RFI from Enphase micro-inverters

Hi,

I am considering having solar panels installed. The main non serial-string-of-panels systems seem to be SolarEdge and Enphase. Solaredge apparently causes RFI; that system uses switching DC-DC converters on the panels and a DC-AC converter inside. Enphase has micro-inverters that produce the mains voltage straight on the solar panel. As these must be switchers, has anyone experience with these with regards to RFI? I am mostly concerned for noise in the range 0-30 Mhz genaral, and 77 kHz specifically (clocks). Can any RFI be prevented by used ferrite clamps like those on mains leads?

Mat Nieuwenhoven

Reply to
Mat Nieuwenhoven
Loading thread data ...

On Wednesday, October 28, 2020 at 8:47:19 PM UTC+11, Mat Nieuwenhoven wrote :

Everybody has run into switching power supplies that generate radio-frequen cy interference. Quite a few of us have worked out how to design them and l ay them out so that they don't generate much. If you are stuck with other peoples designs, there are ways of confining most of the interference to th e immediate vicinity of the power supply itself.

?

Ferrite clamps can can reduce the amount of high frequency current travelli ng down mains leads, but nothing prevents that reduced current from radiati ng. It's not the only way the radio-frequency interference gets out of inve rters. I've run into an inverter that had to be wrapped in a grounded flywi re enclosure (Faraday cage) to stop it from interfering with everything els e electronic in the same rack.

There's one excellent book on the subject - Ralph Morrison's "Grounding and Shielding Techniques in Instrumentation"

formatting link
p/0471838055

I've got a copy of the fourth edition. You could buy the sixth edition.

You'd be much better off finding exactly how much interference the switchi ng inverters that you are thinking of buying spew out, and how they spew it out, but that isn't easy to find out.

--
Bill Sloman, Sydney
Reply to
Bill Sloman

I said "no thanks" to SolarEdge and installed passive panels with a Fronius inverter instead. It's quiet - with the extra ferrites I had installed in it, it's super-quiet.

Can't vouch for Enphase, but for $1000 worth of ferrites and a lot of trouble it's possible to make a SolarEdge system quiet. See the cover story on that in April 2016 QST magazine. It's what convinced me I didn't want to go that way.

Clifford Heath.

Reply to
Clifford Heath

Thanks, good reading. I think I avoid Solaredge.

Mat Nieuwenhoven

Reply to
Mat Nieuwenhoven

I spoke today to a saleman of a company installing solar systems using the Enphase system, and not only did he confirm that those micro-converters generate "some noise", but he immediately said they can fix that using ferrites. I was impressed. I shall get more details, especially after reading the QST article Clifford kindly mentioned. Apparently the panels themselves can also act as antenna somewhat.

Mat Nieuwenhoven

Reply to
Mat Nieuwenhoven

The panels will only act as an antenna if they have a signal to radiate. You will need ferrite between each micro-inverter and its panel to stop that.

The biggest radiator from a rooftop system is the large loop of wire that joins the string. In the QST article, you will see a recommendation that the outbound and return wires should be twisted together, or at very least, run parallel.

Regulations here in Australia require that the return wire is tied to the mounting rail - but my installer wanted to install the panel-to-panel links directly between the panels, about 250mm from the rail. I insisted that each link was returned (as a pair, preferably twisted) to the rail, then twisted along the return wire to the next panel (where it meets the link from a further panel, etc). As it turns out, the message got lost in translation and they didn't do that - but because the panels are passive there is no significant signal to radiate.

In the QST article, the strings were installed twisted like that, but using odds-and-evens, with even panels on the way out and odd panels on the return wire. I have no idea if that would make a difference, but it makes the wiring more difficult to install. It means the wiring is all-links, and each link spans two panels - which doesn't suit the prefabricated link leads used by my installer (they do however have enough extra length to return to the rail).

With an HF sniffer loop on my spectrum analyser, I can see no signal on the roof, and none at 2m from the inverter. It only breaks the noise floor when I hang the loop directly against the inverter body - and that's likely to be near-field coupling from the magnetics.

If you want RF quiet, passive panels are the way to go. I'm lucky to have no partial shading to worry about, so micro-inverters or power optimisers would be no advantage in any case.

Clifford Heath.

Reply to
Clifford Heath

Ferrite beads can help quite a bit but will never eliminate RFI completely. I think you'll find the Enphase is the lesser of the available two evils.

Palomar Engineers: Solar System RFI Notice at the bottom of the page, they sell an RFI suppression kit for Enphase M190/M215/M240/M250/etc micro-inverters. There are other vendors selling these, or you can just buy the parts directly. They use type 75 ferrite, which is a good choice. The catch is that it only works from 200 KHz to 30 MHz. If the switching frequency is below 200 KHz, it will go through the filter like it wasn't even there. To cover 77 KHz and 1-30MHz, you might need two cores, from two different materials. Type 75 for the HF and type 76 for the LF. You'll need filters on both the input and output cables to/from the Enphase inverters.

At 77 KHz, I suspect you'll find far more sources of EMI/RFI than your proposed solar power system. Typically, you'll find plenty of devices with switching power supplies, most of which operate between 40KHz and

150KHz. On a spectrum analyzer, what you'll see is broadband hash as many power supply switchers include "spread spectrum" modulation that reduces the peak power, but smears the average power of a wide frequency range. Chances are good one of these will land in the vary narrow frequency spectrum used by digital clocks at 77 KHz. An easy experiment is to setup a 77KHz resonant loop antenna and plug it into an oscilloscope input. If you're lucky, you'll see DCF77 on the scope, usually between midnight and 6AM. Now, start turning off electronics and watch what happens to the noise on the scope. If it's like my house, it will slowly go down, but never quite disappear. I suspect you'll need to put as much effort into suppressing EMI/RFI from these sources, as you might need to put into suppressing EMI/RFI from your proposed Enphase micro-inverters.

LF (77 KHz) is not radiated by the electronic components. It is conducted by the power lines or magnetically induced by these same wires. The trick is to isolate the noise source (electronics) from the antennas (power wires). One way it to twist the wires together to reduce any differential mode radiation from the wires. Another way it raise the inductance, and therefore the impedance, of the power cables by installing a common mode choke. That's what the toroids in the Palomar kit and the QST article are doing.

Actually, I just realized that filtering 77 KHz with ferrite cores are not going to work for Enphase. Enphase micro-inverters use signaling on 144 kHz. Installing ferrite cores, that blocks communications between devices, is going to cause problems. That means you can probably block RF from 1-30 MHz, but not 77KHz unless you contrive a way to pass 144 KHz through (or rather around) the filters.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
 Click to see the full signature
Reply to
Jeff Liebermann

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.