

I am thinking of kicking my net metering agreement to the curb. I was under the impression that the wattnode could eliminate back feeding?

Most meters, even the "Smart" ones typically sum the power of both legs. That is what mine is doing. I know I have an imbalance, but the meter does not care. I have had it exporting over 5 amps on one leg while importing about 4 amps on the other leg. The utility power meter live display just shows it as -0.12 KW of export.

A possible bigger issue is that even if the system is perfectly programmed for zero export, it will, from time to time have some export. And the smart meters will see this. This happens due to a lag in the control loop. If you have a large load back in the main panel, the Watt Node meter reports it to the Gateway or Insight box. That box commands the XW-Pro inverter to output the desired power to cover that load. This takes a few seconds to happen. During that time, the power came in from the grid, not a big deal. But when that load shuts off, we have the opposite delay. The XW-Pro is still pushing the commanded power to the grid AC1 side. This becomes an export to the grid. Once again, the Watt Node will see this export, send the data to the Insight (or gateway), then it has to send a command to the XW-Pro to reduce the export power to again match the reduced load in the main panel. The update speed on the Watt Node setup does have some adjustment range. Too fast and it could over shoot and swing up and back a bit. I am not using a Watt Node, I have a PLC doing this function. Mine adjusts every 5 seconds, but it can take 2 or 3 adjustments to hit zero. So, it does allow some export for up to 15 to 20 seconds when a large load turns off. The only way to avoid this is putting all the loads on the output side of the inverter. Any load there is instantly sensed by the XW-Pro and it adjusts to cover the load in a few cycles, far less than 1 second.

If you are trying to power loads in the main panel, you really should have a net metering contract. I know some utilities are a rip off on them, so read the fine print. Otherwise, program the system to always have a small draw so you don't get sold out by a smart meter.

 

[BKY2003]

Most meters, even the "Smart" ones typically sum the power of both legs. That is what mine is doing. I know I have an imbalance, but the meter does not care. I have had it exporting over 5 amps on one leg while importing about 4 amps on the other leg. The utility power meter live display just shows it as -0.12 KW of export.

A possible bigger issue is that even if the system is perfectly programmed for zero export, it will, from time to time have some export. And the smart meters will see this. This happens due to a lag in the control loop. If you have a large load back in the main panel, the Watt Node meter reports it to the Gateway or Insight box. That box commands the XW-Pro inverter to output the desired power to cover that load. This takes a few seconds to happen. During that time, the power came in from the grid, not a big deal. But when that load shuts off, we have the opposite delay. The XW-Pro is still pushing the commanded power to the grid AC1 side. This becomes an export to the grid. Once again, the Watt Node will see this export, send the data to the Insight (or gateway), then it has to send a command to the XW-Pro to reduce the export power to again match the reduced load in the main panel. The update speed on the Watt Node setup does have some adjustment range. Too fast and it could over shoot and swing up and back a bit. I am not using a Watt Node, I have a PLC doing this function. Mine adjusts every 5 seconds, but it can take 2 or 3 adjustments to hit zero. So, it does allow some export for up to 15 to 20 seconds when a large load turns off. The only way to avoid this is putting all the loads on the output side of the inverter. Any load there is instantly sensed by the XW-Pro and it adjusts to cover the load in a few cycles, far less than 1 second.

If you are trying to power loads in the main panel, you really should have a net metering contract. I know some utilities are a rip off on them, so read the fine print. Otherwise, program the system to always have a small draw so you don't get sold out by a smart meter.

Thank you for the reply, that was informative.

 

[kevinich]

If you don't have net metering, why are you pushing current to the main panel? How much load is in the main panel you want to cover? Can you move loads to the output side of the XW-Pro? It handles loads on the output side with no issues. To cover loads on the grid AC1 side, you need to add the Watt-Node power meter and program the Schneider system for "Zero Sell". You may still have some imbalance with one leg importing while the other is exporting.

I do not push anything to the main panel. All my loads are on the XW output. But when the inverter converts the energy from the battery it wants to perform a balanced load. So instead of inverting everything on L1, it distributes an EQUAL amount on L1+L2. But since there's no load on L2, it exports the energy from L2 and then compensate the energy by bringing the energy that it exported on L2 by bringing it from the grid on L1.

Example when it sees a L2 delivering a -500W and a L1 delivering 500W, the net is zero. But for the electric company it sees that -500W as nothing, yet it sees 500W used. All in all I end up paying for 500W, but lose on the -500W. So in total 1000W wasted.

On a net-metering meter it will register the export correctly, but with no net-metering this energy is lost. That's why you're not seeing anything wrong with your setup since your meter registers the imports and exports correctly. My meter seems to only register imports and disregards the exports.

 

[GXMnow]

I do not push anything to the main panel. All my loads are on the XW output. But when the inverter converts the energy from the battery it wants to perform a balanced load. So instead of inverting everything on L1, it distributes an EQUAL amount on L1+L2. But since there's no load on L2, it exports the energy from L2 and then compensate the energy by bringing the energy that it exported on L2 by bringing it from the grid on L1.

Example when it sees a L2 delivering a -500W and a L1 delivering 500W, the net is zero. But for the electric company it sees that -500W as nothing, yet it sees 500W used. All in all I end up paying for 500W, but lose on the -500W. So in total 1000W wasted.

On a net-metering meter it will register the export correctly, but with no net-metering this energy is lost. That's why you're not seeing anything wrong with your setup since your meter registers the imports and exports correctly. My meter seems to only register imports and disregards the exports.

Is there any way you could balance up your loads between the two legs? I reduced my balance issue by swapping a few things between L1 and L2.

If everything you have is just 120 volt and all on one leg, you could rewire the XW-Pro for 120 volt only mode. Then it would supply all the loads and have zero out the 120 volt input side. This wires the two transformer secondary windings in parallel and gives you something like 56 amps on a single 120 volt leg.

Going with a center tapped neutral forming transformer would convert all of your 120 volt loas into a 240 volt load, and should give about the same result, while still giving you the option of having some 240 volt loads, but at the cost of some efficiency. Also your new formed neutral can't be ground bonded. No matter how perfect the transformer windings may be, there is very likely going to be a slight voltage difference between the grid neutral and your autotransformer created neutral. It should not be enough to cause any issues, but if you tried to ground bond it, it could cause a lot of current to flow in the ground circuit back to the utility grid transformer. This is why the XW-Pro has a relay to lift it's transformer neutral center tap when it is running on grid.