Modified Sine Wave Converter

[Nic7320]

You have a point. It can work that way.

Just be careful if you connect any equipment on the inverter side to any RV grounds, as this floating side can accumulate some voltage charge.

And adding a GFCI only detects imbalances between hot and neutral, so it won't detect a common-mode charge built up from a floating condition.

THe problem is like this:

Imagine you're in a small boat, floating with a buddy in a pond. There's no dock lines or anchor tying you to ground. You and your (soon to be ex-) fishing buddy get in a fist fight and he throws you hard over to one side. The boat's momentum now has a slight push. You get up slowly, dust yourself off and admit you had it coming. But then you smash him harder and over to the other side. Now the boat has a little push that way. And now you two repeat this thousands of times per second (you're both superheros, by the way). So where does the boat drift off to?

Conservation of momentum in the boat's inertial reference frame doesn't guarantee the boat stays in one place. It's subject to any imbalance in the internal pushes to one side or another, and other influences such as water current and the wind. So the boat's position is uncontrolled and it's likely going to end up somewhere until it bumps into something hard.

Floating circuits do something like this, their common mode (the voltage of all its internal electronics and outputs relative to a completely separate ground) can build up a charge and float off anywhere it wants, until a component somewhere fails or the insulation breaks down. Or a human touches something.

And Inverters (AC inverters in RVs or any inverter in an air conditioners/heat pump or appliance) are different than the old appliances that only ran on 60 cycles, in that they have internally generated high frequency switches pushing things back and forth at 100 kilohertz rates. MOSFETs turn on hard, and then let go. And capacitance couples these signals into places nobody intended.

So the wind and current in this analogy -- are the various leakage currents from other loads because nothing is a perfect insulator. Even an isolation transformer lets a little leakage through.

A neutral to ground bond is generally a good thing when you have a ground, as is a ground to chassis bond. It keeps things within a safe voltage range. Obviously, with all the floating inverters out there, the common-mode charge injection issue must be somewhat limited since they're not routinely zapping people.

YMMV.

[booster]

Here is a link to the inverter specs.


https://samlexamerica.com/documents/...-1119_Hrez.pdf


UL and hospital safety approved for use with no ground as they have no ground on the outlet. To make the 3% voltage spec there really can't be much creep


This is there PST series which has been around a long time in commercial use, I think. There newer, lighter, much less expensive lines are likely not this consistent or robust.

[avanti]

Originally Posted by Nic7320

Conservation of momentum in the boat's inertial reference frame doesn't guarantee the boat stays in one place.

Well, if the influences are stochastic, then Pólya’s Random Walk Theorem does guarantee that the boat will always return to the original location eventually. There is not theoretical limit to how far it can roam, but pretty quickly, the probabilities become insignificantly low.

Consistent drift requires some kind of asymmetry or an external influence. Your "water currents and wind" would be examples, but they are external influences. I am having trouble seeing what kind of internal processes might exist in the system under discussion that would have the necessary asymmetry to build up charges without bound. Could you give some examples?

[Nic7320]

If the high side MOSFET impedance is not the same as the N-channel low side switch, it's not perfectly balanced swing. Add to that any slight timing error in the gate drive circuit that creates a fatter pulse on one leg that's not mirrored on the other. And then there's charge injection from something that swings quickly with a fast rise time but decays slowly creating asymmetrical crosstalk in the circuit board layout. One side won't have the same exact capacitance as the other side. Any inductance in the supply rails also cause the voltage on the rails to bounce.

So it's a noisy circuit. And then there's a small possibility charged particles from space come raining down on the circuitry. Something we have to design for at higher altitudes.

In a circuit full of active devices that are inherently non-linear, combined with uncontrolled leakage currents and device tolerances that are not exact, some rectification effects can add up.

Analog computers had this limitation, as thermal noise would drive the op amp integrators to one end before a computational result was reached. The noise wasn't perfectly gaussian and symmetrical as theory would lead one to believe.

in a perfect world, there is no difference between theory and reality...
But in reality, there is.

[Nic7320]

Actually, now that I think about it, leakage current across an isolation transformer tends to drag the floating side down towards the primary side potential, rather than sending it adrift.

So two wrongs actually make a right there.

But AC loads are usually non-linear, since almost everything converted to DC is done with rectifiers.

[avanti]

Originally Posted by Nic7320

Analog computers had this limitation, as thermal noise would drive the op amp integrators to one end before a computational result was reached. The noise wasn't perfectly gaussian and symmetrical as theory would lead one to believe.

I do not believe that thermal noise is not perfectly random in the sense implied by this claim.

That said, and despite how much I have enjoyed this discussion, I'm afraid I have to put my moderator's hat on and declare this thread to be off-topic. Let's get back on track, please. Any new messages predominantly about physics or statistics will be deleted. Sorry for claiming the last word.

[Nic7320]

I was hoping you'd say that. I don't care to speculate about the physics, it's not my forte.

What I described is part theory, part conjecture, and a lot from from years of practice. Good engineering does not leave things to chance, which is why floating circuits should be avoided.

I briefly mentioned non-linear loads as a possible source of charge accumulation. So let's look back at the boat analogy. If a pendulum swung around in a circle from a floating boat is likely to leave the boat in one place IF the air it encounters is the same in every direction. So sine waves (which are elongated circles) are well behaved when powering purely resistive loads.

But non-linerar electrical loads encounters a different resistance at different parts of the cycle. Take diode rectifiers used when converting AC to DC. They only conduct at the very top of a sine wave, and you see this in the capacitor's ripple voltage as the bump when the cap charges up. Then as the waveform drops, the diode falls out of conduction and the whole circuit has to wait for the next sine wave to reach another peak. The current going into a rectifier looks like a series of short blips, and nothing like the sinusoidal voltage driving it.

So this is like the pendulum bumping into a wind current pushing it one way only as it passes through part of its rotation. And the boat was perfectly well behaved before the load (the pendulum) started swinging (ignoring any drift from the boat's own windage). So this just demonstrates that loads determine the impedance the power source sees. If it's not symmetrical, it's a little push to one side.

And the internal works of a Switched Mode Power Supplies used in most modern power converters now days, are more complicated, since they have lots of noisy, fast-switching pulses that are used to reconstruct the relatively slow 60 Hz sine wave output. These often switch at 20 - 200 kHz.

So, staying on topic here, I have hesitation when anything is ungrounded, unbonded and floating. You're relying on a lot of unknowns and just don't know where it will end up.

[booster]

Going way back to the OP's original question about charging their computer.


I would give same and repeated advice, still. Get a small PSW inverter and run it as a standalone for critical electronics. Most will have a GFCI so quite safe. The big issue is the MSW they have and they got really lucky the computer just won't charge and didn't get anything fried. Just use a 12v accessory plug if you have one or hard wire it in with a switch. I would mount it on something not attached conductively to the van body though if possible to keep it totally isolated electrically.

[Nic7320]

Some interesting GFCI discussion here:
https://www.promasterforum.com/threa...nverter.90127/
and here:
http://maze.airstreamlife.com/wp-con...h_Note_018.pdf

GFCIs are like fuses, they aren't "needed" under normal operating conditions, but they provide extra safety for when things go awry. NEC requires them in some locations and not others.

From what I've seen, very few inverters have GFCI, yet most portable generators do. Of the seven inverters I've had, only one (maybe) had a GFCI, and it went with the PleasureWay when I sold it so I can't verify that. I don't recall the brand.

The Samlex PST-120-12 looks good since it's hospital rated, but it does not have a GFCI. That probably means they have extra isolation to battery ground to get that rating.

Victron Phoenix inverters, EverStart inverters, and my big 2000 W AIMs inverter do not have GFCI. I did install one on my Tesla battery/inverter system.

GFCIs detect current is going through an unintended pathway. This provides an extra measure of safety from electrocution, but it can't prevent other types of faults.

GFCIs give us one more reason to use Pure Sine Wave inverters. GFCIs don't like ratty power.

I had an aquarium light catch fire while on a GFCI. Salt had apparently accumulated in the light socket and bridged the contacts. The socket completely burned away and fell behind the cabinet, which, needless to say... could have been much worse. I woke up in time to put it out. Took four hours to find out why there was a fire.

So why didn't the GFCI catch the problem? Answer; Because there was no ground fault. The arcing was entirely from line to neutral and there was no other pathway to create a ground fault. Electrically, it behaved no different than a hair dryer or any other load.

When I talked to TWO salt-water aquarium stores, they both had similar stories about fires. And look at the Conception dive boat disaster that took so many lives. Likely started in an electrical outlet that was arcing.

What's desperately needed are AFCIs, (Arc Fault Circuit Interrupters) made into pig-tail adapters, so we can easily add one where we need to. AFCIs are comnonly available as circuit breakers, but those need an electrical circuit breaker panel and (technically) an electrician to install one.

So maybe the new combo-AFCI-GFCIs are a good idea in RVs.

[booster]

Originally Posted by Nic7320

From what I've seen, very few inverters have GFCI, yet most portable generators do. Of the seven inverters I've had, only one (maybe) had a GFCI, and it went with the PleasureWay when I sold it so I can't verify that. I don't recall the brand.

The Samlex PST-120-12 looks good since it's hospital rated, but it does not have a GFCI. That probably means they have extra isolation to battery ground to get that rating..

What? You need to get out more.


Nearly ALL of the small, like 350 watt and smaller, ones suggested for the standalone to the OP have a duplex GFCI breaker on them and many up to about 1500 watts do. Take a look at any of the retail inverter sellers to the mobile market.

[Nic7320]

I don't disagree! Getting out more is one reason I have an RV!

A quick search on amazon for "Inverter with GFCI" shows some big ones have GFCI along with a couple Samlex 300 watt inverters, but then it devolves into a random mix of others without GFCI from their fuzzy search algorithm:

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

Then searching for "Inverters" alone yields hundreds of listings without GFCI:
https://www.amazon.com/s?k=inverter&ref=nb_sb_noss_2

Even the one you posted lacks GFCI, but other than that, it looks pretty well designed.

Small ones just don't have room for a GFCI duplex outlet. GFCI devices have two switches to test and to reset. A square grounded receptacle does not alone mean it has GFCI.

So without malice, I say it could mislead readers to suggest most have GFCI when there are so many that do not. Now if you know of a vendor that only carries well-designed PSW inverters that all have GFCI, maybe tell us.

[engnrsrule]

Hello - I also had problems with the limitations of modified sine wave inverters. I was also bothered that the inverter's output was directed to only some plugs and not the full 110 bus (realizing of course the inverter could not power large loads). I just wanted to plug in a computer or charger at any outlet.

Check out my thread for details. I have attached a schematic of my mods here. Also have one correction...My shore tie would routinely trip any GFCI source outlet to which I would connect. A thread I had seen in the past indicated this was possibly a nuance of the tripplite. I have since discovered that a power strip with integral USB charge ports I was using was in fact wired with hot and neutral leads reversed. When I corrected this the GFCI problem went away.

Here is the thread.
https://www.classbforum.com/forums/f...ade-10087.html

Biggest challenge was packing more equipment into very limited space while ensuring adequate space for cooling.

[archeogal2]

I have read all this with interest but with lots of holes in my understanding. However, I don't really want to understand all this; I just want my electrical stuff to work when we're in the van! I know many of you might disapprove of that approach, but. . . Anyway, my question relates to a small inverter to replace a 16 year old Nexxtech 175 watt inverter that we have been using all those years. It's OK, but gets hot, and after reading what has been written above, I am sure I can replace it with something more efficient. I'm looking at the Samlex PSW 120 watt. So, is that large enough since we would use it mainly for phone charging, computer, and printer? And do I really need a GFCI outlet? And, are you people plugging these small inverters into the engine battery or the coach battery? We have a 12v plug in the rear of our van that hooks into the coach battery. We are powered with 300 watts of solar on the roof and a 3000 watt MSW GoPower inverter. Or we could of course plug it into the outlet in the front which goes to the engine battery.


Thank you for any advice you can offer.

[Nic7320]

I would recommend a Victron Multiplus Inverter/charger for all the reasons we just went through. It has an ATS, it is PSW, and it's built like a tank. It also has a nice bluetooth interface and can be adjusted to limit power from a shore power source, which is useful if you have a small generator or camp where the power distribution isn't so good. So if you're looking for a well engineered all-in-one solution, look at the Multiplua.

Booster is more familiar with some other brands such as Magnum, which is also good, so ask him as well.

[engnrsrule]

Originally Posted by archeogal2

I have read all this with interest but with lots of holes in my understanding. However, I don't really want to understand all this; I just want my electrical stuff to work when we're in the van! I know many of you might disapprove of that approach, but. . . Anyway, my question relates to a small inverter to replace a 16 year old Nexxtech 175 watt inverter that we have been using all those years. It's OK, but gets hot, and after reading what has been written above, I am sure I can replace it with something more efficient. I'm looking at the Samlex PSW 120 watt. So, is that large enough since we would use it mainly for phone charging, computer, and printer? And do I really need a GFCI outlet? And, are you people plugging these small inverters into the engine battery or the coach battery? We have a 12v plug in the rear of our van that hooks into the coach battery. We are powered with 300 watts of solar on the roof and a 3000 watt MSW GoPower inverter. Or we could of course plug it into the outlet in the front which goes to the engine battery.


Thank you for any advice you can offer.

There is a lot there...

I am not understanding how you are set up. Sounds like you have a low wattage (175) inverter powered off your engine battery, and a 3000 watt MSW inverter powered off solar. I'm not an expert on solar but that seems odd. In my understanding solar would not go straight to an inverter or service loads, it operates through a controller to charge the coach batteries. The coach batteries in turn power the 12v plugs and the inverter serving the coach space. The 3000w inverter is the one I would consider upgrading to PSW to better serve your boondocking power needs. The smaller inverter on the engine battery would seem to be just for using computer or charging phone while you're going down the road. While it may technically work when stopped, realize that engine batteries generally take less kindly to sustained drains - they are designed for brief high amperage spurts, while coach batteries are typically deep cycle for sustained continuous drain.

As for GFCI, my Roadtrek is set up like a house. there are GFCI in vicinity of sink and head, while other 110v receptacles are non GFCI. I do not see any disadvantage to making all outlets GFCI, but do not think it is a requirement outside wet areas (or exterior outlets).

[Nic7320]

Originally Posted by engnrsrule

Hello - I also had problems with the limitations of modified sine wave inverters. I was also bothered that the inverter's output was directed to only some plugs and not the full 110 bus (realizing of course the inverter could not power large loads). I just wanted to plug in a computer or charger at any outlet.

..My shore tie would routinely trip any GFCI source outlet to which I would connect.

GFCI tripping can be from high frequency (from the stepped waveform's fast edges) noise coming from your MSW Inverter, or perhaps you commingled your neutrals (is your Triplight internally bonded from Neutral to Ground? That can be the problem)

Once you go through a GFCI, that neutral should be kept separate from any shore power neutral. All loads should return current back through the GFCI.

The other thread does not talk about a secondary power systems which you may have have inadvertently created. Once you go through an isolation transformer (which inverters often have), you have an isolated output and IF it is kept separate from your primary power grid (with no common neutral connection, perhaps by running loads off extension cords or plug strips) then it's a separate isolated power system, needing its own neutral to ground bond and electrical ground to chassis ground.

But when you go through an inverter's isolation transformer, you still have the option to keep it isolated into a separate power string, or to defeat the isolation by combining it back with your shore power side neutral. If you combine it, it should follow the SINGLE bonding scheme used on your shore power side.

[booster]

As I mentioned earlier, I would not get a large auto transferring inverter to do light device charging. We have a bit Magnum with all the features and we use small Samlex 120 watt ones to charge our small things and run or TV and DVD.


To answer the questions. GFCI is a very small amount safer but only if the case of the inverter of the inverter is tied to their neutral, which is probably not done on decent inverters. Even then, the only shock you could get would between the hot wire and the inverter case so nothing else in the van would be a hazard unless the unit was sitting on the van bear steel and the cover was not painted. If the unit without GFCI is on a non conductive surface the risk is extremely small, IMO. The Samlex is a good inverter and is also hospital use rated so it will almost certainly not have any case or such issues and will be very safe. It also has 3% accuracy power so better than the 5% most have.



We run all our charging through the 120 watt Samlex and it does it easily. The biggest users is the laptop which uses 90 watts when low battery and not in use. When in use it goes up a bit over 100 watts when on low power stuff that isn't using lots of video stuff. Doing the laptop by itself is normally a thing we do so we don't go over, but phones are not as much.


I think Dan Rowe has them shown for about $100 or so.


Quite a few folks have gone to small PSW inverters for light duty use and low parasitic when not in use, if left connected. IMO, it just makes sense.

[Nic7320]

Maybe this needs a more step-wise description.

Once you send power through a GFCI, it is now a separate power string with its own dedicated neutral. Do not connect it to other neutrals (unless its AC output is well isolated). The GFCI must have the same current returning on its neutral as sent out on the hot side, or it will trip (because that's what it's designed to do).

Some inverters have internal GFCI breakers, so those need special attention if you're commingling their outputs back to the primary side. This is where it gets messy.

Inverters with isolated outputs also need their own bonding and grounding. Most I believe are isolated, but I haven't done an exhaustive survey. YMMV.

Does anyone know the electric code for commercial use or farm use where step up or step down transformers are used? That should provide some guidance what to do with secondary side transformer bonding and grounding.

[avanti]

Originally Posted by booster

Quite a few folks have gone to small PSW inverters for light duty use and low parasitic when not in use, if left connected. IMO, it just makes sense.

That is one of two viable approaches. The other is to aggressively make sure that all light-duty loads are native DC and to make it convenient to turn on the large inverter only when a large load such as the microwave, toaster or coffeemaker is needed.

Both of these approaches are practical. We have discussed the pros and cons previously, but either is fine.

[Nic7320]

Originally Posted by booster

To answer the questions. GFCI is a very small amount safer but only if the case of the inverter of the inverter is tied to their neutral, which is probably not done on decent inverters. Even then, the only shock you could get would between the hot wire and the inverter case so nothing else in the van would be a hazard unless the unit was sitting on the van bear steel and the cover was not painted. If the unit without GFCI is on a non conductive surface the risk is extremely small, IMO.

There's some confusion here as to the role of electrical safety grounds and what Ground Fault Circuit Interrupters (GFCIs) actually do.

An electrical ground (the third pin on a plug) is a pathway to drain away any unintended current from an electrical short so it doesn't find its way through a human. It's basically just a wire that connects to the metal cases of equipment.

A GFCI is another layer of protection. It is an active device with a differential current sensor and a relay to disconnect power during a fault. When it detects a fault, it disconnects power in a fraction of a half-wave cycle!

GFCI won't address every possible electrical fault, but they do a lot more than just protect against shorts to the case; their job is to ensure all current going OUT of the GFCI comes back IN to the GFCI on any connected wiring or loads connected on that string. Otherwise, the current is going somewhere unintended, like through YOU.

You can identify a GFCI by its two buttons TEST and RESET. Any additional outlets connected to the same GFCI SHOULD get a little sticker that indicates they are also GFCI protected (but this is often not done).