18 month DIY project finally done: 39kW inverter power off grid system

[agtcovert]

I think in general NEC requires emt or mc for pv DC.

My understanding was schedule 80 PVC above ground and not inside the residence, EMT inside.

That's my current plan, I have some short runs in the ground where I'll put my wire in schedule 40 PVC, transition to schedule 80 above ground. I'll have an LB that's threaded to transition to EMT inside my house for the run to an inside disconnect, then to inverters.

At least that's the plan. If that's not right, then I guess I need to return some stuff. We get hot in the midwest, but I had planned to attach the above ground right below my siding.

 

[fmeili1]

I think in general NEC requires emt or mc for pv DC.

EMT or MC is only required for inside usage of PV DC - outside it's ok to go with PVC, regarding the requirements.

 

[timselectric]

PVC expands and contracts due to temperature changes.
Without expansion fittings and the ability to move freely, it will pull itself apart.
And if it gets hot enough, it will soften and discolor.
The roof is not a good place for it.
Sadly, I have had to learn a lot of things the hard way. Back before youtube and search engines.

 

[fmeili1]

PVC expands and contracts due to temperature changes.
Without expansion fittings and the ability to move freely, it will pull itself apart.
And if it gets hot enough, it will soften and discolor.
The roof is not a good place for it.

Yepp! Exactly like you've described - unfortunately.

Sadly, I have had to learn a lot of things the hard way. Back before youtube and search engines.

But as you see, even in times with search engines and youtube... it happens

It's not nice that manufactures will rate things as outdoor, direct sunlight approved with 25-40 years lasting time... and it will be rotted just after one hot summer here in Arizona... this is false labeling

 

[timselectric]

Yepp! Exactly like you've described - unfortunately.


But as you see, even in times with search engines and youtube... it happens

It's not nice that manufactures will rate things as outdoor, direct sunlight approved with 25-40 years lasting time... and it will be rotted just after one hot summer here in Arizona... this is false labeling

I have locations where I installed PVC outside in direct sunlight, 20 years ago. And it still looks like it did then. But I have also replaced many thousands of feet, that should have never been installed on a roof.
On a roof, is a horrible environment for any wiring. But some materials are better suited for it.

 

[jahyde]

In the dessert you would want EMT on the roof, and put insulating foam for hot water pipe over it so it doesn't de-rate as bad... plan on putting new foam every 5-6 years, or wrap the foam in something like eternabond tape for longer lifespan.

 

[fmeili1]

In the dessert you would want EMT on the roof, and put insulating foam for hot water pipe over it so it doesn't de-rate as bad... plan on putting new foam every 5-6 years, or wrap the foam in something like eternabond tape for longer lifespan.

That's a good idea with the insulating foam and the eternabond!

Without insulation, the metal conduit would be really hot and I fear about the insulation of the PV wire (even if it's made for this). I've found that metal exposed to direct sun could be up to 180-240°F (82-116°C) - here in the Mojave I think it will be trend to it's maximum.

 

[jahyde]

They also sell a fiberglass pipe insulation that will last you longer, but wind might be an issue, you might use the fiberglass, then wrap that with a roll of galvanized wire like for stucco lathe.

fiberglass pipe insulation - Google Search

You could buy the long pre-formed lengths, or the roll with aluminum layer might last longer - then wrap with wire to secure it from high wind.

 

[fmeili1]

I think, I need to install more PV modules and batteries since my wife finished the Christmas decoration



But in reality, it's not toooooooo bad for so many (LED) lights... just 1500W more consumption for about 6 hours daily - but anyway, in winter time, the 9kWh more energy consumption is a challenge

 

[AdamsSolarProject]

Congratulations for you fantastic solar system!

 

[fmeili1]

It's time to post some updates after the system has been in operation for 8.5 months now.

In the meantime I've finished a rebuild to change the system from using AC-in's for grid support to an "online double conversion" version, it's in "production" since January 12th now. Since this time, the house is always and exclusively driven by the off-grid inverters (no more bypass grid usage).
Many details about this rebuild are documented in the the DIY chargenectifier thread (following my posts starting from here).
In one of my posts I explained why I did this conversion.

The result looks like this (other pictures are in the DIY chargenectifier thread):


The main parameters are:

• 4x Huawei R4875G1 rectifiers with 75A max. charging current (4x 4kW)
• Usually they are running with 40A (~50% of max. power) to keep the fan noise low and this is enough grid support for 99% of my use cases
• placed a pair of these units on both of my battery racks
• CAN bus controlled via ESP32 controlled via ESPHome application and MQTT coupled to my smart home system (OpenHAB)

These units are extreme reliable and I like them a lot!

Some data about my solar harvest over the last 8.5 months:


Overall, my solar production declined to the lowest value in December (as expected). The forecast for January (only 6 days so far) looks like it will start to increase again in January. The grid usage value in the table for January are not longer accurate because Solar Assistant only shows grid usage via AC-in's of the AIOs which is not longer used since 01/12/2025. Instead I need to add the grid usage via the new R4875G1 rectifiers (via smart home integration).
The house load had a minimum in November because low usage of cooling and heating of the central heatpump.

The solar power over the last 90 days shows this minimum around Christmas time and started to increase at the beginning of the year.


My fan modding and the active cooling working flawless and the AIOs temperatures stay way below 50°C most of the time.


The 12 batteries are still very nice balanced between each other with low and high loads (charging and discharging), here is one example with low charging situation:


The max. cell deltas over all batteries look like this:


The average max. cell delta is between 20mV and 30mV for most of the time. A peak of about 85mV is a situation where the batteries reached 100% SOC while charging, which is ok for this (short) situation.

Here is a snapshot of the inverter view of SA for a typical situation with central heat pump running - I will say a typical load situation:



The system has no issues and the EG4-6500EX are running flawless. No dropouts no faults/errors and nearly no LED pulsing problems (sometimes a very little bit at extreme low load conditions, nearly not noticeable ).

I still think the EG4-6500EX (Axpert Max” MKS2-6500/8000) is better than its reputation. My feeling is that EG4 has sunset this inverter too early. One reason for this may be the unannounced hardware and firmware changes from Voltronic which caused a lot of problems and confusion with the EG4-6500EX about 2 years ago... who knows. For sure, the design of this Axpert Max MKS2 is really not the newest and today better options are available but it's a mature system design which usually just works after correct setup... just my 5 cents.