Hybrid and Grid-Tie Solar Inverters From China: What Actually Matters

I have bought inverters for a few small projects over the years, and the lesson that keeps repeating is that the sticker on the box tells you almost nothing. What decides whether you are happy in year three is three unglamorous things: how the MPPT is set up, how the efficiency behaves across the real operating range, and how the thing handles the grid disappearing at 6pm when the load is heaviest. If you only really care about one number, you are probably buying the wrong inverter.

MPPT configuration is the first thing to ask about

A "hybrid" inverter with one MPPT input sounds fine until you realise you now have to make every string on the roof match. One input means one voltage window, and if part of your array is shaded in the afternoon, the whole array suffers. Two independent MPPTs let you run two arrays with different orientations, different shading profiles, or different panel counts without one dragging the other down. For a residential roof that gets any shade at all, two MPPTs is a real picture, not a luxury.

Ask the factory directly: how many MPPT inputs, what is the MPPT voltage range, and what is the max input current per MPPT. These three numbers, not the headline kW, tell you what you can actually wire. A common trap is an inverter rated for a big array but with an MPPT current limit that refuses to accept the panels the salesperson quoted you.

Ask for the efficiency curve, not the peak efficiency

Every brochure says "up to 98% efficiency." That is the peak, measured at a single sweet-spot load, and you will rarely sit there. What you want is the weighted or European efficiency, and ideally a curve showing behaviour at 5, 10, 20, 50 percent load. A cheap transformer-based inverter that hits 97% at half load but drops to 88% at 10% load will quietly cost you more energy than a proper transformerless unit over a year, especially on cloudy mornings and evenings.

Also ask about the self-consumption at night. A hybrid inverter that idles at 30W overnight runs the battery down for no reason. That is 0.7 kWh a day of nothing.

The grid-tie and battery transfer behaviour

When the grid drops, a hybrid inverter has to switch over fast enough that computers and routers do not reboot. Ask them for the transfer time in milliseconds. Good units are under 20ms; some cheap ones are 50ms or even 100ms, which is enough to trip a desktop PC. Ask whether the unit can output unbalanced loads across phases if it is a three-phase unit, and whether it supports generator input, because a lot of hybrid installs end up with a backup genset.

For the battery side, ask what chemistry and voltage range it accepts, and whether the BMS communication is actually open. Some manufacturers only talk to their own battery over a proprietary protocol. If you want to swap batteries in two years, that is a problem worth catching now.

Certifications and the paperwork trap

The inverter must carry the grid-code approvals for where it is installed, and these vary wildly: a unit that is fine in one country can be rejected outright in another. For Australia, look for AS/NZS 4777.2. For the US, UL 1741 and often the newer UL 1741 SB for grid support functions. Europe wants EN 50549 and the relevant country grid code. Ask the factory for the certificate and check the certificate number and the exact model listed on it. A generic "CE" sticker on an inverter is not a grid approval, and I have seen buyers get stuck at customs or with inspectors because the certificate did not name their exact model.

Ask who owns the certificate. If it is a trading company selling a factory's inverter, the certificate is in the factory's name, and you may not be able to use it for your own brand or warranty claims. This comes up more than people expect.

The things worth walking away from

If the supplier cannot tell you the MPPT count off the top of their head, that is not a good sign. If the efficiency answer is only the peak number, ask again for the curve. If they promise "all certifications available" without naming the standard, get the actual files. And if the price is dramatically below everyone else for the same claimed specs, you have almost certainly found a unit where one of those three things was quietly sacrificed. I would rather pay 15% more for an inverter with honest documentation than gamble on one that looks identical and costs less.

Finally, get a sample unit and actually run it under load with a meter for a few days before committing to volume. The datasheet is a marketing document. The meter does not lie.