Modified Sine Wave Converter

[GeorgeRa]

Originally Posted by booster

…………………………Both of the small inverters are (now) permanently bonded because they are isolated from the van wiring and ground, which is what George has also, I think. ……………………………..

I followed the installation manual from Mornigstar. The Morningstar AC out is floating, I picked one of these two AC wires and bonded to the ground, it became neutral. Both outlets are GFCI. I only use either shore/1000W Magnum or Morningstar 300W. I think, accidental turning both wouldn't cause a safety problem. Galley's GFCIs green indicators show me which AC circuit is on, Morningstar or Magnum/Shore.

This is well known source for RV safety. https://www.amazon.com/No-Shock-Zone.../dp/0990527913

His video about generators bonding

https://www.youtube.com/watch?v=M-bTLdMjuqU

[booster]

Interesting that the cheapo outlet tester will find it easily, didn't know that.


He didn't address the double bonding but he was talking about plug in to use generators.


The grounding plug is essentially what I did with the tiny 120 watt inverters which don't even have room for a grounding plug and come with two prong. I removed the plug and installed a 3 wire cord to an outlet strip, and when I wired it I connected neutral and ground inside the little inverter. Works great and is never on any outlet other than the strip.

[Nic7320]

Hold on. Now it gets really tricky.

If your secondary inverters are galvanically isolated with internal isolation transformers, then they need their own bond to their own neutral or they also end up floating. But those bonds should NOT be switched off while on shore power, because the shore power bond is not in the secondary circuit, it's on the primary side, before any isolation transformer. Switching it off causes the secondary circuit to float.

So maybe the rule should be rewritten as "every transformer secondary gets one neutral to ground bond."

That might actually make things easier.

[booster]

Originally Posted by Nic7320

Hold on. Now it gets really tricky.

If your secondary inverters are galvanically isolated with internal isolation transformers, then they need their own bond to their own neutral or they also end up floating. But those bonds should NOT be switched off while on shore power, because the shore power bond is not in the secondary circuit, it's on the primary side, before any isolation transformer. Switching it off causes the secondary circuit to float.

So maybe the rule should be rewritten as "every transformer secondary gets one neutral to ground bond."


On edit: I think there would be some other stuff going on that way. When we had to do isolation transformers on sensitive equipment on work the outputs could never be grounded except to the equipment chassis, IIRC, because if you use the standard AC grounding setup you could get feedback noise and defeat the purpose of the isolation transformer. The problem in the van is that the if you bond the output of the isolation transformer in an isolated inverter, you tie it to the the chassis, which on shore power is going to be tied to shore power neutral and dirty power, defeating the purpose of the isolation transformer. I am not an electrical engineer, but I think to do the output isolation you would need to switch the ground of the chassis off of the shore power ground. Of course you then would give up the safely of having an earth ground path for the chassis when on shore power so the chassis could go hot with a short. There may be more involved than that, but at least on the Magnum and others I have seen, the neutral goes straight through on shore power and is bonded and through on inverter so no isolation. As we saw in George's video link, the floating output of the needs bonding to work, but there is no shore power feed through on that and the output would be going through a transfer switch if permanently installed, like an Onan which autobonds and then goes through ATS and the inverter/charger treats it just like shore power. I assume the Honda with bonding plug would be the same if it was hardwired in.



I guess I think you would need to isolate the grounds for the outlets off the chassis is you want to run isolated AC power on inverter with bonding on the output, but then you would need to tie the outlet grounds back to chassis when on shore power for safety. If it were an inverter only, not a feedthrough to the same isolated output, you could leave the outlets on the chassis as long as the inverter chassis wasn't grounded, but I don't think you could with AC feedthrough and shared grounds.


That might actually make things easier.

Why would the output of an inverter, isolated or not, be in the circuit anyway? Or for that matter the primary either. That is what transfer switches are for? If the output is to the 110v of the van, that is where there shore power is going to wind up, isn't it? If so there would be two bonds in the system, I think.

[Nic7320]

Now I see the problem. I had to draw out a diagram to see where the pathways are.

If you have an isolated outlet string that does NOT cross-connect to the neutral of the RV (like running an extension cord from an unbonded inverter to an appliance), then it's floating, and needs its own neutral to ground bond full time because it does not benefit from primary side bonds. It's really a separate circuit.

But if the neutrals are tied together, then you can only have one bond on that secondary winding. And when I say "secondary winding," that can be the park's distribution transformer, the distribution xfmr going to your home, a built-in RV gen set or an inverter. Each can have a secondary winding.

But that leaves a question; "Are there any inverters that don't isolate the outputs?" I cracked open my little cigarette lighter inverter and it has a transformer, but that still doesn't guarantee it is isolated. So I tried checked resistance from 12V input side ground to... "the AC electrical ground," but it doesn't even have one! There's a hole for a ground pin that doesn't go anywhere. Not even a contact.

Where it can get really messy is when a new TV or something is added to this supposedly isolated circuit, one can mistakenly connect a cable to something on the other (primary) side and that's when sparks fly. Or it quietly destroys your HDMI input on something plugged in elsewhere.

So this goes back to the point I was making, most inverters and some generators under 2000 watts ignore bonding, They're playing Russian roulette.

[booster]

My guess is that the inverters do float in general. That is the part I don't understand about the Magnum description where they say the bond needs to be released or you can get a shock hazard because of voltage difference between the inverter/van neutral and shore power when plugging in. There would have to be a delay built in to the inverter charger that makes sure to shut off the inverter with the transfer switch and unbonds the neutral before the shore AC is fed through or that difference wouldn't go away when plugging in. I do know the Magnum does not start immediately when you plug into shore power as it checks the power quality before engaging, so that may also be when the are doing the delay and unbonding before connecting to the output.

[Nic7320]

What they're saying makes sense. The Magnum internal bond needs released before you connect to another bond, or you end up with neutral current running through your ground wires and the grounded chassis of the RV.

Once current runs through your electrical ground it can create a IR drop (voltage difference) relative to another ground. Like the one you're standing on.

Some DC to DC converters are isolated and some are not. Considering all the complications and safety issues with AC inverters, I would be shocked (no pun intended) if anyone sells inverters that are not galvanically isolated from DC input ground to AC output ground. But I don't know that for a fact.

Some manufacturers are pretty ignorant when it comes to proper design practices. And they're not always east Asian imports... Early in my career, I worked for an electrical safety device company in California that had an engineering manager that could not do grade school math. So I was caught between our engineering manager and the customer's engineers at Square D who would preface things by saying "when you're on the witness stand, how do you prove... (this design was done properly.... )" The same engineering manager did not want to design in thermal disconnects. They are now required under C62.41 or UL1449.

A Cal-OSHA rep told me "most safety regulations are made in hindsight."

[booster]

No doubt about the make/break on the bonds and ground current, but they specifically say voltage difference on the neutral. The shore power neutral is at earth as it is bonded at the post, so the difference has to be all in the inverter and would have to be fairly large to be of hazard. In the industrial safety stuff we had to take arc-flash protection only after 50 volts.


The confusing part is that you always have a voltage difference on the hot leg when you plug in and that doesn't worry anyone, so it would likely have do with something related to exposed plug contact, but you would also have hot on the hot line, too.



It would be interesting to take voltage readings of hot and neutral, and ground also on the end of the shore cord and see what you get. I would expect that the hot would be zero as the transfer switch wouldn't have switched yet if the inverter is on, and neutral and ground the same at some higher voltage because they are still bonded?

[avanti]

Originally Posted by Nic7320

A Cal-OSHA rep told me "most safety regulations are made in hindsight."

All building codes and other safety regulations are essentially an accumulation of lessons-learned, usually the hard way. The failure modes that motivate them are generally quite obscure. If they were obvious, nobody would do them in the first place. Most of the stuff in the NEC is the result of somebody dying. Exit doors in ballrooms didn't open outward until a panicked crowd fleeing from a fire surged forward and couldn't open the inward-opening doors.

They say in the aircraft industry: "The regulations are written in blood."

[Nic7320]

Originally Posted by booster

... you always have a voltage difference on the hot leg when you plug in and that doesn't worry anyone, so it would likely have do with something related to exposed plug contact, but you would also have hot on the hot line, too.

It would be interesting to take voltage readings of hot and neutral, and ground also on the end of the shore cord and see what you get. I would expect that the hot would be zero as the transfer switch wouldn't have switched yet if the inverter is on, and neutral and ground the same at some higher voltage because they are still bonded?

Sure, there's a voltage drop on the hot wire as well once current is flowing, but that's the side nobody ever comes in contact with. Or at least let's hope so, on our rigs. (meanwhile,' I'm rewiring an IEC power inlet on my charger)

Because neutral returns connected to safety grounds have potential for human contact (or pets), any voltage there relative to another ground becomes a problem. So we don't want any current flowing in electrical safety grounds.

[Nic7320]

Originally Posted by booster

...In the industrial safety stuff we had to take arc-flash protection only after 50 volts.

Lowest recorded electrocution I know of took place from a standard "D" cell battery. Subject put pins under the skin of his fingers and tried to measure the resistance from arm to arm. Turns out some early Simpson 260s do put out 100 mA on low ohm range and that's all it took. He probably didn't get his reading.

Another electrocution took place on 12 volts. Victim was trying to retrieve something from an electroplating tank. His hip waders developed a leak and as he exited the tank he reached over to the bar that had 12 volts on it. Again, this was extraordinary circumstances where a highly conductive solution had contact with his... let's just say... vulnerable parts.

So 50 volts assume you have a healthy unbroken and fairly dry skin layer to protect you.

[GeorgeRa]

Originally Posted by Nic7320

Lowest recorded electrocution I know of took place from a standard "D" cell battery. Subject put pins under the skin of his fingers and tried to measure the resistance from arm to arm. Turns out some early Simpson 260s do put out 100 mA on low ohm range and that's all it took. He probably didn't get his reading.

Another electrocution took place on 12 volts. Victim was trying to retrieve something from an electroplating tank. His hip waders developed a leak and as he exited the tank he reached over to the bar that had 12 volts on it. Again, this was extraordinary circumstances where a highly conductive solution had contact with his... let's just say... vulnerable parts.

So 50 volts assume you have a healthy unbroken and fairly dry skin layer to protect you.

These are gargantuan corner cases. 12V from an electroplating, wow, there would be a lot of stochiometric ratio of hydrogen and oxygen mix, good for an excellent boom boom. Perhaps they were electroplating kryptonite, it needs a lot of oomph.

[Nic7320]

Originally Posted by GeorgeRa

These are gargantuan corner cases.

Absolutely agree on that. But ask a surgeon about electrical safety with medical instruments and that's where some of those corner cases take place.

Our skin keeps our body's electrical resistance above a few thousand ohms and it 'takes about 100 mA to cause ventricular fibrillation. So anything below 50 volts is generally considered safe. But once you contact something under the skin, that resistance turns into an aqueous solution of ions.

And maybe don't quote me on 12 volts... that was what was reported, but who knows if it's correct. A lower value really underscores my point, once the skin is breeched, higher currents can flow.

Back to inverter safety, before I forget. Electrical code (NEC?) allows use of GFCIs on ungrounded outlets, such as in older homes. In spite of its name, the "Ground Fault Circuit Interrupter" circuitry does not itself use a safety ground, it's simply looking for any difference in hot currents in and neutral return current out to decide to trip. The ground pin is just passing through the unit.

So if you're going to use a floating output or ungrounded inverter in an RV, a GFCI will give you some measure of safety for any connected wires, loads and people, but not necessarily if you touch the inverter case itself. It's better than nothing.

[booster]

Here is a link to a Magnum install manual:


https://www.magnum-dimensions.com/si...0Series%29.pdf


It goes through all the bonding stuff in text, but it was still unclear as to how the ATS worked with it all and the timing.


The diagrams on pg 42-43 show the unit in inverter and shore power configurations, including the internal transfer switch connections and bonding relay connections. They also describe the timing to an extent there, mainly that whenever shore power applied the unit goes into standby mode while it qualifies the power and then does, or does not, do the transfer switch and bonding movements.


As I suspected it would be, the shore power is a direct pass through and has nothing to do with the inverter section output, including bypassing the transformer, which on shore power reverses direction and is used a charger for the batteries.


The automatic bonding relay is in the neutral section of the ATS and actually does more than just bond the neutral and ground. The shore neutral and inverter neutral are tied together as a single output from the transfer relay, but when inverting the shore neutral is disconnected and the the output wire, which is connected to inverter neutral is switched to ground. Since when on shore power the shore AC neutral is connected to the same output that is for the inverter output, but apparently the rest of electronics switch off the inverter and turn on the charger so the transformer starts running in reverse, turning the inverter output into the charger input. This makes it a lot more clear how they could make all of it work, to me.



The inverter output when inverting goes through a transformer, but it is for voltage modification and not isolating, it appears, and shore power to the output is not run through a transformer at all so no isolation. I think this explains why there are not any issues with having both AC grounds, in and out, and the DC grounds all on the chassis as isolation isn't an issue and it works out that way easily.


It also gets more clear how they can seamlessly handle things like a glitch in shore power, I think. The AC input from shore would either be dead or the electronics shut it off, but if the inverter section is turned on when it happens, it should go right the inverting and then back to shore if the power came back or cleaned up. I haven't tested that at all, so not certain.


The way the shore neutral is turned off and the bonded made by a single contact makes it essentially impossible to have the two bond problem, which certainly would make the safety guy, engineers, regulators, and insurance company happy.


This doesn't answer about the "what if" of an isolation transformer in the circuit, but I think to be that type of setup, you would need to also isolate the grounds after the inverter when you bonded it, it you really needed clean power isolation. That would make it a standalone power source, I think, in the eyes of code, just like a standalone in inverter with it's own outlets that is bonded internally, which is exactly what we have with our two small, standalone, bonded inverters.

[avanti]

Here's a nice, simple diagram from my Outback installation manual that shows how a system with the inverter plus shore power plus genset is supposed to be set up. The "double-pole/double=throw switch" is an external transfer relay. Nothing new, just perhaps easier for folks to follow:

Untitled 9.png

[Nic7320]

That's an amazing manual, and rare to see something that is that thoroughly written. But after reading a bit, it made sense why they do all the things they do, they are catering to the marine market that has some extremely harsh requirements.

I heard of a guy with a brand new boat he left in a slip, not knowing he had an electrolysis problem with the dock's grounding system. By the time he returned to his brand new boat, a leakage current had reverse-electroplated his outdrives into a matrix of spongy metal. After they pulled the boat from the water, they found you could actually squeeze oil out of the outdrive cases they were so deteriorated. Of course, he wanted to sue everyone. MerCruiser makes an active protection circuit (MercCathode) to prevent this sort of problem.

But something also looks really familiar with that sloped shape of the Magnum box, take a look at Spartan Power (Reno, NV) and tell me if Magnum is rebranding the Spartan product line. I met the Spartan guy when I visited Inverters R Us, his retail distribution hub, and he seemed to know what he is doing.

It's not clear why Magnum needs to use a separate bonding relay. Fig 2-20 and 2-21 refer to "neu-gnd relay K1," but a simpler solution is to permanently bond every neutral at the source (like booster suggested) and use a 3-pole double-throw transfer relay to isolate each source when it goes offline (switching line-1, line-2 and the neutral with its ground bond all at once). So each neutral bond to frame ground connection is not a double bond when the transfer switch isolates it entirely from the RV circuit.

The battery charger circuit in Figure 2-19 is not much different than an isolated inverter circuit, since DC chargers have isolation transformers in them as well. It shows the output side needs a DC ground to frame bonding connection, but of course, most RV house batteries are tied to ground (except the upper 6 volt batteries in a series 12 V string). And inverters REALLY need this bonding, since they are internally pulse-width modulating at kiloHertz rates and generating high frequency noise.

So it's generally a bad idea to leave any circuit floating, and that's why ground bonds are there. When a circuit is floating, special precautions should to be done to ensure safety that are beyond the scope of most of us, yet that's what most inverters sold without bonding systems do, and plenty of other consumer devices as well. (Drills with two pin cords are a floating load since they don't have a ground, so they rely on double insulation to be safe).

And to something I said earlier, if you mistakenly connect neutrals together from the RV shore power side to the output side of an inverter, that will not drive any current through a bond connection, since the only return current path for the inverter output is only back to the inverter's isolation transformer itself. Again, it's an isolated circuit so the only way to get current to flow anywhere (to the input side) is to make two connections, and a neutral connection with ground bond is only one.

So have we totally hijacked this thread into a bonding topic?

[Nic7320]

I've been thinking out this failure issue with MSW inverters and the spikes booster saw.

If you only fit one or a few wave cycles onto an oscilloscope screen and still see spikes, those are real and not measurement artifacts (from probe skew or ground bounce). Skew generates tiny nanosecond delays and won't show up when you've got milliseconds shown on your screen. Same with ground bounce, it is created by fast rise-times through an inductance in the circuit, and those are usually small compared to the longer waveform you're looking at.

So it appears the MSW wave forming cycle is actually this: 0 volts, off, 170 volts, off, 0 volts, off, -170 volts... and then it repeats the cycle. And it's the "off" state that creates this problem. The "off state" is there to prevent any shoot through currents in the H-Bridge switching transistors; they are given extra time to turn off before the next transistor is turned on. And in sloppy designs, this "off" state means the voltage becomes uncontrolled in between pulses.

So is there any energy during these spikes that can harm the load? From the inverter's perspective, no. It's "off"... or at least it thinks it is off. But once you add any inductive load to the inverter, the current in an inductance doesn't cease during the "off state," its voltage flies up to meet this current demand. That's what generates the spikes, and also why the spikes are suppressed when the inverter is heavily loaded.

Catch diodes can contain any spikes, but fast switching Schottky diodes are expensive.

And the GeorgeRA's Nikon charger that got damaged, may vey well have become the suppression element, absorbing all the spikes. Was it the only load plugged in? If so, then it might have inductance built in to the charger itself.

My first thought was that an MOV based surge suppressor can help, but MOVs aren't designed for continual energy absorption more than one watt or so per disc. And they degrade with use, so relying on them for MSW spike suppression is a bad idea.

So I'll retract my statement that MSW isn't a problem. It's their "off-state" that is the problem. And PSW inverters don't have an off state, so they don't let inductive kicks get out of control.

[booster]

Can't say about Spartan, but they look like Magnum. They have different specs and are almost half the price, so unlikely the same guts. Sensata bought Magnum a while so maybe they own Spartan also. Magnum does private label, sort of, with Sensata equipment for Volta RV lithium systems,


I don't think you can do the backward transformer use without the automatic bonding as they both in the same box and you need one bonded, one not, plus direction change on the transformer for charging.


The combined inverter/chargers have some big advantages in space, simplified wiring, unified remote, etc compared to discreet components.


As far as the 3 pole connections are concerned, you don't really need all 3 by most accounts. The Magnum ground, of instance is always straight through whether on shore charging or inverting.


Many have just gone to adding a transfer relay on hot and neutral, which is readily available. There don't seem to be any 3 pole around any more, anyway. If you have a generator also, you would have shore and genny on first one, and the output from that one on second ATS along with the the output from the bonded inverter. All chassis ground go together because the neutrals can only bond one at a time.


Engineersrule had a long, long, discussion on here you may want to look at, including why the transfer switches are probably better than plain relays. Things like robustness, mobile ratings, on and switching delays for make break, power checking before connecting. Lots of steps he went through to get where he wanted to be.


Personally, I still think that the best thing for many RVs is to have a good quality PSW 3-400 watt full standalone inverter with it's own outlets, bonded without any connections to the chassis except DC input with a switch so you can use only when needed for things like charging electronics. You can get an Aims 300watt for $85 on Amazon, so with wire, a switch and a fuse you are under $100 total. Safe for everything and low parasitic even when in use. It will also run most TVs and DVD players these days of low power, so all good, no downside. We use less than 50 watts if on the TV and DVD together. Our biggest use is charging DW's laptop which is a gaming ASUS (not for gaming but is inexpensive way to get lots of capacity for reasonable cost). The brick is 180 watts, but when charging a dead battery and running non game normal stuff, if won't even pull 100 watts. If we want to run big stuff like the microwave or a hair dryer, then we just turn on the Magnum inverter for a bit. All the rest is done with the Samles 120 watt PSW inverters. We have two so one can be mounted in the audio cabinet and one in the area we have for charging stuff or running the laptop when using if we don't use battery. It is the norm to plug into the Samlex outlet instead of the van shore power even when on shore power because shore power can be much dirtier sometimes. These Samlex are tiny and of the more industrial line so a bit more money at about $100 each. The downside is only one two prong outlet, so I hardwired them with 3 wire cords to outlet strips.

[Nic7320]

Originally Posted by booster

The downside is only one two prong outlet, so I hardwired them with 3 wire cords to outlet strips.

Add a GFCI. They can be found as short pigtail adapters. Maybe cut off the ground pin off the male end (bad advice in any other circumstance, but dedicate its use only on this inverter).

https://www.amazon.com/s?k=GFCI+pigt...f=nb_sb_noss_2

[booster]

Originally Posted by Nic7320

Add a GFCI. They can be found as short pigtail adapters. Maybe cut off the ground pin off the male end (bad advice in any other circumstance, but dedicate its use only on this inverter).

https://www.amazon.com/s?k=GFCI+pigt...f=nb_sb_noss_2

I could, but decided against it. Certainly not a requirement or at least wasn't in the past as most of the outlets in the 07 were not GFCI. When on shore power or large inverter all are now GFCI, so feel plenty safe. The entire AC wiring for the small inverters is no more than a cord about a foot long, so very, very few places for issues.