What I'm learning about DIY Enphase and Hoymiles Microinverters

[ncsolarelectric]

Update, my Enphase system has been running flawlessly since turning it on. No complaints. Other than the weather because it's been overcast or thunderstorms nearly every day. However, today my system hit a new production record, 25.19 kWh, or 2.52 kWh per solar panel, per day. It is outperforming. Hopefully, the weather will cooperate for a little while so I can see what it can really do.

 

[svetz]

Hopefully, the weather will cooperate for a little while so I can see what it can really do.

Have you run SAM to get the expected max for your setup for various times of the year?

..., today my system hit a new production record, 25.19 kWh, or 2.52 kWh per solar panel, per day. It is outperforming.

The 1000 W/m² at sea level depends on "average" air mass, but it's not always constant and changes seasonally (SAM has a graph for your area). This is the output from my solar sensor the other day, note the max was 1187.5.
And yes, my inverters clipped out the solar bonus dag nab it! ; -)

.......

WOOT! ~41% to that first megawatt hour! Planning a party? Or are all your friends already sick of hearing how great solar power is? ; -)

 

[ncsolarelectric]

Have you run SAM to get the expected max for your setup for various times of the year?


The 1000 W/m² at sea level depends on "average" air mass, but it's not always constant and changes seasonally (SAM has a graph for your area). This is the output from my solar sensor the other day, note the max was 1187.5.
And yes, my inverters clipped out the solar bonus dag nab it! ; -)
View attachment 102726

WOOT! ~41% to that first megawatt hour! Planning a party? Or are all your friends already sick of hearing how great solar power is? ; -)

According to my simulation, July should average 18 kWh per day. So far, July is averaging 19.2 kWh. Outperforming the estimate.

 

[svetz]

According to my simulation, July should average 18 kWh per day. So far, July is averaging 19.2 kWh. Outperforming the estimate.

My setup consistently "overperforms" too. I've also noticed that while none of my microinverters run under 285W (their rated continuous max) it's not uncommon to see some around 305W for extended periods. Enphase has a top-of-the-line pricetag, but they don't inflate their specs.

Have you decided the layout for Hoymiles yet? Is it a testbed, going on a different house, or will it be on the same microgrid as Enphase?

 

[ncsolarelectric]

My setup consistently "overperforms" too. I've also noticed that while none of my microinverters run under 285W (their rated continuous max) it's not uncommon to see some around 305W for extended periods. Enphase has a top-of-the-line pricetag, but they don't inflate their specs.

Have you decided the layout for Hoymiles yet? Is it a testbed, going on a different house, or will it be on the same microgrid as Enphase?

My Hoymiles inverters and cables should arrive on Thursday. I can't afford to buy any solar panels for myself right now, so I'm thinking of replacing one of the IQ8+ with an HM-350NT, on a separate branch and see how it performs. It has a 319W continuous output with a 350W peak, compared to the 290W / 300W peak of the IQ8+. Then I can set up the DTU and see how well that works as well.
I need to make some videos!

 

[ncsolarelectric]

I ordered the Chint Din-rail meter used with Hoymiles DTU-Pro to control the Export current. The inverters can be configured for zero-export. My first experiment is to connect the meter to the output of an off-grid inverter and then control the microinverters so they do not backfeed into the off-grid inverter. Then they should run in parallel with the microgrid and be stable, without shutting down randomly.

My second experiment is to connect the microinverters to a 48Vdc battery instead of solar panels and see how well they perform electrically and thermally.

If both experiments are successful, we will be able to create our own AC-coupled batteries with a backup microgrid. Who wants to donate an inverter and battery for this test? Or just buy a pallet of solar panels, that would help me out immensely!

 

[BiduleOhm]

You (all people in this thread) may find that video interesting:

 

[Cubedweller]

You (all people in this thread) may find that video interesting:

Thanks. This really has me thinking to forego the Enphase batteries and just using something else for the remaining 14 panels. SolarEdge or just plain string stuff with EG4 or SOK batteries.

 

[ncsolarelectric]

Thanks. This really has me thinking to forego the Enphase batteries and just using something else for the remaining 14 panels. SolarEdge or just plain string stuff with EG4 or SOK batteries.

I'm not sure I understand your reasoning based on what was in the video. He clearly showed that for 14 microinverters, the nighttime power consumption was only 8 watts, or less than 0.6 watts per microinverter, or less than 0.2%.

 

[Cubedweller]

I'm not sure I understand your reasoning based on what was in the video. He clearly showed that for 14 microinverters, the nighttime power consumption was only 8 watts, or less than 0.6 watts per microinverter, or less than 0.2%.

He said 18 watts apparent power, each, or 280VAr for 14, which is no problem with the grid, but using batteries at night instead, could be an issue. This explains some Enphase customers with batteries that have been discharging at night with no loads on them. Especially in winter this could be an issue. Those batteries also have IQ8s strapped on top of them. Dave doesn't have batteries yet, so that should be interesting if he does.

 

[ncsolarelectric]

He said 18 watts apparent power, each, or 280VAr for 14, which is no problem with the grid, but using batteries at night instead, could be an issue. This explains some Enphase customers with batteries that have been discharging at night with no loads on them. Especially in winter this could be an issue. Those batteries also have IQ8s strapped on top of them. Dave doesn't have batteries yet, so that should be interesting if he does.

The batteries do not deliver VArs, they are DC and can only deliver Watts. So unless Enphase has a huge AC ripple voltage riding on top of the DC voltage, which I doubt is the case. The Vars are circulating in the inductance and capacitance of the inverter, not the battery. That power comes from the utility grid when on-grid, but when running on battery, the inverter inductors and capacitors have losses which add up to the total efficiency of the inverter while running on battery. This is true of any inverter running on battery and most are much less efficient. Also, yes, the Enphase IQ8BAT inverters operating on battery at night will have losses even if there is no load attached to it, simply because inverters are not 100% efficient. This is a non-issue, IMO. Energy is conserved, it's Newton's law.

 

[svetz]

You (all people in this thread) may find that video interesting...

The PV microinverters go offline when there isn't enough sun, AFAIK they're not in "standby", probably why Enphase doesn't report their standby power consumption. Every night my Envoy reports communication loss to the microinverters, and every morning the event clears. They are powered by DC from solar.

So his setup and what he's actually measuring seems questionable. It could be tons of things.

The IQ8s in an IQ Battery are in standby and they do consumer power (but he doesn't have a battery). A quick look at my system and it seems to be ~3.3 watts, probably on par with other systems.

Shouldn't be hard to verify his results, just need someone with a battery to run an experiment at night: disconnect the house, disconnect the grid and then see if there's a >200 watt drain on the battery. Since the house is disconnected, the only thing the batteries should be powering is the envoy and wireless dongles.

 

[ncsolarelectric]

That test is easier said than done. The Envoy runs from the backup sub-panel, so it is powered by the battery when on battery. So if I turn off "all" the load, I don't have an Envoy to gather data.

 

[BiduleOhm]

The PV microinverters go offline when there isn't enough sun, AFAIK they're not in "standby", probably why Enphase doesn't report their standby power consumption. Every night my Envoy reports communication loss to the microinverters, and every morning the event clears. They are powered by DC from solar.

So his setup and what he's actually measuring seems questionable. It could be tons of things.

The IQ8s in an IQ Battery are in standby and they do consumer power (but he doesn't have a battery). A quick look at my system and it seems to be ~3.3 watts, probably on par with other systems.

Shouldn't be hard to verify his results, just need someone with a battery to run an experiment at night: disconnect the house, disconnect the grid and then see if there's a >200 watt drain on the battery. Since the house is disconnected, the only thing the batteries should be powering is the envoy and wireless dongles.

I don't know if he's right here even if he's a very knowledgeable and respected EE, I just stumbled onto this video and thought it could be useful for some people here.

Also, I'm pretty sure he's right about the 19 VA but I'm pretty sure VAR aren't pulling current from the battery, only W. So even if the reactive power is super high it probably doesn't matter at the battery level, need confirmation with real world figures tho.

 

[svetz]

That test is easier said than done. The Envoy runs from the backup sub-panel, so it is powered by the battery when on battery. So if I turn off "all" the load, I don't have an Envoy to gather data.

Agreed about the difficulty, if I turn off the load the wife will be annoyed. ; -)

I don't know if he's right here even if he's a very knowledgeable and respected EE...

Just because the microinverters aren't in standby as he assumed doesn't mean they aren't somehow interacting with the grid. But, it could just as easily be something to do with the SMA he had in the circuit. I definitely want to try it to see what sort of results I get.

 

[svetz]

From: https://www.mcelectrical.com.au/hoymiles-microinverter-review/ Thanks for posting the link, ignorant as I am about Hoymiles it really helped to understand the excitement about it. I did spot some things that didn't ring true about Enphase... ...with Enphase, you can’t look down on your app and see how much your entire system is producing right now... But, you can... in fact Enphase's animated depiction is very elegant in comparison. Enphase has cellphone apps, 3 different web views of the data, an official remote API, and an unsupported local API... tough to beat them in this category.

Hoymiles

Enphase

There was also the MPPT on shade, they stated that Enphase was significantly better.

Enphase just has a superior maximum power point Algorithm

That makes no sense to me, there's nothing arcane or mysterious about the MPPT algorithm. I suspect if anything they just didn't wait long enough for the Hoymiles to zero in on it (or perhaps Enphase's speed is the advantage?).

Hoymiles has a 12-year warranty on the micros and Enphase has a 10-year warranty.

Possibly an Australian thing, the U.S. warranty lists it as 25 years.

Probably some of it is based on their bias, it's easy to not look for things when you dislike something:

... let me be clear from the outset. I’m not the biggest fan of microinverters.

Still, an interesting read to be sure.

 

[svetz]

...Shouldn't be hard to verify his results, just need someone with a battery to run an experiment at night: disconnect the house, disconnect the grid and then see if there's a >200 watt drain on the battery. Since the house is disconnected, the only thing the batteries should be powering is the envoy and wireless dongles. Looks like I just need to flip two breakers in the Enpower to do the test, the mains and the loads. The battery and combiner box (panels and Envoy) would be still be live. Probably need to leave if off for 30 min or so. As my wife won't buy the "in the name of science" I'll have to sneak it when she's sleeping.

If you have any ideas about how to make the test better let me know. I know we're primarily after the battery watts in the dark with 0 load, but it would be fun to measure the reactive power too. Normally I can get the reactive power from https://envoy.local/ivp/meters/readings which gives:

[
    {
        "timestamp": 1658242342,
        "apparentEnergy": 30645196.700,
        "reactEnergyLagg": 8115296.496,
        "reactEnergyLead": 12.106,
        "instantaneousDemand": 5364.222,
        "activePower": 5364.222,
        "apparentPower": 5398.430,
        "reactivePower": 490.053,
        "pwrFactor": 0.993,
        "voltage": 244.130,
        "current": 44.228,
        "freq": 60.000,

But, with the load off the PC won't run. Might be able to do something with the cell phone, no way I can type in a 128 byte 6 month token.

 

[BiduleOhm]

So he made a new (and far clearer) video on the subject:

(the link brings you directly to the interesting part but you can watch the whole video if you want).

So, the TLDR is: reactive power isn't drawn from the battery as it goes back and forth but the current has to be generated and there's losses (mainly in the inverter) and that's what will eat some power from the battery. So yea, not eating the full reactive power figure from the battery but still eating some part of that.

 

[svetz]

So he made a new (and far clearer) video...

Thanks for posting, it was interesting.

All grid-tied inverters are going to have capacitors on the AC side, so every inverter (not just microinverters) should have the same issues. Really nice that the math added up with the three 330 nF capacitors. Wonder how it compares with a comparable 8 kW string inverter? Has anyone tested this?

It strikes me that much of the conversation is about IQ8s rather than the previous generations of microinverter. For example, I've never experienced the bit at 25 minutes with my IQ7s where he says they can stay on at night and do updates or provide power factoring for the microgrid. My updates always come during daylight hours and they always lose communication at sundown. AFAIK only the IQ8s are bidirectional. If anyone has experienced IQ7s being bidirectional or getting updates at night please post about it.

The 15 µA consumption of standby current he got from Enphase is interesting. It seems ridiculously low when compared to other inverters. It's too low from my measurements yesterday (although, those measurements do include powering the other devices in the IQ battery (e.g., BMU, control board, communications, LEDs). Guess it's possible that's just the IQ8 portion.

...So yea, not eating the full reactive power figure from the battery but still eating some part of that.

From his video he thinks the battery losses are copper losses (which since it's less than 1/2 watt I'll just ignore) and the inverter efficiency which seems to be the big hitter. For his system it's 1.1 amp at 245V = 270W x 0.04 ~= 10W. For a 16 hour night, that's 160 Wh on his system.

Unfortunately, that's too low to measure accurately with the test in #67 (earlier it was reported as 270W which would easily be measurable). Still curious as to what it really is. I do have a CT on the PV inputs, but it shows 0 at night time via the Enphase tools. Of course, I also have IQ7s so it could be different.

I doubt it would actually be as high as 160 Wh as he stated because there are capacitors in the battery's inverter too... so isn't most of the energy just reflected around? That is the actual battery power is only supplying the copper loss plus capacitor losses with the inverter inefficiency penalty?

Ideally, the way to measure it would be to get current reading from the DC side. I won't get around to that until the next decade when the warranty is expired. ; -) Might be raw data available if I ever get around to cracking the zigbee API.

If anyone has thoughts on how to measure it I'd love to hear them.


Q-Relays
That bit at the end was interesting. In the US the microinverters are UL 1741 compatible, so no additional devices needed. But, in Australia to accomplish what sounds like the same thing, rather than reprogram the voltages/frequencies/times Enphase already has for UL1741 into a special AU profile; they make their customers buy a Q Relay.

I suspect the decision was made that way as they found a device they could rebrand and it was cheaper and faster than software development and AU lab testing. But it's sad that it's needed as it seems like another point of failure.

 

[BiduleOhm]

I doubt it would actually be as high as 160 Wh as he stated as there are capacitors in the battery's inverter too... so isn't most of the energy just reflected around? That is the actual battery power is only supplying the copper loss plus capacitor losses with the inverter inefficiency penalty?

Most of the losses of an inverter with a light load is switching losses so I bet that would be the main concern here.

 

[ncsolarelectric]

My new microinverters just arrived. Video (No ads)

 

[Herb]

When I learned about the IQ8 Microinverters and their ability to run off-grid with "Sunlight Backup", I had to try it. I was going to put in a SolarEdge system in 2020, but due to the plague, it never happened. I installed a backyard ballasted ground mount, using Unirac RM10 racks and CMU blocks. Jinko 400W PERK panels, and Enphase IQ8Plus-72-2-US inverters. Sunlight Backup and Battery Backup with Enphse both require their System Controller 2, which has the split-phase "neutral forming" transformer and the automatic transfer switch. A pricey piece of equipment!
So on Thursday 6/9/22, my system passed the final inspection, and the utility granted PTO the same day. The next morning when I tried to start my system, I couldn't because of these factors that I missed, when I went through the Enphase training. It's in there, but it's confusing. Also a word of warning to DIYers. Enphase's 25 Year Warranty requires that the system be installed by a Certified Enphase Installer. If you don't go through their training, you'll void your warranty.

1. A grid-tied solar system only requires Wifi, but when using the System Controller 2 for either Sunlight or Battery Backup. A cellular modem is required to commission the system. The App will not let me provision the system without a Cell modem. I didn't buy one until yesterday.
2. A Sunlight backup system is required to have 2 Load Controllers. I only have two 20A circuits, so I only bought 1 LC. The App will not let me provision the system without 2 LCs. So I would need to buy something I don't need, to make the system operate. (OUCH!)
3. A Battery backup system does not require the 2 LCs. So rather than waste money on a second LC I don't need, I invested in the IQ Battery 3T which I had intended to do eventually from the beginning. I just didn't expect it to be "required" to commission the system.

The last thing I learned is, that there are new alternative microinverters that are UL Listed, CA Rule 21 approved, and are less than half the price of Enphase. They are very user-friendly for DIYers and practically plug-n-play.

All true, I've had an Enphase PV system for around 5 years and they are great, so far no issues, but it does nothing when the power goes out! that is a total shame in my book. Anyways I ended up making my own 15kw battery backup system, is not huge but it does more than the Enphase system when the lights go out!

 

[fromport]

All true, I've had an Enphase PV system for around 5 years and they are great, so far no issues, but it does nothing when the power goes out! that is a total shame in my book. Anyways I ended up making my own 15kw battery backup system, is not huge but it does more than the Enphase system when the lights go out!

IQ6/7/8 all do the frequency shifting power modulating. Can't you hook them up to your battery backup system than can use that to help power the load & charge the batteries in case of no grid?

 

[Ampster]

All true, I've had an Enphase PV system for around 5 years and they are great, so far no issues, but it does nothing when the power goes out! that is a total shame in my book. Anyways I ended up making my own 15kw battery backup system, is not huge but it does more than the Enphase system when the lights go out!

It looks like you missed an opportunity to AC couple your Enphase to your hybrid inverter. The Enphase was never intended to be just a GT system which by its nature will not produce power unless it is AC coupled to an hybrid inverter.

 

[svetz]

...Enphase ... by its nature will not produce power unless it is AC coupled to an hybrid inverter.

Enphase's IQ8 microinverters are hybrid inverters (e.g., work while the grid is up or down even without battery) and older microinverters can now work offline too without a separate hybrid inverter via their Ensemble solution.