PV Hybrid Inverter Technology

[cods4]

Hi All
I am designing a hybrid solar power system, which is made up of solar panels, hybrid inverter, and batteries, and can send/receive power to/from the grid.
The local standards require that my inverter has simple separation between the AC grid connection, and the 48V DC battery terminals. And they say one example of simple separation is a transformer.
There are a plenty of different solar inverters, but they all seems to either be high-ferequency or low-frequency type. The low frequency ones use a large transformer between the AC grid connection, and the inverter circuitry, so this would obviously provide the separation between the AC and DC.
However these LF inverters are far more expensive than the HF alternative.

My question is around the HF inverters and how they operate. I understand they use a switching frequency in the order of 40kHz, and they put this through a small transformer to efficiently transform the voltage to the correct level. And I'm trying to determine if this transformer would provide simple separation between the battery DC side, and the AC?

I have found  some photos of the inverter I'm planning to use which show the small toroidal transformers. But I can't follow the traces and work out what is going on exactly.
I'm hoping that someone who understands inverter design could identify the circuit type and let me know if this should provide separation between the AC and DC or not.

https://fccid.io/2AV6USG01LP1/Internal-Photos/Internal-Photo-4723808

[uer166]

How much power electronics design experience do you have? Such a thing would take a few years of some serious study and trial and error, with 10's of k$ spent on prototypes and test gear. Based on your questions and wording I assume not much, although could be wrong.. There really is no such thing as a "low frequency" inverter anymore, at least not anything modern.

What's the power levels that you need? What in/out voltages? What standards are you making it to? Will it be grid tied? Islanded? What will the pack look like? BMS? Why 48V and not something more appropriate?

[NiHaoMike]

I would go for a 400V or so battery, overall costs are going to be lower unless you're only looking for a little capacity. Keep in mind the PV strings are typically around that voltage so that'll make things easier on that end.

You can consider a Prius inverter as a starting point, they're available for around $100 and already take care of the isolated gate drives for you. They also have a large bidirectional buck/boost converter which could be used for interfacing to a somewhat lower voltage battery around 200V.

[cods4]

Sorry guys, wasn't clear enough in my first post.
I'm designing the system using off the shelf inverters, panels etc.

I'm just trying to determine if that Deye inverter in my Link, and similarly designed inverters have electrical separation between the AC and DC terminals.

I wouldn't have nearly enough power electronics experience to design an inverter from scratch.
Electrical engineer by trade. But I work with big stuff not electronics very often.

[T3sl4co1l]

The toroidal transformers are all common-mode inductors.  The two-winding one (page 3) I'm not sure about, it's much too fine wire to be handling much power (and doesn't seem to be wired to much of a load?); the rest are three windings, so, I take it this unit is made for 3-phase conversion?

Wait, four pins on the far side, I think that may be a potential transformer (i.e., used to sense the voltage on the primary side, by a calibrated ratio), that could explain the high turns ratio.  Still, no idea why the sectoral wind.

Also the nearby (quick connect) terminals are labeled PVn(+/-), which sounds like DC (photovoltaic panel connections?).  Can't tell how they're connected to the transformer-thing.  So, dunno.

Anyways, the smaller E-core transformers are probably for power to various subsystems, up to a hundred watts or so (bottom p5, yellow rectangular thingies).

If there's any mains isolation in this thing, it's probably the potted mess on page 4.  The boards are all about managing, handling, monitoring, switching, etc. the PV and mains supplies, and the actual power conversion connecting between them, I'm guessing, was wired here.

Without labels on the lugs, and seeing where they were bolted to, I can only guess this is the case.  The box does seem long enough for three transformers, which would be needed for the three phases.  There should be some filter inductors somewhere, or maybe they're internal somehow (e.g. resonant converter).

Can't tell what top of page 5 is; backside of the transformers box?  Looks embedded in a heatsink, or maybe that's just a cutout.  Clearly bolts to the main power switching board (bottom p6 / top p7).

That WiFi module is all kinds of jank, looks hand made. Completely unwashed, gooped up with rosin.


As for ratings, what does the manual say?  What standards does it conform to?  You can learn 100% of what you need to find in there.  If it doesn't say it meets some standard, assume it doesn't!  Think like a lawyer -- this is a legal question more than electrical.  If it doesn't specify isolation, like, it shouldn't even be sold probably, and definitely won't meet your system requirements!  Now, you will have to check standards and what they mean, for that to make any sense, but such is the responsibility of the system designer.

Tim

[cods4]

Awesome, thanks so much for the detailed analysis.
I'm slowly going through it all and googling some of the terms you've used

I just did a bit more digging and realised this is the American version with split phase. Hence the L1, L2 and Neutral.
Although that still doesn't explain to me the 3 winding toroidal chokes.
But anyway, it doesn't sound like there's any obvious isolation between the ac and battery dc.

The toroidal transformers are all common-mode inductors.  The two-winding one (page 3) I'm not sure about, it's much too fine wire to be handling much power (and doesn't seem to be wired to much of a load?); the rest are three windings, so, I take it this unit is made for 3-phase conversion?

Wait, four pins on the far side, I think that may be a potential transformer (i.e., used to sense the voltage on the primary side, by a calibrated ratio), that could explain the high turns ratio.  Still, no idea why the sectoral wind.

Makes sense, they will need to monitor the voltage on these for the MPPT to work. They must also be measuring current somewhere too.

Also the nearby (quick connect) terminals are labeled PVn(+/-), which sounds like DC (photovoltaic panel connections?).  Can't tell how they're connected to the transformer-thing.  So, dunno.

Anyways, the smaller E-core transformers are probably for power to various subsystems, up to a hundred watts or so (bottom p5, yellow rectangular thingies).

If there's any mains isolation in this thing, it's probably the potted mess on page 4.  The boards are all about managing, handling, monitoring, switching, etc. the PV and mains supplies, and the actual power conversion connecting between them, I'm guessing, was wired here.

Without labels on the lugs, and seeing where they were bolted to, I can only guess this is the case.  The box does seem long enough for three transformers, which would be needed for the three phases.  There should be some filter inductors somewhere, or maybe they're internal somehow (e.g. resonant converter).

Can't tell what top of page 5 is; backside of the transformers box?  Looks embedded in a heatsink, or maybe that's just a cutout.  Clearly bolts to the main power switching board (bottom p6 / top p7).

Yeh, I didn't even notice there must be a bit going on inside that metal heatsink/enclosure.



That WiFi module is all kinds of jank, looks hand made. Completely unwashed, gooped up with rosin.


As for ratings, what does the manual say?  What standards does it conform to?  You can learn 100% of what you need to find in there.  If it doesn't say it meets some standard, assume it doesn't!  Think like a lawyer -- this is a legal question more than electrical.  If it doesn't specify isolation, like, it shouldn't even be sold probably, and definitely won't meet your system requirements!  Now, you will have to check standards and what they mean, for that to make any sense, but such is the responsibility of the system designer.

Agreed. One thing I find strange, these Inverters (or atleast similar ones from the same manufacturer) were approved by Australian clean energy council to say they meet the inverter standard AS/NZS4777.2:2015 (which i havent read). But when I read the battery energy storage standard AS/NZS5139 it says you shall not use an inverter without simple separation between ac terminals and battery terminals if your battery is in DVC-A category (less than 60v) which is exactly what this inverter is designed for.

Also, there are a lot of other hybrid inverters using similar designs in the same situation.

Tim

[T3sl4co1l]

Oh, split phase? Sure, that would do it too.

(There's also a 4-winding kind, for, you guessed it, 3-ph with neutral.)

Did you find the manual?  For some reason their website seems a bit broken for me, but I was able to guess the search key,
https://www.deyeinverter.com/download?search=SG01LP1

Diagrams seem to show battery being a standalone thing.  More likely it's tied with the PV circuit, not mains, but it seems unlikely in any case to be completely isolated from both.

Tim

[Phoenix]

While I haven't digested the full topology I'm 100% sure it's isolated battery input (and non isolated PV).

On page 4 the grey box of gunk contains 2 DC boost main chokes for PV and two AC inverter main chokes.

On page 5 is a weird PCB mounted into the house. That is is the isolation transformer. Lots of fat wire for the 190A battery rating. And temperature monitoring by the looks of it (or could also be a cancellation or sense winding).

On page 7 there is a hell of a lot of devices for a non isolated topology. Some must be for a DC DC isolator. Plus a bunch of isolation slots. If not isolated it's one hell of a boost converter... Not feasible!

On page 6 there are two large banks of capacitors. The small bank in on the isolated battery side.

Finally, if memory serves me correctly, since it's low voltage battery it must be SELV isolated by the standards anyway.

[Phoenix]

The two-winding one (page 3) I'm not sure about

I'm thinking it's arc fault detection. Mostly because the PCB label has "arc" in it.

Originally my mind went to leakage current but it doesn't encompass all the PV windings. The leakage detection is done in the AC filter area. Much more sensible as it protects the islanded load too.


If anyone is interested I can mark up the photos as best I know.

[T3sl4co1l]

The two-winding one (page 3) I'm not sure about

I'm thinking it's arc fault detection. Mostly because the PCB label has "arc" in it.

Originally my mind went to leakage current but it doesn't encompass all the PV windings. The leakage detection is done in the AC filter area. Much more sensible as it protects the islanded load too.


If anyone is interested I can mark up the photos as best I know.

Ah, that could do it too.  I'm unfamiliar with arc detection circuits; I'd have imagined just some AC coupling to the line.

Tim

[NiHaoMike]

I'm designing the system using off the shelf inverters, panels etc.

I would suggest MPP, they have the option of daytime backup without batteries, but (unlike SMA) allow adding batteries later on with minimal additional hardware.

Does your area have net metering? If not, you might be better off with a zero export system and bypass the hassle of exporting to the grid.

[TheMG]

I think it would in fact be difficult to find a PV inverter that doesn't have an isolated AC output. If there is such a thing, I've never encountered one.

[Phoenix]

I marked up the photos with some details. The transformer should be "3" not "4".

[Phoenix]

I think it would in fact be difficult to find a PV inverter that doesn't have an isolated AC output. If there is such a thing, I've never encountered one.

The opposite is true. Basically every domestic PV inverter on the market today will be non-isolated/transformerless - they are simply more efficient, smaller and cheaper. They do require additional protection such as redundant relays, leakage detection and array insulation resistance monitoring.

[TheMG]

I should specify, I'm referring to any PV inverter for off-grid use or hybrid (grid tie with battery backup).

No reason for the isolation on a purely grid-tie inverter of course.

[Phoenix]

I should specify, I'm referring to any PV inverter for off-grid use or hybrid (grid tie with battery backup).

Hybrid inverters (grid tie plus battery backup) also have non-isolated PV inputs. If you look at my analysis of the one in this thread I identify the DCDC transformer and isolation barrier... the PV inputs are on the HV (read non-isolated grid) side. This is also consistent with other hybrid inverters that I'm aware of.

[T3sl4co1l]

Oh, I was looking at this just the other day; that's probably a matrix transformer, and three not four of them.  12 transistors, 3 pairs of primary terminals, 2 secondary terminals + 2 jumper pads for 4 total nodes connecting 3 secondary windings.  No phase shift or anything (with the series secondary, everything has to be in phase), purely done to distribute the current over a wider area on the primary side.  Transistors must be H-bridge, I guess?

Tim

[cods4]

Thanks very much for the analysis guys. It's extremely helpful.

I'm out at site commissioning an HV switchboard today and tomorrow. So will spend some more time looking at all this over the weekend.

But what I'm reading is there is decent isolation between the battery and AC terminals,
Battery and PV terminals,
But not the PV and AC terminals.

So if I put a 500 or 1000v megger between the battery and ac terminals, I shouldn't let the smoke out?

I'm thinking I might have to get one of these to have a play with.

[Wolfram]

I think it would in fact be difficult to find a PV inverter that doesn't have an isolated AC output. If there is such a thing, I've never encountered one.

These exist, they are usually called "transformerless solar inverters". Depending on power levels and relevant standards, it can be cheaper to provide safety isolation on the PV side, compared to adding galvanic isolation between the solar array and mains.

[David Hess]

This page implies that the AC side is isolated from the battery and solar panels.  I do not see how it could not be galvanically isolated.

[Phoenix]

This page implies that the AC side is isolated from the battery and solar panels.  I do not see how it could not be galvanically isolated.

Are you referring to this?

"48V low voltage battery, transformer isolation design"

Yes, the battery is isolated.
No, the PV is not isolated (see the FCC teardown photos).

What's wrong with it not being isolated to the PV?

[Phoenix]

and three not four of them.

Oops, you're right. I corrected my picture in my text.

Transistors must be H-bridge, I guess?

I can trace the tracks to be a H-bridge, but something is missing. Those screw points must go somewhere and I don't see any AC coupling capacitors or current sensors to avoid flux walking.

[cods4]

Thanks for all your help guys, I have just heard back from the manufacturer, and they are saying they have isolation between the battery terminals and AC terminals, and this is achieved using a transformer. Unfortunately they couldn't provide me any drawings (not surprised).

[NiHaoMike]

Thanks for all your help guys, I have just heard back from the manufacturer, and they are saying they have isolation between the battery terminals and AC terminals, and this is achieved using a transformer.

Not at all surprising they're using a transformer for isolation, it's pretty much the only technology that can practically isolate that much power. A distant second would be a motor turning a generator, but that's basically a transformer split in two and then joined with moving parts.

[jonpaul]

Coes4, bonjour

1/ power inverters and PV design takes years of experience, and special instruments to debug. Beware of shock, arc, injury and lethal voltage and current.

2/ Toroïdal shapes are not used generally as PV inverters.
we used round center leg EER, PQ, and other modern ferrites, TDK or Philips /Epcos.
Bobbins are needed for isolation and shields.

3/ A simple solution,  if you have a hybride EV,  my old  friend Richard Factor's PriUPS.com site started a decade ago

http://www.priups.com/

Bon courage

Jon