EEVblog 1518 - Hoymiles Solar Microinverter TEARDOWN

[EEVblog]

Depotting, teardown, and some reverse engineering analysis of the 400W Hoymiles solar microinverter.

[Ranayna]

The video is still unlisted. Is that intentional?

[uer166]

Schematic doesn't seem to be 100% correct. I think on output side you have a H-bridge "unfolder" running at 60/50Hz, and the flybacks are generating the half-wave rectified mains waveform on output side, which is why the output caps are so small, they sit across mains output and aren't DC.

 Edit: something like attached. A very cool topology, although would probably struggle to maintain sine output at no/low load if it was islanded. Good thing it's grid tied!

[Red Squirrel]

That was a cool tear down. A video on how these work would actually be really cool. Ex: how do they sync with the grid, and how do they detect if it goes down. Especially when there are many of these on the grid, how does one inverter not think another inverter is "the grid" and keep going, for example. I presume it goes by current draw, if grid goes down it will try to pull massive current?

[EEVblog]

FYI

[maelli]

Took an involar 250W microinverter apart, since it failed.
The gunk on that one was much harder to remove.
Topology is fairly similar.
IRFS4321 150V 85A mosfets
The 2-pin To220 are 1200V Sic Schottky diodes, CREE C2D05120. Interesting
Two MC56F8xxx, one on the solar side, one on the line side.

[maelli]

Why the large electrolytic capacitor bank 4 x 2.7mF 63V?
Since power delivery must be at power factor 1, there will be a 2 x 50Hz ripple in the power flow (800W line voltage maximum, 0W during zero crossing).
Since there is not other large storage of energy in this microinverter, the low voltage side caps have to handle this energy storage.
It should be possible to measure some 100Hz ripple in the PV module voltage, mostly at full sun when current is high, voltage low.
Cannot measure this on the involar unit I took apart, since it is Kaputt.

I personally do not like to have that many electrolytics on my roof, time will tell how long this lasts. At least they use 105°C types.
My string inverter is in the basement where it is cold. And there are very few capacitors in my inverter, since it is 3-phase :-)

[BrianHG]

That was a cool tear down. A video on how these work would actually be really cool. Ex: how do they sync with the grid, and how do they detect if it goes down. Especially when there are many of these on the grid, how does one inverter not think another inverter is "the grid" and keep going, for example. I presume it goes by current draw, if grid goes down it will try to pull massive current?

I know a lot say PLL style.  But, since I'm sure they can see the amount of current between their inverter driver and also monitor the voltage of the grid to determine the wattage sent out, it is probably as simple as driving a positive amount of power out withing the current limit of the inverter.

And since power = V*I, when seeing a negative voltage, they send a greater negative current, when they see a positive voltage, they send a greater positive current.  When they see no voltage, they don't feed anything as this may energize the grid during a power failure.

There are also probably limits and a degree of timing in place to keep a clean sine wave recognizing 50 or 60 hz system and when to swing the current drive in reverse during the second half of each waveform.

[thm_w]

The simplest passive anti-islanding method is just measuring the voltage and frequency of the mains and shutting down if its out of spec (1.1% for voltage, ~2% or less for  frequency).
The chance of the grid shutting down and a few microinverters being able to backfeed power within that range is low.

There still are other more complex active methods though.
https://ieeexplore.ieee.org/abstract/document/4778680/
https://www.sciencedirect.com/science/article/abs/pii/S0306261910001820
https://forum.solar-electric.com/discussion/19429/microinverters-and-anti-islanding-how-does-it-work-technically

[RileyKA]

Dave, I love the solar content. I find the micro inverters very interesting. I was wondering if you think you would ever be able to get your hands on a US version of one? I'm curious how they handle creating and managing the split phase power we use in the US. I see that enphase does make a US split phase inverter

https://enphase.com/store/microinverters/iq8-series/iq8plus-microinverter

[f4eru]

The simplest passive anti-islanding method is just measuring the voltage and frequency of the mains and shutting down if its out of spec (1.1% for voltage, ~2% or less for  frequency).

1.1% on voltage ?
Nope.
Voltage actually varies quite a lot because of the impedance/resistance of the network and cables.

[thm_w]

The simplest passive anti-islanding method is just measuring the voltage and frequency of the mains and shutting down if its out of spec (1.1% for voltage, ~2% or less for  frequency).

1.1% on voltage ?
Nope.
Voltage actually varies quite a lot because of the impedance/resistance of the network and cables.

Yeah looks like the actual spec is +10/-6%, and preferred +6/-2%. Not sure where I got 1.1.

[EEVblog]

Dave, I love the solar content. I find the micro inverters very interesting. I was wondering if you think you would ever be able to get your hands on a US version of one? I'm curious how they handle creating and managing the split phase power we use in the US. I see that enphase does make a US split phase inverter
https://enphase.com/store/microinverters/iq8-series/iq8plus-microinverter

Didn't know a 120V split phase inverter was a thing, thought they were all 240V?

[RileyKA]

Dave, I love the solar content. I find the micro inverters very interesting. I was wondering if you think you would ever be able to get your hands on a US version of one? I'm curious how they handle creating and managing the split phase power we use in the US. I see that enphase does make a US split phase inverter
https://enphase.com/store/microinverters/iq8-series/iq8plus-microinverter

Didn't know a 120V split phase inverter was a thing, thought they were all 240V?

I didn't either until I looked for it. Most systems I've for the US market seem to be string inverter systems, hadn't seen micro inverters until you featured them in your install video of the enphase system.
After seeing this teardown video, and the talk of the grid disconnect feature, i got to wondering how all that works here. For single-phase 240v countries it seems like it would be easier to manage creating, monitoring and synchronizing with the grid than it would be here.
I guess in reality it would be the same process only doubled and at the lower 120v and 180 degrees out of phase.
Or maybe I'm just over thinking this and its just a 240v inverter with a center tap off the transformer.

[thm_w]

The Hoymiles MI-1200 has a 120/240V split phase diagram in the manual.
It looks like the same circuitry multiplied by 3 inside.



https://fccid.io/2ARNB-MI1200/Internal-Photos/Internal-photos-4133195
https://fccid.io/2ARNB-MI1200/User-Manual/User-Manual-4133200.pdf
https://fccid.io/2ARNB

[Phoenix]

Yeah looks like the actual spec is +10/-6%, and preferred +6/-2%. Not sure where I got 1.1.

The domestic voltage spec is not the passive anti-islanding limit. I will discuss (to the best of my memory) Australia.
In terms of over voltage disconnection there are 2 fast limits and one 10 minute average limit.
>275V instantaneous - 200ms
>265V for 1s - 1s
>258V (adjustable) for 10 minutes - 3s
In terms of under voltage two limits plus specific ride through requirements
<180V for 10s - 1s
<70V for 1s - 1s
I don't remember ride through it's complicated.

In terms of frequency the disconnect range is 47hz and 52hz, with similar delays. Oh and 4hz/s rate of change of frequency.

(note that reconnection limits are different, and based on the grid spec).

The regions I've studied (Aus, Europe, North America) all have similar limit set ups, of course with different values.

Most regions also require active anti islanding (except UK). This requires the inverter to inject a disturbance signal into the grid and monitor the effects. This allows it to detect scenarios where, when an accidental islanded load matches the inverter generation, the voltage and frequency limits are still in range.

[el737]

Hi,
I hope it is ok to ask here - can anyone please tell me which wire diameter would be safe to use from the Hoymiles 800 to the outlet? I have read about it, but it would be nice to hear from someone who is using this or similar inverter for a while.
Thanks

[thm_w]

Hi,
I hope it is ok to ask here - can anyone please tell me which wire diameter would be safe to use from the Hoymiles 800 to the outlet? I have read about it, but it would be nice to hear from someone who is using this or similar inverter for a while.
Thanks

It says 12AWG in the manual, no?
Anything else will depend on your local rules and regulations.

[el737]

Didn't get the manual. But the confusion always comes from the fact that in the original package the manufacturer included no-name no-label stranded cable that seems even thinner than 14AWG 
I thought of using 12, just wanted to confirm since that is one of the rare electric things I've never worked with (solar systems).
Thanks

[thm_w]

Didn't get the manual. But the confusion always comes from the fact that in the original package the manufacturer included no-name no-label stranded cable that seems even thinner than 14AWG 
I thought of using 12, just wanted to confirm since that is one of the rare electric things I've never worked with (solar systems).
Thanks

https://www.hoymiles.com/wp-content/uploads/downloadupload/User%20Manual_HM-600-700-800_Global_EN_V202206.pdf