How does your inverter deal with ground.

[Leon]

They probably expect you to share the Ground on the output with ground on the input. What does the manual say?

That's what I assumed, but the manual shows the output ground wire going to the screw on the chassis.

 

[FilterGuy]

SMA Sunny Island (SI-6048US, etc.)
Has two ground lugs for up to 4 awg, one AC input and one AC output. They are connected together and to chassis.
It has a larger (2/0 or so) lug for battery grounding connected to chassis. SMA says grounding the 48V battery is optional, and can be positive and negative.

NEC may want the battery grounded, but SMA doesn't care.
It has PTC fuses (0.5A) on both battery positive and negative, for use powering the two dry-contact SPDP relays it controls (or they can switch AC.

Neutral is not bonded to ground internally. Inverter is supposed to be hard wired, with neutral bonding outside.

Good info. I downloaded and scanned through the manual and did not see where it talks about N-G bonding....but your description makes sense for that style of inverter. I did notice that it said the Neutral in and out are tied together. I will try to get it added to the resource soon.

NEC may want the battery grounded, but SMA doesn't care.

This is not a surprise. The only time I have seen an inverter that cares is when they have built in SCC with Ground Fault detection.

BTW: Do you have an SI-6048US? How do you like it? It looks like a nice unit and Sunnyboy is a well known brand.

 

[FilterGuy]

That's what I assumed, but the manual shows the output ground wire going to the screw on the chassis.

The ground is all tied together so the only thing that matters is that the AC out ground is connected to one of the spots. They probably show it going to the chassis lug because the AC-in connection can't fit both ground wires

 

[Hedges]

Good info. I downloaded and scanned through the manual and did not see where it talks about N-G bonding....but your description makes sense for that style of inverter. I did notice that it said the Neutral in and out are tied together. I will try to get it added to the resource soon.

This is not a surprise. The only time I have seen an inverter that cares is when they have built in SCC with Ground Fault detection.

BTW: Do you have an SI-6048US? How do you like it? It looks like a nice unit and Sunnyboy is a well known brand.

"The Sunny Island is designed for indoor use only and may only be used as stationary equipment." (new European model is rain tight for outdoor.)

If I used it for a mobile application with shore-power cord, I would either add a relay to bond neutral to ground when input voltage not present, or park the plug in a socket having neutral wired to ground.

I took advantage of the two neutrals and the two grounds being common. I was running out of conduit capacity for pulling wires so AC line (hot) is wired to both AC input and AC output from breaker panels for "grid" and "island", but for neutral and ground just has one wire (branched at "island" breaker panel rather than at Sunny Island").

I have an SI-5048US with 100 AH AGM, fed by MSTE "Sunny Island Charger 40" (communication link between them, also works with Midnight Classic using an adapter), and 120/240V transformer.

I have four SI-6048US configured 2s2p with 405 Ah AGM, acting as UPS between the grid and my house. Sunny Boys are on a panel fed by Sunny Island, and a 100A load-shed relay disconnects the house at 70% DoD. I plan to further automate disconnecting less important and heavy loads (A/C, laundry, etc.) at a higher SoC. For now, I do that by timer (A/C) or manually.

Yes, I like it. It is robust and has worked fine. Probably I only needed two of them 2s, but I was able to get four which doubles feed-through current to/from grid to 112A in case I happen to draw more. I've done a little testing with running off-grid and varying loads. Sunny Island never complains about anything. Sunny Boy doesn't like VFD which drives 3-phase pool pump (capacitor/diode front end.)

One morning the power was out and I didn't even notice until I walked over the garage. Sunny Island was powering my 10kW electric duct heater (furnace). That would have drained the batteries in 90 minutes, so I switched to gas. One of the things I want to automate.

Sunny Island can be connected 4p for 120V single phase (220V with European model), 2s2p for 120/240V split-phase, or 3y for 3-phase. Either grid-backup or off-grid.
With 3-phase, multiple of these island clusters can be interconnected, at least in an off-grid system.

 

[Aquanuts]

Interesting. What model do you have? According to the manual for the LV2424 inverter/charger this is conceptually the way it should work:

View attachment 57346

(Unlike many companies, Aims documented the LV2424 grounding scheme pretty well.)
As you can see, When on battery, the LV2424 creates the N-G bond and when on shore power it does not. This is pretty standard for a good inverter/charger because you always want one and only one N-G Bond and shore power should always have a N-G Bond.


That seems like a really bad design for an inverter. It completely prevents one of the primary purpose of the 'ground' wire in the AC wiring from working: A short to the ground wire will not pop the breaker and clear the fault.

++The model is PWRI150024S. I don't have a schematic, probably should have asked for one, but I found the explanation to be credible. If there is a ground short, there would be a difference in the current flow between line and neutral which would allow the inverter to sense it. The inverter does sense ground and will trip for that. I'm not ready to say it's a bed design. but I would think that the system should still have a chassis ground. Then again, I,m not a licensed electrician but I have stayed at a Holiday Inn before.

 

[FilterGuy]

The model is PWRI150024S

I just looked up the manual for this and it is striking how poorly documented it is compared to the Aims LV2424. I am pretty sure this is a rebrand of another companies inverter that is targeted at the lower price market.

 

[Aquanuts]

Interesting. What model do you have? According to the manual for the LV2424 inverter/charger this is conceptually the way it should work:

View attachment 57346

(Unlike many companies, Aims documented the LV2424 grounding scheme pretty well.)
As you can see, When on battery, the LV2424 creates the N-G bond and when on shore power it does not. This is pretty standard for a good inverter/charger because you always want one and only one N-G Bond and shore power should always have a N-G Bond.


That seems like a really bad design for an inverter. It completely prevents one of the primary purpose of the 'ground' wire in the AC wiring from working: A short to the ground wire will not pop the breaker and clear the fault.

The PWRI150024S does not have an AC input, so there is no way for the unit to sense that there is another source. I ran both shore power and the inverter through a relay that switches the lines and the neutrals. When on shore power, as in your drawing above, the neutral to ground bond is made at the shore power panel. I do have the shore power ground bonded to the chassis and whenever the cord is installed the ground will be applied to the chassis but the neutral will float unless there is power on the shore power line, then the relay will switch to shore power. There is about a 10 - 15 sec time delay that is to let the generator stabilize. So far, I have only used grid power with it. The chassis of the inverter is tied to ground as is the DC negative.

 

[Aquanuts]

I just looked up the manual for this and it is striking how poorly documented it is compared to the Aims LV2424. I am pretty sure this is a rebrand of another companies inverter that is targeted at the lower price market.

Yes, it seems the schematic is a well guarded secret. As to the price, it was on that big internet site that sells real cheap, who's name I wont mention here.

 

[Leon]

A bit of an update on this, perhaps some usable information for your resource:

This is for the unbranded all-in-one MPS-V PLUS 3.5kW



I have only tested with a battery connected - no PV or AC input.

lets walk through it.

Ground:
Every inverter I have looked at or have the information for has the Chassis Ground, AC input Ground, and AC output ground all tied together. You can check this with an ohm meter.

As a 240 V unit, you will still have a neutral and a Hot.... I do not have the manual so I can't really say how it is marked.

Neutral-Ground Bond.
This gets a little tricky because you don't know what relays are internal to the inverter and what the 'non-powered' state of the relay is.
When the inverter is hooked up you can put an AC voltmeter between neutral and ground. If there is a voltage higher than a few milivolts, there is probably not a N-G bond. If the voltage is zero or just a few milivolts, there probably is a N-G bond.

Since it is a 240V Single phase output, my guess is that you will find that there is no N-G bond. However, if you are putting in an off-grid system you need to decide what your grounding scheme will be. The key question will be 'Will the breaker trip if there is a short between the Hot and ground/safety wire. One way to accomplish this is to create a N-G bond at the inverter. In the event of a short, the bond will alow a high current that will trip a traditional breaker and clear the fault. However, people have reported that some manufacturers say not to do this or it will damage the inverter. (That seems like a bad design to me and I sometimes wonder if the call agent does not really know what they are talking about).

• As you predicted, AC in and chassis ground have continuity.
• In off grid mode, powered only by the battery, I get 115V between AC out neutral and ground, while line to ground gives 117V. Is it normal for ground to be halfway between line and neutral like that?

Full voltage between line and neutral.


Here is my arrangement for breaker and ground-neutral bond. Neutral and ground bus-bars have a copper link between them. The cable on the left goes to my earth stake, in the clay soil on the other side of the wall. The bus outside the white enclosure allows for any other grounding. I have not yet bonded the DC side of the system (24V) to ground. The breaker is a RCBO (residual current circuit breaker with overcurrent protection). I think this is similar to what you call a GFCI in North America? It will trip with current over 20A, or if there is an imbalance between active and neutral of over 30mA (Actually less in my testing - two 4.7k resistors in series between active and earth will trip it, which should be about 25mA. It would not trip with 3 resistors, so I think the true trip value is between 16-25mA)

Another way to 'clear the fault' is to use CGFI breakers that will trip if the current in the hot and neutral become unbalanced.

The key is to have one and only one NG-bond in the system. If you have an inverter set up and there is an external N-G ground, you can check to see if there is an internal N-G bond by putting a clamp on ammeter on the ground wire between the inverter and the external N-G ground while there is an AC load on the system. If there is current on the ground wire, then the inverter almost certainly has an internal N-G ground and the external N-G ground should be removed. (Note: That current on the ground wire is a safety issue and is why you should not have more than one N-G ground.

I tested this and found no current on the ground wire. Seems like there is no internal NG bond, at least the way I am using it.

Note: In all of the discussion so far in this post, the term 'ground' has almost nothing to do with earth ground. It is talking about the grounding wire that is throughout the AC system (In the US, this is called the Equipment Grounding Wire) The tests are the same regardless if the ground wire is tied to earth.

The primary purpose of tying the ground wire to earth ground is to ensure the system does not 'float' to a voltage that is significantly higher than the earth.

Since you will only be using it as an inverter this does not apply, but the issue gets more complicated if you are using 'shore power'. Since shore power [almost] always has a N-G bond, there should be no additional N-G bond when on shore power. However, when you are not on shore power, there needs to be a bond. Some inverters handle this with relays. Some don't handle it at all.

 

[Hedges]

A bit of an update on this, perhaps some usable information for your resource:

This is for the unbranded all-in-one MPS-V PLUS 3.5kW

View attachment 58052

I have only tested with a battery connected - no PV or AC input.

• As you predicted, AC in and chassis ground have continuity.
• In off grid mode, powered only by the battery, I get 115V between AC out neutral and ground, while line to ground gives 117V. Is it normal for ground to be halfway between line and neutral like that?

Full [230V] voltage between line and neutral.


Here is my arrangement for breaker and ground-neutral bond. Neutral and ground bus-bars have a copper link between them. The cable on the left goes to my earth stake, in the clay soil on the other side of the wall. The bus outside the white enclosure allows for any other grounding.



I tested this and found no current on the ground wire. Seems like there is no internal NG bond, at least the way I am using it.

When AC output from a transformer or inverter is isolated, it is often approximately symmetric around ground because each leg has parasitic capacitance to ground. But if you load it (e.g. with an old-fashioned light bulb) from neutral to ground (alternate test to battery negative) usually no current can be delivered and voltage drops to zero. So that's how I suggest testing an inverter, to see if it can tolerate strapping neutral to ground. Some cheap inverters in the U.S. deliver +/-60V relative to battery. Grounding neutral would cause AC to appear on battery terminals.

Once you do ground neutral, that capacitive coupling can drive a small amount of ground current. We're studying that at work, seeing around 2 mA to 10 mA or so. This could be a problem for GFCI/RCBO if a load connected downstream of the breaker had such capacitance to ground. EMI line filters are one such source of capacitance to ground.

We use 5 mA GFCI trip level for human safety with separate GFCI per branch circuit (rather than whole house at 30 mA as in Europe.) We may have 30 mA on some circuits to protect equipment.

In your picture, because there is only one ground terminal you have the output cord connected to a screw in the sheet metal. An issue with that is if input goes to utility and output has a short to ground, the ground current has to pass through sheet metal to trip breaker. We are supposed to make a connection to cabinet frames and sheet metal, but also have continuity through wires, not depending on metal structure. I think solar panel grounding now has an exception to that, things like "WEEB" grounding devices which have toothed washers that bite into metal components so they can be used for ground conductors.

At least here in the U.S., we bond neutral to ground wire only at one place (usually at service entrance from utility grid, where we also have a ground rod.) Problem with having a second ground at your inverter is that return currents in neutral from your loads or others (like motors somewhere on the property) can go through neutral to your chassis and back through your ground wire. Also, open circuits in some locations could make your chassis hot.

If your inverter is source of power and you have a neutral-ground bond at the inverter, no current in ground wire. If utility is source of power feeding through inverter and there is neutral-ground bond both at inverter and at utility connection, ground wire from inverter to load will carry zero current (except any capacitive coupling) but from utility connection to inverter, the return current will split between neutral and ground. If you saw zero ground current, that was due to where power came from and which ground you measured.

Inverters for mobile (RV, boat) applications have a particular UL listing and automatically bond neutral to ground when unplugged from the grid, isolate when connected to grid because bonding occurs in the wiring circuit feeding it.

 

[Leon]

Thanks for the input.

When AC output from a transformer or inverter is isolated, it is often approximately symmetric around ground because each leg has parasitic capacitance to ground. But if you load it (e.g. with an old-fashioned light bulb) from neutral to ground (alternate test to battery negative) usually no current can be delivered and voltage drops to zero. So that's how I suggest testing an inverter, to see if it can tolerate strapping neutral to ground. Some cheap inverters in the U.S. deliver +/-60V relative to battery. Grounding neutral would cause AC to appear on battery terminals.

I just tried this with a 275W incandescent heat lamp, the only non-LED bulb I have around. When connected across neutral and chassis ground, with nothing else connected to the inverter, the voltage drops to 49mV. My meter still recognises the 50hz frequency. I can not measure any current flow using a clamp meter. With the NG link in place, I tested for AC across the main battery terminals and got nothing. I have not grounded either side of the DC relative to the inverter. Does this sound ok?

Once you do ground neutral, that capacitive coupling can drive a small amount of ground current. We're studying that at work, seeing around 2 mA to 10 mA or so. This could be a problem for GFCI/RCBO if a load connected downstream of the breaker had such capacitance to ground. EMI line filters are one such source of capacitance to ground.

When you say a problem, do you mean nuisance tripping? Or is this ground current dangerous?

We use 5 mA GFCI trip level for human safety with separate GFCI per branch circuit (rather than whole house at 30 mA as in Europe.) We may have 30 mA on some circuits to protect equipment.

Seems like your standards are safer for humans, at least in that regard. I'd prefer to conduct 5mA than 30. I might see if I can find a breaker with a lower trip rating, though 30mA seems to be standard.

In your picture, because there is only one ground terminal you have the output cord connected to a screw in the sheet metal. An issue with that is if input goes to utility and output has a short to ground, the ground current has to pass through sheet metal to trip breaker. We are supposed to make a connection to cabinet frames and sheet metal, but also have continuity through wires, not depending on metal structure. I think solar panel grounding now has an exception to that, things like "WEEB" grounding devices which have toothed washers that bite into metal components so they can be used for ground conductors.

Yeah, ok. While I won't be using utility AC input with this system, I could easily connect the ground wire to the AC input ground terminal and add a jumper to the chassis screw. Does that make sense?

Regarding my solar panel grounding, I have run a wire from the racking and tied it in to the ground wire that goes to my earth stake.

At least here in the U.S., we bond neutral to ground wire only at one place (usually at service entrance from utility grid, where we also have a ground rod.) Problem with having a second ground at your inverter is that return currents in neutral from your loads or others (like motors somewhere on the property) can go through neutral to your chassis and back through your ground wire. Also, open circuits in some locations could make your chassis hot.

If your inverter is source of power and you have a neutral-ground bond at the inverter, no current in ground wire. If utility is source of power feeding through inverter and there is neutral-ground bond both at inverter and at utility connection, ground wire from inverter to load will carry zero current (except any capacitive coupling) but from utility connection to inverter, the return current will split between neutral and ground. If you saw zero ground current, that was due to where power came from and which ground you measured.

Understood. So if I was to connect AC input, I would need to remove the NG bond at the inverter and then replace it for off-grid use. I have read that some all-in-one units do this internally, but I have no indication whether mine does or not. Seems like it would be a terrible design for a unit to be able to switch from utility to battery source, but not handle the NG linking. It's a good thing I have no plans for utility connection, I am at the bleeding edge of my knowledge as it is.

Inverters for mobile (RV, boat) applications have a particular UL listing and automatically bond neutral to ground when unplugged from the grid, isolate when connected to grid because bonding occurs in the wiring circuit feeding it.

Sounds like a standard that ought to be applied in any inverter with the provision for taking supply from the grid for charging or pass-through.

Thanks again for your help, I really appreciate it.

 

[Hedges]

Sounds OK.

Ground current could trip a GFCI. In our case, we're chasing down interference in an analog instrument. Such currents shouldn't be a safety hazard (except in a hospital, where separate grounding is used.)

In the U.S., I think you can find single and two-pole GFCI breakers for 5 mA from 15A to 50A.
At 5 mA, you're protected and you can let go. I think at 30 mA (say 29, so it doesn't trip), you might be left grabbing the wire and unable to let go, but it doesn't kill you.

Yes, copper joined to copper, connected to the ground input and the ground screw sounds good. They don't like us to put two wires under one setscrew for some reason. Except when it is a setscrew "wire nut" rated for multiple wires. Could attach a busbar to the ground screw and land several wires in it. I've used split-bolts to splice wires, with one stripped longer to fit in a terminal.

Solar panel frame ground wire should provide copper path back to ground connection of charge controller. It will also get to the ground rod, but can't only go to a ground rod (unless you want to catch nightcrawlers.)

Switching from utility to battery while hard wired, still have neutral/ground bond at utility panel. Shore power plug of RV/boat, that's where the bond (which hopefully exists) is disconnected when you unplug. And then you hope hot/neutral are wired correctly. "Trust, but verify."

I'm still learning. Dealing with 60 Hz noise in instruments currently, but working with another old timer who has made it his life's work.
I've got a scope with clamp current probe on one wire, and I've set up switches, plugs, transformers, chokes. Oops, my laptop power supply was too close and plugged into the same power strip. Need to unplug, run on batteries, and repeat my noise measurements. I expect to be spec'ing some custom line filters.

The basic functioning of transformers/chokes I've got, but what the hysteresis of BH curves really means for performance is new to me. I managed to measure it with my variac and scope at home, but mapping it to a useful SPICE model is a challenge. Especially when the simulation crashes whenever I push it toward saturation. But my bench tests don't crash

Hysteresis - Wikipedia

en.wikipedia.org

Saturation (magnetic) - Wikipedia

en.wikipedia.org

 

[kite]

lets walk through it.

I bought an AMPINVT 2000W (12VDC/120VAC) inverter charger off Amazon (open box but appears new, $350 shipped). I'd like to understand how to safely use it as a backup power supply during power outages. Battery will be 560Ah 12V.

The manual makes ZERO mention of grounding or N-G bond. I've found a handful of threads mentioning these inverters but no relevant details. But @FilterGuy outlined nice instructions in post 18. Following those, I tried to figure out what's going on and how I should set things up. I'd appreciate input!

Ground:

Every inverter I have looked at or have the information for has the Chassis Ground, AC input Ground, and AC output ground all tied together. You can check this with an ohm meter.

CONFIRMED, via ohm meter and by chasing wires inside. Chassis ground strap is present.

Neutral-Ground Bond
Check my conclusion based on the below measurements. But it seems like this unit does not have an internal N-G bond in any modes (line priority, battery priority, etc). The only N-G bond I've detected is in my house panel.


AC in connected (grounded circuit) - no DC connected

• AC out N-G voltage = 68mV
• AC in N-G voltage = 68mV (this and the prior reading wander around, but remain the same)
• N-G beeps when continuity checked

AC in connected via non-grounded extension cord - no DC connected

• AC out N-G voltage = 39V
• AC in N-G voltage = 39V
• N-G - open circuit

DC in only

• AC out N-G voltage = 8V
• N-G no continuity

AC in connected (grounded circuit) - DC connected

• AC out N-G voltage = 130mV
• AC in N-G voltage = 130mV
• N-G beeps when continuity checking

So, for my use case, would the below be appropriate configurations?

1. keep the unit plugged into a grounded wall outlet even when the grid is down (take advantage of the panel N-G bond)
2. if using away from grid/wall outlet:
   1. install a N-G jumper in one of the AC outlets on the inverter (or into the screw terminals, though this leaves more room for error during configuration changes)
   2. run all loads through a GFCI outlet
3. my neighbor has a cheapo generator I could charge from if needed. I should check if it has a N-G bond and make sure only one N-G bond is connected (generator or my jumper)

 

[FilterGuy]

keep the unit plugged into a grounded wall outlet even when the grid is down (take advantage of the panel N-G bond)

That sounds right.

if using away from grid/wall outlet:

1. install a N-G jumper in one of the AC outlets on the inverter (or into the screw terminals, though this leaves more room for error during configuration changes)

That would work.

https://www.amazon.com/Southwire-Company-LLC-44400-Generator/dp/B07F4R7BDL/ref=sr_1_5?crid=3N89DV02S70VF&dchild=1&keywords=neutral+ground+bonding+plug+for+generator&qid=1627694132&sprefix=Neutral+Ground+%2Caps%2C243&sr=8-5

You could also build a female version of the above N-G plug and when you are not plugged into the generator, plug cord that goes to the generator into the Female N-G plug to provide the bond.

run all loads through a GFCI outlet

That works too.

my neighbor has a cheapo generator I could charge from if needed. I should check if it has a N-G bond and make sure only one N-G bond is connected (generator or my jumper)

Most portable generators do NOT have an NG bond.

 

[Leon]

Sounds OK.

Ground current could trip a GFCI. In our case, we're chasing down interference in an analog instrument. Such currents shouldn't be a safety hazard (except in a hospital, where separate grounding is used.)

In the U.S., I think you can find single and two-pole GFCI breakers for 5 mA from 15A to 50A.
At 5 mA, you're protected and you can let go. I think at 30 mA (say 29, so it doesn't trip), you might be left grabbing the wire and unable to let go, but it doesn't kill you.

Yes, copper joined to copper, connected to the ground input and the ground screw sounds good. They don't like us to put two wires under one setscrew for some reason. Except when it is a setscrew "wire nut" rated for multiple wires. Could attach a busbar to the ground screw and land several wires in it. I've used split-bolts to splice wires, with one stripped longer to fit in a terminal.

Solar panel frame ground wire should provide copper path back to ground connection of charge controller. It will also get to the ground rod, but can't only go to a ground rod (unless you want to catch nightcrawlers.)

Switching from utility to battery while hard wired, still have neutral/ground bond at utility panel. Shore power plug of RV/boat, that's where the bond (which hopefully exists) is disconnected when you unplug. And then you hope hot/neutral are wired correctly. "Trust, but verify."

I'm still learning. Dealing with 60 Hz noise in instruments currently, but working with another old timer who has made it his life's work.
I've got a scope with clamp current probe on one wire, and I've set up switches, plugs, transformers, chokes. Oops, my laptop power supply was too close and plugged into the same power strip. Need to unplug, run on batteries, and repeat my noise measurements. I expect to be spec'ing some custom line filters.

The basic functioning of transformers/chokes I've got, but what the hysteresis of BH curves really means for performance is new to me. I managed to measure it with my variac and scope at home, but mapping it to a useful SPICE model is a challenge. Especially when the simulation crashes whenever I push it toward saturation. But my bench tests don't crash

Hysteresis - Wikipedia

en.wikipedia.org

Saturation (magnetic) - Wikipedia

en.wikipedia.org

Thanks again. I'm now more confident my setup has a lower risk of electrocuting my family while burning the house down.

Always more to learn. I'd love to get some sort of official training in systems like this, but the only way seems to be a four year apprenticeship. Not really practical while working full time. Maybe it's time to look at online engineering degrees.

 

[Hedges]

Thanks again. I'm now more confident my setup has a lower risk of electrocuting my family while burning the house down.

Always more to learn. I'd love to get some sort of official training in systems like this, but the only way seems to be a four year apprenticeship. Not really practical while working full time. Maybe it's time to look at online engineering degrees.

There are for-profit schools offering one or two year programs in various technician and nursing fields, for about $30,000.
Similar programs in junior colleges, for a lot less money (but popular programs are sometimes full.)

Universities like Cal State have more advanced programs at higher cost than a J.C. but cheaper than (officially) for-profit schools. Easy enough to qualify to be admitted. The big universities are more selective. But both kinds will let people take some classes without being admitted, space available. (Even Stanford, but you have to pay for a full load even if only taking one class.) Berkeley, I noticed undergrad quantum physics was always full, but graduate level had space available. (If I'd had the initiative to self-study and master the undergrad I might have considered attending. The units could have counted toward a degree I got elsewhere.)

But if you just want to learn, not get a sheepskin, check out "MIT Open Courseware", video of lectures and copies of assignments. Based on prior year texts, which you can find on Amazon much cheaper than this years texts.

MIT OpenCourseWare | Free Online Course Materials

MIT OpenCourseWare is a web based publication of virtually all MIT course content. OCW is open and available to the world and is a permanent MIT activity

ocw.mit.edu

Or if you want a degree, maybe your employer will pay?

I went full-time to SJSU during my year of unemployment, then took one or two classes at a time, often skipping a semester, until I finished. (Tuition covers 2 courses, more costs extra. Can skip a semester without being dropped.)

Some big employers even pay the university to teach a tailored program on-site (about triple the tuition cost.)

 

[kite]

That sounds right.

Thanks for all the feedback!

You could also build a female version of the above N-G plug and when you are not plugged into the generator, plug cord that goes to the generator into the Female N-G plug to provide the bond.

Smart!!

 

[Fpoet]

This thread comes at a good time, as I recently installed a Samlex PST-2000-12 inverter (hardwired to a Samlex STS-30 transfer switch) in my van. The setup works fine on both shore and inverter power, but I would not mind confirmation with regards to grounding.

The Samlex manuals are quite detailed and I followed them the best I could. The inverter does have internal neutral-ground bond, and the manual specifically describes how to connect both the devices (section 8.6 in the inverter’s manual).

So in the van, the inverter is hardwired to the transfer switch, which is wired to the breaker panel, which feeds only GFCI outlets. To have a single grounding point, the only grounding connection I have is through the inverter’s ground lug, to the chassis.

Is this correct?

TIA

 

[Leon]

There are for-profit schools offering one or two year programs in various technician and nursing fields, for about $30,000.
Similar programs in junior colleges, for a lot less money (but popular programs are sometimes full.)

Universities like Cal State have more advanced programs at higher cost than a J.C. but cheaper than (officially) for-profit schools. Easy enough to qualify to be admitted. The big universities are more selective. But both kinds will let people take some classes without being admitted, space available. (Even Stanford, but you have to pay for a full load even if only taking one class.) Berkeley, I noticed undergrad quantum physics was always full, but graduate level had space available. (If I'd had the initiative to self-study and master the undergrad I might have considered attending. The units could have counted toward a degree I got elsewhere.)

But if you just want to learn, not get a sheepskin, check out "MIT Open Courseware", video of lectures and copies of assignments. Based on prior year texts, which you can find on Amazon much cheaper than this years texts.

MIT OpenCourseWare | Free Online Course Materials

MIT OpenCourseWare is a web based publication of virtually all MIT course content. OCW is open and available to the world and is a permanent MIT activity

ocw.mit.edu

Or if you want a degree, maybe your employer will pay?

I went full-time to SJSU during my year of unemployment, then took one or two classes at a time, often skipping a semester, until I finished. (Tuition covers 2 courses, more costs extra. Can skip a semester without being dropped.)

Some big employers even pay the university to teach a tailored program on-site (about triple the tuition cost.)

Plenty to think about. I am in Australia. Most of our universities derive most of their funding from international students. Since the borders closed, this has almost totally collapsed so I suspect these institutions will be much less selective about who they take. My day job is teaching high school engineering and related technical subjects, so there is a chance my employer would be supportive.

 

[Riderice]

I have an AIMS 24vdc to 115vac 1500W/3000W pure sine wave. I have been having an issue do to it being under powered, unrelated to this thread but during the course of troubleshooting the problem I discovered that a circuit checker indicates an open ground. AIMS Tech Support says that is a normal reading as the neutral line carries voltage, (I didn't get the value), and must be left to float from ground. He also said that connecting the neutral to ground would damage the inverter.

Same problem here, just got my new AIMS 48v 6000w inverter last night, hooked it up, saw the open ground. So I thought, oh I'll just bond the ground-neutral externally. Emailed Aims to make sure that's okay, and good thing I did because I got this reply:

"You CAN NOT bond the neutral and ground on this inverter as it is a high frequency style inverter and creates 60V on the neutral and 60V on the hot line. Bonding these will force the power to the hotline and cause damage to the inverter that is not covered under warranty."

So now I have to waste hundreds of $$$ on shipping and "restocking fees" to return this garbage, and get something with a real 120 hot wire, and that can be bonded properly.

 

[Lincgen]

View attachment 33722

EPEVER UP3000-HM10022
Hybrid Inverter

Just Been through this exact same senario with Epever, initially contacted them about their output Neutral Ground /Earth bonding and after confirming its suitability i found out that infact when supplied by a Mains TNC-S supply( Neutral Earthing bond at the Main switchboard ) operating in Bypass Mode the earthing system configuration passes through to the output correctly', but when placed in inverter mode the inverter output creates an IT system /isolated supply -unbonded system , therefor in Australia the outputs downstream RCD protection would not be configured correctly for correct operation as its looking for a TNC-S supply

In simple terms the output should be neutral earth bonded but cannot be because when in bypass it already is and then placing a second neutral earth bond at the output places this in parallel to the first, should the inverter supply be by a RCD / GFCI circuit then nuisance tripping will occur whilst protective earth would share load current with the neutral that's not permitted

So moral to this story , understand your local standards and the accepted earthing system to be employed , because there are a multitude of earthing systems and the language varies by country , ie GFCI,RCD, Neutral Earth Bonded, MEN system , TNC, TNCS , IT , Earthing Conductor , Grounding Conductor, being very clear with your inverter supplier of what you require in the manuals i have read, this topic is definitely not clear when you start dealing with Clone type inverters and Languages or in Australia we say the Lingo

The solution to the above is now not that easy , an inverter with a configurable earthing system is what's required or a system as shown in filter guys inverter drawings where the output is automatically reconfigured, please look at the variances and i congratulate the Filter Guy for trying to clarify and hope i have added some further incite

 

[mrdavvv]

Anyone has info about Solark / Deye configuration=?

 

[FilterGuy]

Anyone has info about Solark / Deye configuration=?

The manual does not explicitly describe the grounding and bonding, but from what it does say I am pretty confident in the following:

1) The SolArk does *not* create a Neutral-Ground Bond in any situation.
2) The Neutral is common throughout the system. (there is no difference between AC neutral in and AC neutral out)
3) The Ground is common throughout the system. AC Ground in, AC Ground Out, and the Sol-ark chassis are all tied together.

For a home installation, the sol-ark is pretty straight forward. For a mobile situation, there is a need to create a Neutral-Ground Bond when not on shore power. Conversely, when on shore power there should not be a bond. This could be done with a 120V relay that is powered by shore power.

Here are a few things from the manual:



The requirement that the Solar PV is not grounded is because the built in ground fault protection is providing grounding of the PV circuit. However, the frames of the panels still need to be grounded (See below)


Notice that in the diagram above, the loads and the Solark are both tied to the main panel neutral bar.



The following is in a section about wiring the solar panels

 

[eabyrd]

@FilterGuy I've read your 2nd paper, and you may recall me as the guy trying to figure out a grounding issue with the MPP LVX6048 on another thread. Setting that aside for a moment, and assuming (yeah I know, don't say it) that their wiring is like SOL-ARKs given the also have you connect both neutral & ground to both AC input & AC output sides. Does the following seem right. I've never bothered to ground my Battery before or my PV input lines.

Please take a look and offer any feedback you think appropriate.

I appreciate all you do here sir, you're willingness to give is inspirational

Thanks in advance

 

[FilterGuy]

Mostly correct..... but....



A small nit: I would draw (and implement) the wire that is grounding the DC as green and connect it to the grounding bus. In your setup it is electrically equivalent but if you use a white wire and connect it to common, it might be confusing to anyone looking at the system.

I found this in the manual:



Many/most of the all-in-ones have built in PV Ground fault protection these days and I am pretty sure the 6048 does as well. You should NOT connect the PV- input circuit to ground (The PV circuit will be grounded by the internal ground fault protection and doing another external ground connection will probably either disable or trigger the internal ground fault protection)

As is the case with most manuals from MPP, they say precious little about how it handles grounding. Therefore I can not say for sure if the battery circuit is grounded internally or should be grounded externally. (NEC does require 48V battery circuits to be grounded) My working assumption is that it should be grounded externally as you kinda show.

The documentation also does not discuss bonding so I can't say for sure what should or should not be done for N-G bonding on the critical loads panel. My working assumption is that it should *not* have an N-G bond.

BTW: I assume your 200A main breaker panel is tied to earth ground. That should be the only place in the system that gets tied to earth ground. (Do not tie the critical loads box to a separate earth ground.

BTW2: The neutral bus in the critical loads panel must be electrically isolated from the metal box of the panel. The metal box of the panel should be tied to the grounding circuit.