EG4 6000EX - Double Ground/Neutral

[Zwy]

Hi Zwy,
Thanks for your suggestions. For reasons I don't understand, the Tesla universal mobile charger (UMC) only works when I connect the EGC to the main service panel ground and bond the N-G in the critical loads panel. I previously had a separate ground rod for the off-grid subsystem only (see rev 2 diagram below - I've since revised the wiring to what's shown in rev 3, which is the version you were commenting on). I'm now only using the auxiliary ground rod to ground the panels and it is not connected to the critical loads EGC busbar.

With the rev 2 wiring, the Tesla UMC kept faulting and current limiting to 8A. At one point, I had to use both the separate auxiliary ground rod and the main service panel cold water pipe ground to get the charger to work but it still current limited to 8A.

How far is the distance from critical loads panel to the ground rod?

After disconnecting the PV+ and PV- per Summit's suggestion as shown in the rev 3 diagram, I was then able to charge at up to 25A. However, the Tesla UMC still faulted occasionally and throttled back to 8A before jumping back up to 22A on its own for unknown reasons. I also have a Juicebox EVSE charger but it won't work with any of the ground, NG combinations. It used to work when tied to the grid.

BTW, in the rev 3 diagram, the dashed green lines are not connected (forgot to annotate that).

At this point, I'd begin to determine if the inverter is bonding N-G. Most likely it is, removal of the N-G bonding in the critical loads panel, then attempting to charge off the inverter might reveal that indeed the inverter is bonding N-G.

Or open up the unit and determine if the ground screw is present.

 

[bobphdee]

It is sounding like your system has 3x NG bonds: inside the EG4, CRpanel, Main-panel. Confirm also that AC volt L1-N and L2-N, are the same. Since the EVSE is connected to the CRpanel, I would do some baby step tests:

1. remove CRpanel NG bond, 2. Check EG4 NG bonding (volt then ohm), 3. Check PV+/- G-bond also

My system has only 1 NG bond in the Main-panel when it's tie to the house wiring. In off-grid mode, the CRpanel has NG bond with G still sharing Main-panel G-rod, but no N connection

Confirm also that AC volt L1-N and L2-N, are the same.

L1=119.7, L2=119,8

1. remove CRpanel NG bond, 2. Check EG4 NG bonding (volt then ohm), 3. Check PV+/- G-bond also

With CRpanel NG bond removed:
Inverter off: Vng = 0.097V; Rng = 1.6 Mohm; measured at inverter
Inverter on: Vng = 33V to 38V (fluctuating, sometimes O.L.); measured at inverter without the CRpanel bond, so there is no internal bonding screw?

Tesla EV doesn't charge at all without CRpanel NG bond (faults with loss of ground error on LED)

what's the resistance reading PV-pos & PV-neg with respect to the EG6000 chassis ?

Inverter off:
PV+ to chassis: open with and without the PV wire connected
PV- to chasis: open with and without the PV wire connected

Is it ok to measure PV+/- to chassis while the inverter is ON?

My system has only 1 NG bond in the Main-panel when it's tie to the house wiring. In off-grid mode, the CRpanel has NG bond with G still sharing Main-panel G-rod, but no N connection

 

[bobphdee]

How far is the distance from critical loads panel to the ground rod?

At this point, I'd begin to determine if the inverter is bonding N-G. Most likely it is, removal of the N-G bonding in the critical loads panel, then attempting to charge off the inverter might reveal that indeed the inverter is bonding N-G.

Or open up the unit and determine if the ground screw is present.

How far is the distance from critical loads panel to the ground rod?

There's 107 ft of 6 awg copper wire to the main service panel and another ~30 ft of 6 awg copper to the cold water pipe clamp at the other side of the house.

At this point, I'd begin to determine if the inverter is bonding N-G. Most likely it is, removal of the N-G bonding in the critical loads panel, then attempting to charge off the inverter might reveal that indeed the inverter is bonding N-G.

Vng = 33V to 38V (fluctuating, sometimes O.L.); measured at inverter without the CRpanel bond, so there is no internal bonding screw? The Tesla charger won't charge at all without the critical panel N-G bond. The Juicebox EVSE charger won't charge at all even with the critical panel N-G bond.

Signature Solar told me that the warranty is voided if I open the unit, unless I give them a statement from an electrician saying that the screw needed to be removed.

 

[Zwy]

How far is the distance from critical loads panel to the ground rod?

At this point, I'd begin to determine if the inverter is bonding N-G. Most likely it is, removal of the N-G bonding in the critical loads panel, then attempting to charge off the inverter might reveal that indeed the inverter is bonding N-G.

Or open up the unit and determine if the ground screw is present.


There's 107 ft of 6 awg copper wire to the main service panel and another ~30 ft of 6 awg copper to the cold water pipe clamp at the other side of the house.

That is some distance. Is there any way you can install a ground electrode at the critical loads panel? This would allow N-G bond at the critical loads panel without all the resistance of the current ground circuits.

Vng = 33V to 38V (fluctuating, sometimes O.L.); measured at inverter without the CRpanel bond, so there is no internal bonding screw? The Tesla charger won't charge at all without the critical panel N-G bond. The Juicebox EVSE charger won't charge at all even with the critical panel N-G bond.

You have voltage potential and it fluctuates. It's quite possible the inverter has the screw and/or some type of fault internally.

Signature Solar told me that the warranty is voided if I open the unit, unless I give them a statement from an electrician saying that the screw needed to be removed.

I'd contact them about your voltage readings for N-G bond and let them make a decision about letting you open the unit or sending out another unit as this one might have a fault. They should be able to tell you if the unit bonds N-G.

If the unit wasn't designed with dynamic N-G bond, I'd install one at the critical loads panel and bond there. Keep the critical loads panel separate from the main service panel and treat it as an off grid system.

 

[summit]

L1=119.7, L2=119,8

was the CRpanel NG bonded removed when L1/L2-N were measured above ? because with Vng fluctuating, it's surprising the L1/L2 are stable\

With CRpanel NG bond removed:
Inverter off: Vng = 0.097V; Rng = 1.6 Mohm; measured at inverter
Inverter on: Vng = 33V to 38V (fluctuating, sometimes O.L.); measured at inverter without the CRpanel bond, so there is no internal bonding screw?

Tesla EV doesn't charge at all without CRpanel NG bond (faults with loss of ground error on LED)

looks like EG46000 does not have an internal NG bond, given the Vng; which explains the EVSE's fault.

Inverter off:
PV+ to chassis: open with and without the PV wire connected
PV- to chasis: open with and without the PV wire connected
Is it ok to measure PV+/- to chassis while the inverter is ON?

since the schematic indicates the chassis is always grounded, the PV are good. I've measured Vdc PV+/- to ground with inverter on and gotten -Vdc between PV- and G; I think i need to scope it to see the PWM wave

 

[summit]

confirm the 2 problems: 1) can't charge with PV connected, 2) can't maintain 22A output. I think the EG6000 is problematic. Do a test with setting the Tesla charge rate to 20A. For sure the CRpanel needs NG bond

 

[bobphdee]

If anyone's interested, I collected some current and voltage plots while charging with the Tesla EVSE at 5A, 10A, 15A, 20A and 22A. I used a custom built IoT box with a current sensor attached to one of the hot lines in the critical load panel going to the charger. Interesting to see that the current waveform has some distortion and noise at the peaks at the higher currents. The plots were generated using the Arduino serial plotter, if you're familiar with the Arduino IDE. The output scale is in analog to digital converter (ADC) counts since I haven't calibrated it yet, but the plots will give you an idea if the waveforms are noisy.

Tesla charger ON: 5 A (inverter LCD @ 20% load; 1200W, battery current = 28A)

View attachment 130915

Tesla charger ON: 10 A (inverter LCD @ 40% load; 2400W, battery current = 54A)
​

View attachment 130916


Tesla charger ON: 15 A (inverter LCD @ 59% load; 3600W, battery current = 80A)

View attachment 130917


Tesla charger ON:20 A (inverter LCD @ 79% load; 4800W, battery current = 106A)​

That is some distance. Is there any way you can install a ground electrode at the critical loads panel? This would allow N-G bond at the critical loads panel without all the resistance of the current ground circuits.



You have voltage potential and it fluctuates. It's quite possible the inverter has the screw and/or some type of fault internally.

I'd contact them about your voltage readings for N-G bond and let them make a decision about letting you open the unit or sending out another unit as this one might have a fault. They should be able to tell you if the unit bonds N-G.

If the unit wasn't designed with dynamic N-G bond, I'd install one at the critical loads panel and bond there. Keep the critical loads panel separate from the main service panel and treat it as an off grid system.

That is some distance. Is there any way you can install a ground electrode at the critical loads panel? This would allow N-G bond at the critical loads panel without all the resistance of the current ground circuits.

Yes, that 137' ground wire probably has a voltage drop. I looked up the resistivity of 6 ga copper wire: 0.407 ohms/1000 ft or 0.05 ohms for 137 ft. Voltage drop = 0.05V per 1 A of current, so there shouldn't be much of a voltage drop, assuming my EGC current is low, if any.

I have an auxiliary ground next to the garage that's roughly a 25' run. I was originally using it for the system ground, but the EVSE chargers kept faulting with that ground rod - finally got the Tesla UMC to work after I disconnected the auxiliary ground rod and tied the critical panel EGC to the main panel water pipe ground. However, the Tesla UMC only works with the critical load panel N-G bond and the PV+/- disconnected. The auxiliary ground rod serves as the PV panel frame ground.

Yesterday, I disconnected the neutral on the critical loads panel output going to my air conditioner and the Tesla UMC charged up to 22A without faulting, even with the PV+/- connected, so that was a breakthrough. The Juicebox EVSE still wouldn't work, however. Problem is, I won't be able to run my 120V air conditioner without the neutral. I don't charge the car and run the air conditioner simultaneously. Need to go back and check my air conditioner neutral wiring.

I'd contact them about your voltage readings for N-G bond and let them make a decision about letting you open the unit or sending out another unit as this one might have a fault. They should be able to tell you if the unit bonds N-G.

Signature Solar tier 1 tech support was very responsive but couldn't resolve my issues so it was escalated to tier 3 tech support. I submitted 11 pages of photos, measurements and wiring diagrams 10 days ago. Haven't heard a word since then! That's why this forum is so great!

 

[bobphdee]

was the CRpanel NG bonded removed when L1/L2-N were measured above ? because with Vng fluctuating, it's surprising the L1/L2 are stable\

looks like EG46000 does not have an internal NG bond, given the Vng; which explains the EVSE's fault.

since the schematic indicates the chassis is always grounded, the PV are good. I've measured Vdc PV+/- to ground with inverter on and gotten -Vdc between PV- and G; I think i need to scope it to see the PWM wave

was the CRpanel NG bonded removed when L1/L2-N were measured above ? because with Vng fluctuating, it's surprising the L1/L2 are stable\

Yes, the CR panel NG bond was removed when I measured the fluctuating Vng.

looks like EG46000 does not have an internal NG bond, given the Vng; which explains the EVSE's fault.

That's my conclusion as well. So I'm leaving the NG bond in the critical load panel.

Yesterday, I disconnected the neutral on the critical loads panel output going to my air conditioner and the Tesla UMC charged up to 22A without faulting, even with the PV+/- connected, so that was a breakthrough. The Juicebox EVSE still wouldn't work, however, which is a bummer because I'd like to make the Juicebox my primary EVSE charger. Problem is, I won't be able to run my 120V air conditioner without the neutral. I don't charge the car and run the air conditioner simultaneously. Need to go back and check my air conditioner neutral wiring.

My system has only 1 NG bond in the Main-panel when it's tie to the house wiring. In off-grid mode, the CRpanel has NG bond with G still sharing Main-panel G-rod, but no N connection

Sounds like we have the same off-grid setup: CR panel has NG bond with G sharing main panel G-rod, but no N connection. In your off-grid mode, since you say you have no N connection, are you not running any 120V loads?

Inverter off:
PV+ to chassis: open with and without the PV wire connected
PV- to chasis: open with and without the PV wire connected
Is it ok to measure PV+/- to chassis while the inverter is ON?

since the schematic indicates the chassis is always grounded, the PV are good. I've measured Vdc PV+/- to ground with inverter on and gotten -Vdc between PV- and G; I think i need to scope it to see the PWM wave

Do you have a 6000EX schematic that shows the chassis is grounded? Don't see one in the manual. Was referring to measuring the resistance between PV+/- to ground with the inverter on: in that case, I think we're measuring the parallel resistance of the PV panel output and MPPT charger input. Would be interesting to see the PWM waveform.

When you say you measured -Vdc between PV- and G, is -Vdc = - PV string voltage? I assume your PV- is not tied to ground, otherwise it would be zero volts.

 

[summit]

Does the Vng still fluctuating with the air conditioner disconnected ?

In off-grid mode, there is no N to main panel. In outage back-up mode, CR-N/G are connected to Main-N/G respectively. I have both 120/240 loads.

Vdc is just voltage in DC. My voltage between PV+/- is as expected ~2xx volt. The unexpected/unexplained is (PV+ to G) is +1xx volt and (PV- to G) is negative volt. I plan to put a scope on it but haven't get around to it.

As mentioned earlier, my system was throwing an overload fault. It went away when I disconnected the PV- from an existing grid-tie inverter, so now both PV+/- are going straight to the hybrid inverter. My hybrid inverter manual specifically says it does not work with grounded PV (being "non-isolated")

 

[bobphdee]

Does the Vng still fluctuating with the air conditioner disconnected ?

In off-grid mode, there is no N to main panel. In outage back-up mode, CR-N/G are connected to Main-N/G respectively. I have both 120/240 loads.

Vdc is just voltage in DC. My voltage between PV+/- is as expected ~2xx volt. The unexpected/unexplained is (PV+ to G) is +1xx volt and (PV- to G) is negative volt. I plan to put a scope on it but haven't get around to it.

As mentioned earlier, my system was throwing an overload fault. It went away when I disconnected the PV- from an existing grid-tie inverter, so now both PV+/- are going straight to the hybrid inverter. My hybrid inverter manual specifically says it does not work with grounded PV (being "non-isolated")

Hi Summit,

Does the Vng still fluctuating with the air conditioner disconnected ?

Without the NG bond in the critical loads panel, and with the air conditioner completely disconnected (no L1 or neutral), Vng fluctuates between 39 to 51 VAC.

In off-grid mode, there is no N to main panel. In outage back-up mode, CR-N/G are connected to Main-N/G respectively. I have both 120/240 loads.

Are you using a transfer switch with a switched neutral? Which model?

I ran more tests yesterday with the NG bond in the critical loads panel. I ran the Tesla Universal Mobile Charger (UMC) up to 20A but noticed the plug was getting warm, so I used a thermal camera to monitor it. Looks like the UMC was throttling back the current when the plug reached 115 deg F, so that would explain why it was fluctuating on its own between 22A and 15A the other day. After letting it cool down, I attempted another charging session but the UMC kept throwing an "adapter" fault and current limited itself to 8A. I suspect my UMC plug adapter is faulty so I scheduled a service call for next week.

Having said all that, I still can't get the Juicebox charger to work with the NG bond in the critical loads panel so I still have a ground fault somewhere that I can't figure out.

Thanks for your comments/questions and appreciate everyone's inputs!

 

[summit]

Hi Summit,

Without the NG bond in the critical loads panel, and with the air conditioner completely disconnected (no L1 or neutral), Vng fluctuates between 39 to 51 VAC.

since the EG6000 Neutral is floating, NG bond is essential in the critical load panel, and any stand alone connection

Are you using a transfer switch with a switched neutral? Which model?

My system is not yet completely set-up; my plan is to connect CR-N to Main-N and CR-G to Main-G, with NG bonded in Main panel. This configuration was tested good as expected. The CR-N to Main-N is disconnected while other test/configuration are being tried out. Am a bit paranoid.

I ran more tests yesterday with the NG bond in the critical loads panel. I ran the Tesla Universal Mobile Charger (UMC) up to 20A but noticed the plug was getting warm, so I used a thermal camera to monitor it. Looks like the UMC was throttling back the current when the plug reached 115 deg F, so that would explain why it was fluctuating on its own between 22A and 15A the other day. After letting it cool down, I attempted another charging session but the UMC kept throwing an "adapter" fault and current limited itself to 8A. I suspect my UMC plug adapter is faulty so I scheduled a service call for next week.

is this UMC the only Tesla charger you have right ? it does make sense about the temperature throttling. I've found that it is often warmest at the plug, and not along the cable. Check the 40-50 (?) outlet as well, it could have gotten hot and transferred the heat to the UMC plug

Having said all that, I still can't get the Juicebox charger to work with the NG bond in the critical loads panel so I still have a ground fault somewhere that I can't figure out.

have heard that JuiceBox can be finicky. Does it work from other Main-panel's outlets ?

 

[bobphdee]

since the EG6000 Neutral is floating, NG bond is essential in the critical load panel, and any stand alone connection

My system is not yet completely set-up; my plan is to connect CR-N to Main-N and CR-G to Main-G, with NG bonded in Main panel. This configuration was tested good as expected. The CR-N to Main-N is disconnected while other test/configuration are being tried out. Am a bit paranoid.

is this UMC the only Tesla charger you have right ? it does make sense about the temperature throttling. I've found that it is often warmest at the plug, and not along the cable. Check the 40-50 (?) outlet as well, it could have gotten hot and transferred the heat to the UMC plug

have heard that JuiceBox can be finicky. Does it work from other Main-panel's outlets ?

Yes, the Tesla UMC is the only Tesla charger I have. The Juicebox worked fine when I had 240V from the grid main panel to the garage. Unfortunately, to avoid pulling another 100+ feet of wiring (x4), I used that wiring for the off-grid 240V. So I no longer have grid-tied 240V at the garage to run the Juicebox.

 

[OnGrid]

I have an off-grid 6000ex (no grid in/out) that has the AC out connected to a main panel/load center. The panel has a ground/neutral bonding screw like the one shown in this image:
View attachment 121300
In attempting to learn more about ground neutral bonding I discovered that the LVX-6048 does ground/neutral bonding when in battery mode ( so for me all the time ) and assumed this model would be similar. (Signature Solar hasn't responded to my email yet)

If the two models do handle g/n bonding the same:
Do I have a double ground neutral bond? If so, what type of problems would that present in my system?

Did anyone ever answer HLD's question. Does he have a double ground neutral bond that he needs to be worried about? If so should he have removed the ground in the EG4 Inverter?

Feels like the thread has gotten of course.

 

[PreppenWolf]

If it were me I'd remove the bond in the load panel. The power source should have the NG bond, IMO for onsite power generation.

 

[summit]

Did anyone ever answer HLD's question. Does he have a double ground neutral bond that he needs to be worried about? If so should he have removed the ground in the EG4 Inverter?

the LVX-6048 is from MPP (Taiwan) whereas the 6000ex is from EG4 (China); two different companies. I have an U5648gk from MPP is there is a grounding screw to be removed. Earlier posts indicate some 6000ex are shipped without and some with the NG bond. Simply check Vng and Resistance-NG to determine whether yours has the internal bond. If there's an internal NG bond, then remove the NG in the critical load panel

Feels like the thread has gotten of course.

am kinda waiting for bobphdee to update whether his 6000ex is malfunctioning as he's working with SSolar

 

[SoFar]

We have kinda gone off into the weeds. It is fun but I would like to bring the discusion back on topic.

If someone with a unit could do the following test it would be very helpful

Conditions: Put the inverter into inverter/battery mode with no load on the output.
* No AC input,
* Inverter on and running from battery,
* no loads or other connections on the output

Test:
What is the voltage from neutral-out to ground?

Results:
* If the voltage is more than a volt or so, it would indicate (prove?) there is no bond while in inverter mode. (A bonded neutral would definitely show zero or near zero volts)

(Previous tests in pass-through mode showed a 24V differential between neutral and ground. That test proved there is no bond while in passthrough mode. If there is no bond while in inverter mode and no bond in passthrough mode, then there is no dynamic bonding)​

* If the voltage is close to zero, it is a weak indication that there is a bond but does not prove very much. If the neutral is bonded to ground, the voltage would zero. However, the voltage of a floating center tap could be also be zero....or it could be quite high.


ALSO: Has anyone opened a unit up? Has anyone seen neutral bonding screws?

I have an EG4 3kW and I measured the voltage as you requested and I got 62Volts. You stated if the voltage is more than a volt or so, it would indicate there is no bond while in inverter mode.

Signature solar had me test my unit for continuity a few days ago and told me if I had continuity, the inverter had an internal bond. I did what they did and it had continuity when the inverter was powered on and the on/off switch was on. (Note: i had disconnected my ground and AC outs to do the test.

I also have my ground and neutral (bonded) in the little panel i have that is fed by my inverter. Just like the creator of the thread HLD. So is this an issue for me too?

 

[FilterGuy]

Signature solar had me test my unit for continuity a few days ago and told me if I had continuity, the inverter had an internal bond. I did what they did and it had continuity when the inverter was powered on and the on/off switch was on. (Note: i had disconnected my ground and AC outs to do the test.

Testing continuity when the inverter is on can/will result in a bogus answer. The existing voltage will interfere with the continuity tester.

I have an EG4 3kW and I measured the voltage as you requested and I got 62Volts. You stated if the voltage is more than a volt or so, it would indicate there is no bond while in inverter mode.

That is a sure sign the neutral is not bonded to the ground.

I also have my ground and neutral (bonded) in the little panel i have that is fed by my inverter. Just like the creator of the thread HLD. So is this an issue for me too?

When you measrued the 62 volts, was the inverter connected to this panel? It must not have been or the bond in the panel would have zeroed out the floating voltage.

The bond in the output panel is a problem. When the inverter is in pass-through mode there are two bonds.


The system needs the bonding screw added back in.

 

[summit]

I have an EG4 3kW and I measured the voltage as you requested and I got 62Volts. You stated if the voltage is more than a volt or so, it would indicate there is no bond while in inverter mode.

Signature solar had me test my unit for continuity a few days ago and told me if I had continuity, the inverter had an internal bond. I did what they did and it had continuity when the inverter was powered on and the on/off switch was on. (Note: i had disconnected my ground and AC outs to do the test.

I also have my ground and neutral (bonded) in the little panel i have that is fed by my inverter. Just like the creator of the thread HLD. So is this an issue for me too?

what model is your 3kw ? 240vac output or 120vac ?

 

[SoFar]

Testing continuity when the inverter is on can/will result in a bogus answer. The existing voltage will interfere with the continuity tester.


That is a sure sign the neutral is not bonded to the ground.


When you measrued the 62 volts, was the inverter connected to this panel? It must not have been or the bond in the panel would have zeroed out the floating voltage.

The bond in the output panel is a problem. When the inverter is in pass-through mode there are two bonds.


The system needs the bonding screw added back in.

No, i disconnected the feeds to the panel (including the ground) when I checked the voltage. I just had the battery on and the inverter on.

Yes, when I measured the continuity, it created a sound on the meter, but the meter was showing different readings (oscillating).

 

[summit]

I have been doing these drawing long enough and have a large enough libray of drawings that I can usually just modify something I already have to quickly generate a new layout.

Everyone else is doing the heavy lifting. They are doing all the testing to determine what the hell the inverter is actually doing. I just put it all together based on what they figured out.

You are still absolutely correct..... It is criminal negligence that they do not provide the information we need to know how to properly use the product.

hi FilterGuy - I find your drawings very clear and useful. Mind sharing your various block diagrams ? looks like they were done in PowerPoint ? your knowledge and contribution are very much appreciated