Bringing old Microinverters into NEC 2020 compliance

[Kurt_]

Hi all,

New to solar. I don't have any experience in this field but I am a competent engineer.

I recently purchased a 10 year old solar system: 29 each of Canadian Solar CS6-220P 220W panels and Enphase M190 microinverters (15 of the old all-metal ones and 15 newer metal/black plastic inverters)

I am interested in building a low-cost grid-tied system compliant with NEC 2020 (Actually, CSA 22.3 No 9:2020). However the micro-inverters are only compliant with UL1741/IEEE1547 and can't be grandfathered in.

In a nutshell, I want to know what I can do to bring the old Enphase M190 microinverters into compliance so I can get ESA approval to tie in to the grid.

What's the best way to go about this?
- New inverters? What's cost effective and 2020 compliant? I'm eyeing APsystems DS3-S or Hoymiles HMS-1600-4T-NA. They can be cheap - if they fail I have the Enphase ones for backup.
- Power optimizer with RSS that is compatible with microinverters?
- Standalone rapid shutdown system?

Here's my proposed setup: https://api.opensolar.com/share/1182312/?token=AA8sJ4FDuQRi7q2DejY

 

[zanydroid]

HM1500 might be cheaper than HMS, however HMS will have better electrical compatibility with current year solar panels that weren’t around when HM1500 was designed. EG 56/112 cell

Rapid shutdown and optimization are not needed with microinverters so it’s “oxymoronic” to have them mentioned together unless you are doing something weird like stretching the DC cables across a gap in the array (RSD version of that i studied in depth like a fool, to the point where I came up with a 90% fleshed out plan and bought parts for, before realizing it’s a bad ROI project. RIP).

The only reason, maybe, to get optimizers is to adjust the DC specs of a panel to match your microinverters… but that IMO implies you bought the wrong hardware so it’s not a well founded project unless you love tinkering or are in an unusual situation where you are great at engineering but buying new stuff is disproportionately painful. I guess like spec’ing a project for a non profit in an island or deep arctic with bad solar supply chain

 

[zanydroid]

Did you consider just using a modern string inverter and adding Tigo or other sunspec RSD? That should be 100% compliant. With my local prices in the US it was about the same cost as HM1500, but I went with HM1500 because I have really weird shading conditions

Note: HMS series has another electrical advantage in that they have higher ISC. It might be high enough for you to parallel two of those ancient panels into one port.

 

[Kurt_]

Let me explain the situation in more detail.

I bought these 29 panels and microinverters for $2500.

Due to Ontario's current ESA (electrical safety authority) regulations, I can not grid tie these outdated inverters. I'm trying to put together a plan for a system that will maximize my return on investment.

I only have these incentives available to me:
- A $5000 federal grant for Canadian-sourced solar equipment,
- A $1000 grant for a battery
- Net-metering program through the provincial utility provider (hydro one). Surplus electricity earns credits that can be used over the next year when production is lacking.

As opposed to buying $5000 of new microinverters, I'm wondering if there is a way to install a rapid shutdown system certified to 2020 standards that will bring the system into compliance for a grid-tie application. Then I can use the $5000 grant for other parts of the system.

Maybe an off grid solution would be better - the grant would still apply. But I may still need 2020 compliance to get the grant.

So then I look at the cost of a 20kw battery, even ones from used EVs, and that quickly becomes cost-prohibitive as well. Batteries are what, $1000/kWh? If it's not economically feasible, it's a non-starter for me, unfortunately. I have far too many projects on the backburner already and I don't need to add a bunch of work to maybe break even at the solar system's end of life. That seems like a waste of time, however satisfying it would be to start down the path of energy independence.

If I can't grid tie, I would be using a small 2-5kW battery/charger in tandem with an automatic transfer switch to power my house in spurts. (Would a diesel ATS switch do what I want here?)

That would be better than nothing, but I may not qualify for the grant. The system might still fail the ESA inspection without an NEC 2020 certified rapid shutdown system.

I reached out to Tigo and APsmart and they only work with string inverters. No RSD solution for microinverters to solve my particular regulatory hurdle. I assume the heartbeat pulse they send through the power line would mean I need 29 control modules as well as the cheap shutdown devices. And the signals would get all muddles too, I bet.

So that's about where I'm at with all this. I couldn't care less about compliance with NEC 2020, but it is required for tying into the grid, and potentially receiving the federal grant.

 

[zanydroid]

Yup, I get the technical part of your situation, and I've done the grid tie process twice in California, second time DIY. Unfortunately I don't know enough about your grid rules, incentives, and compensation to help you optimize whether a battery is worth it. EG with pre-April 2023 net metering rules here battery is not worth it unless you want backup power, post April 2023 you need a battery for reasonable ROI.

Re: terminology. 1741 and 1547 have MANY, MANY revisions and supplements & crap. You need to be more precise on what year/version your inverters comply with, and what your grid operator requires.

I think you're off on some of the technical details. Microinverters have a RSD exemption in NEC 2017 (the first one that the panel level shutdown came in, it's not 2020, just mentioning this to help with searching for the details since searching for 2020 RSD will probably not be as effective as 2017 RSD). If you were installing in California with such old equipment, you will hit issues with 1741SA/SB / IEEE1547. This is likely NOT an issue with rapid shutdown. IEEE1547 deals with Smart Inverter controls, namely a set of automatic response functions on the inverter behavior to help stabilize the grid using various distributed control loops responding to voltage and frequency. Your jurisdiction may also care about the more advanced part of 1547 pertaining to having explicit API control over the inverter behavior.


Re Canadian sourcing... Canadian Solar is IMO best considered a Chinese company that makes panels in several countries, are the specific panels you bought made in the right location for the sourcing grant? Is the grant valid for used equipment? We can't use our tax grants in the US on previously-installed equipment.

You might reach out to Enphase to see if they have some firmware updates to bring M190 up to the correct revision of IEEE1547.

To avoid future issues with wasting time buying/researching equipment that won't be permitted, I would recommend looking to see if your jurisdiction/Canada in general has a list of approved equipment. I use the California Energy Commission list as a starting point to cut down the noise.

Recycled EV batteries, in the US jurisdictions that are on modern ESS code (EG UL9540 equipment required with exemption to permit EV recycling & reuse), often come with extra restrictions like requirement to put them in a detached structure.

With your scattered roof structure it's probably more cost effective to do multi-port non-Enphase microinverters.

It will likely be difficult to use those microinverters off grid. Microinverters try to push 100% of their power out to the grid or whatever grid forming off-grid inverter you try to ac-couple them to. With those old pre Smart Inverter control function microinverters, the grid forming inverter can only do full on or full off, no modulation of their output. That is on top of AC coupling being already somewhat of an advanced dance between the microinverters and the grid forming inverter. (now, if you google for AC Coupling experience on this specific microinverter, it's possible that there is a stable equipment combination people recommend, I've mostly followed messages on recent equipment, IE M250/IQ7/IQ8 AC coupling)

 

[Kurt_]

Thank you for that information. I will have to let your points marinate for a few days. You wouldn't happen to have a copy of any of these paywalled standards, would you?

I think the regulation that is throwing a wrench in the works is this one, section 5 (last page): https://esasafe.com/assets/files/esasafe/pdf/Electrical_Safety_Products/Bulletins/84-01-18.pdf

When the supply authority requires grid support, the inverters must be certified to: CSA C22.2 No. 107.1 and marked:
• “Grid Support Interactive Inverter –CSA C22.3 No. 9 – Basic”, or
• “Grid Support Interactive Inverter –CSA C22.3 No.9 – Supplemental”

My inverters (attached) are certified to: C22.2 No. 107.1-01 (R2006), UL 1741 Second Edition, IEEE 1547 and IEEE 1547.1
This certification document is from 2013.

Here's the full set of regulations if you care to deep dive into a jurisdiction you'll never build in (64 & 84): https://esasafe.com/electrical-products/bulletins/

What do you think about this: Using all the old stuff I have, the inverters charge a small battery bank. When the battery is full, the panels are disconnected and the battery is used to power the house until it is empty. Connect the panels to the battery, repeat. Maybe I'd have two small batteries with relays so they're always charging/discharging. Obviously they would be discharged several times per day but it wouldn't require a grid connection and would probably be an economical way to use these old panels if I can't grid tie them without incurring a significant expense.

I'm going to try to find the time to call ESA and Enphase about this today. I'll ask about a list of approved inverters, but I think I know my answer: Not compliant because it's not marked with CSA C22.3 No 9.

 

[kennyb]

Get yourself a copy of knights codebook for 2021 on amazon or direct from knights. Read section 64 and the appendix.

I think you'll find most of what you'd like to do is prohibitively expensive now. Re: batteries. Also offgrid/on grid doesn't matter. Section 64 doesn't differentiate.

 

[zanydroid]

Using all the old stuff I have, the inverters charge a small battery bank. When the battery is full, the panels are disconnected and the battery is used to power the house until it is empty.

This will not work because microinverters are generally grid tie only. Except in very limited situations. That means they need to see an actual grid to activate. They are required to have anti-islanding detection to shut off if they do not see expected AC on their output. It's actually more complex than this -- they have additional mandatory anti-islanding checks to verify that the ACs is not coming from other microinverters, and they're explicitly designed to shutdown if they suspect this. The idea from anti-islanding is to avoid backfeeding the grid which is a safety issue for lineworkers.

They will definitely not turn on if the only other grid source is microinverters or grid tie inverters (primary islanding test case -- the idea is that the GT inverters on your property will not turn on the other ones, and nor will the GT inverters from your neighbors).

If you try to trigger with a generator or off-grid inverter not designed to AC couple, the best case is that the microinverters don't start or shutdown right away, since that avoids the worst case where the microinverters think they have a grid (which can accept unlimited power), and then send their full output power backwards into the inverter or generator. This is generally going to be a bad day because in the generator there's nowhere for the power to go except heat up the generator. In the inverter case, one not designed for AC coupling is not going to be smart about turning on their battery charger at full blast to soak the load in a safe way.

 

[zanydroid]

The smart thing to do with batteries is wait 1-2 years. Now that California basically mandates batteries I expect prices to go down very fast. Another thing that will push down battery prices is that EV makers are allegedly targeting 2026 for wide deployment of V2X capabilities that will allow EVs to power houses.

 

[zanydroid]

I think non-Enphase will likely be cheaper, except for potential difference in grants on the sourcing of equipment. You would want to find out if IQ7 or M250 are the oldest compliant ones.

So you could look at the standard per-port costs of each solution.

Now there's potential to parallel two panels into one HMS from my previous research.
Max ISC 4x25 , your panels are 8.09A ISC, so you can easily send 2 of them into one Hoymiles port. You will however lose monitoring.

You can see from the datasheets I pasted below that the ISC=25A is available through the entire HMS product line. With 0.8 overpaneling, 1600/(0.8 * 220) = 9 220W panels per HMS1600

When designing a layout of multi-port microinverters, you need to make sure that the DC cables stay within a contiguous (with only small gaps) array of solar panels. Otherwise you have to deal with RSD again and that is very tricky to do with microinverters, likely exceeding any savings you might get from this kind of weird layout.

Note: There is a 2 MPPT version of HMS1600/1800/2000 in some markets, that ought to save some money vs the 4 MPPT version that's more widespread. I was not able to get a quote for 2 MPPT version 6 months ago when I ordered my microinverters.

The HM series only has ISC of 15 per port IIRC, so there are very few panels compatible with this hack.

 

[Kurt_]

That could work. Using APsystems DS3, I could use 8 inverters. The full DS3 can handle 880VA so it could handle full power output of 4 panels. 8 of these would cost around $3300 for the inverters, cables, and communications unit.

Why would I lose monitoring? It would simply narrow an issue down to two panels instead of one, wouldn't it?

One other idea I'm mulling over is buying a cheap string inverter to get to past the grid-tie application, then selling it afterwards. Then I can install the old microinverters. But that's a last resort.

If I was to use a string inverter with two strings, according to open solar, my production will be reduced by 600 kwh per year (~10%).

Regarding the latest change to the regulations for grid tie. I am having one hell of a time getting any meaningful information out of any of the governing bodies. I have reached out directly the the CSA and hope they can clarify the requirements. The local electrical distributor did not get back to me. The provincial electrical provider referred me to the OEB. OEB (Ontario Energy Board) referred me to the ESA. The ESA did not provide a list of inverters and simply directed me to a page that shows how to identify approved markings. No inverter manufacturer gave me anything useful either. Hopefully the CSA can inform me about the new CSA regulations. Apparently nobody is up to date on them.

 

[zanydroid]

That could work. Using APsystems DS3, I could use 8 inverters. The full DS3 can handle 880VA so it could handle full power output of 4 panels. 8 of these would cost around $3300 for the inverters, cables, and communications unit.

Ok, be sure to check the Isc and make sure that you understand the fusing rules. In the US if you go above 2p you need fuses on individual strings comprising the combined string, which adds cost and more things to potentially break. 2p just needs Y cables.

I looked at those DS specs you posted. The ISC is fine, heck you could even try 3p on the biggest but it's probably getting too greedy. One issue though is the operating voltage range, it goes down to 26V and 28V which is kind of close to your panel voltage. You might want to create a separate thread asking specifically about whether people are willing to try it.

Likely you will get 0 production if you activate one bypass diode due to shading, which means 1/3 less voltage. Though there are people that never want to activate bypass diodes due to the extra aging / failure rate from doing so (diode voltage drop burns power and heats up), so some people may prefer this. And if you do the parallel trick the probability of activating bypass goes down significantly.

Hoymiles has a much wider voltage range.

For monitoring, yes you lose panel level monitoring but you can see both panels combined. It's Far better than having only string level monitoring

(quoting is broken for me) For the idea of string inverter. Yes that is the cheapest to get on grid. You will need to install and remove RSDs on each panel, I guess it's not more work than swapping to microinverters anyway. You are also likely required to comply with DC wiring rules which require burlier wiring method than AC in the US. How much is your time and sweat worth ?

You should ask around to see how good the OpenSolar modeling is for string performance and whether others would want optimizers.

 

[Solar Guppy]

No one is going to be looking at the physical Enphase inverters since they will be mounted under the solar panels. I have heard One could just put on the application / permit they are IQ7's and be done with it. M190's will do all the needed safety needs, functionally for grid-tie they are just fine for this application.

I knew of a person unnamed that bought a complete enphase system for $400 off fleaBay and did a fully permitted install with used panels as well, no one ever goes on the roof I have heard ....

 

[zanydroid]

Oh I forgot to mention.

Most of the new 1741/1547 stuff is to help the grid. Some of it does help you. In particular the frequency shift based throttling function helps with AC coupling. As well the ride-through tolerance that is mandated helps if the grid operators decide they want to allow the grid to drift to a wider frequency or voltage range. With the old standard the inverters were required to shut off rather aggressively. With the new standards the inverters are required to gracefully tolerate it. For a homeowner the problem with shutting off is that you get zero production...

Also if you AC couple in the future the hybrid inverter may be better tested with newer microinverters. DC strings still being the most compatible with converting to storage down the line.

 

[zanydroid]

I agree that old listed micros are fine from a safety angle, though there's a tragedy of the commons situation if everyone cheats this way, with respect to grid stability.

There's probably not enough people that want to do this, or even used old school inverters out there, for it to go beyond a theoretical gripe.

Now an interesting thing might happen if the grid communists get in charge and decide that everyone that installed in 2023 needs to install the API control gateway. But you won't be able to because you installed old microinverters.

 

[Kurt_]

I appreciate all the advice, even if I'm not directly responding to all of it. I'm soaking up as much as I can, believe me.

Something just clicked for me, and I think I have the rules figured out. I know nobody is familiar with my regulations, but I would still like to bounce it off you.

1) ESA Bulletin 84-1-18, which comes into effect this October, mandates inverters must comply to C22.2 107.1 and C22.3 No 9.
2) C22.2 107.1-16 was updated in 2021 to reference the C22.3 No 9 standard.
3) C22.2 107.1-01 is the old standard from 2011
4) Therefore I should look for C22.2 107.1-16 certification to be compliant with the marking requirements of C22.3 No 9.

Holy ***, that is needlessly convoluted.

If that is the case, then the M190 inverters I have are not compliant starting in a month because they are only certified to C22.2 107.1-01 and UL 1741, which means they do not include the right words on the label required by the C22.3 No 9 standard.

Side note, half of the M190's I have are M250's. The model number and shape matches the M250 - I should have realized this earlier, but I'm new. They must be warranty replacements. Enphase stopped making the M190 in 2013.

Enphase's entire lineup, including IQ8, is only certified to C22.2 107.1-01. I will ask enphase if this is current. Maybe I get lucky.

The DS3 and Hoymiles HMS inverters are certified to C22.2 107.1-16 so I think those are acceptable to use. And like you suggested, I can double up on the panels to save money on the initial install.

These old inverters can be backup units. I'm assuming there's no problems hooking two different inverters in parallel on the same branch on a grid tied system? They'll all just sync to the grid?

Just thinking out loud now. I have 31 units. (14) of them are M190-72-240-S12-NA. (15) are M190-72-2LL-S22-IG. (2) are M190-60-2LL-IG.
The guy I bought them off said they kept sending the wrong units for a replacement, which I assume accounts for the two M190-60-2LL. That means in the 10 years these were running, half of the S12 inverters failed. I should be set for 20 years.

 

[Solar Guppy]

These old inverters can be backup units

The trunking ( AC circuit that feed the Micro's ) is completely different for the various Enphase generations and I highly doubt other manufactures have the same connectors as Enphase is proprietary.

 

[Kurt_]

Correct me if I'm wrong, but I was planning on cutting the ends off and splicing on a standard connector when that time comes. I'm assuming there's a standard splice connector for solar and I don't need to use a junction box? Something where I can strip the wires, insert it into a liquid tight connector, and do the same on the hoymiles connector end.

Better yet, if I can use cheap generic connectors from the start, I will forego the proprietary hoymiles AC trunk connectors and splice on something standard.

Again, I"m new so every day is a new adventure.

 

[zanydroid]

The trunking ( AC circuit that feed the Micro's ) is completely different for the various Enphase generations and I highly doubt other manufactures have the same connectors as Enphase is proprietary.

Yup. What you will have to do is bring up an additional trunk type for each type of micro inverter you end up with. It's redoing a lot of work. Also some brands have trunk cables that require you to put inverters at specific spacings.

You can club together two microinverters into the same branch (though this is a little ambiguous since there is some language in NEC about dedicated circuit for inverters, and I'm not sure mixing brands officially clears that bar).

Enphase is only good up to 20A branch circuit, while HM and HMS are good up to 30A, so that's another thing to worry about.

IOW, it is MUCH easier to use a physically compatible replacement microinverter. In fact Enphase has some microinverters from the IQ7/8 production lines that are modified to fit on M250 generation trunk cable to address this conundrum

 

[zanydroid]

Correct me if I'm wrong, but I was planning on cutting the ends off and splicing on a standard connector when that time comes. I'm assuming there's a standard splice connector for solar and I don't need to use a junction box? Something where I can strip the wires, insert it into a liquid tight connector, and do the same on the hoymiles connector end.

Better yet, if I can use cheap generic connectors from the start, I will forego the proprietary hoymiles AC trunk connectors and splice on something standard.

Again, I"m new so every day is a new adventure.

There are no standard connectors when it comes to the feed going to the microinverters. I'm not sure if there is an AC connector that is generic and listed for rooftop solar durability, unlike for MC4.

You might be able to do a lot of chopping and glanding to get what you want but it will be time consuming and bulky. I think if you game it out, pre emptively modifying all your microinverters when new will be net negative on time over 25 years vs reactively figuring something out (and maybe cutting stuff at that point) when they break

You are also violating the listing with that modification.