SMA Sunny Island -- Didn't Triple-Check Torque

[Disturbedfuel15]

Well, now I know. I'll say that, it was a stupid mistake.... But I feel very lucky I caught it before big damage occurred.

My dual SI-6048s have been running my whole home great for a couple/few months now. The other day, I noticed lights flickering. Then, my wife was using the microwave when it restarted...

Went out and checked the inverters. I could hear 'crackling' inside the slave. Immediately shut everything off, running on grid power (passthrough is setup for maintenance). Removed the cover and oh sh*t...



Now I have to figure out if I have to replace:
1. Connector (best case)
2. Motherboard
3. Entire Sunny Island (worst case)

I found this thread https://diysolarforum.com/threads/schneider-wiring-failure.77976/#post-996967 which is nearly identical. I particularly wanted to highlight this person's helpful post:

I learned (from YT https://www.youtube.com/@ElectricProAcademy)

torque once
loosen
move conductor back and forth
torque again

do that 3 times

it allows strands to fill voids

and yes, ferrule on fine strands all the time

I have some research and work to do. I hope this helps others to remember to triple-check their torque values!

 

[Hedges]

On Sunny Island, the two neutral connections (AC1 and AC2) are simply in parallel. Same for ground.

If you're not using AC2, simply cut off damaged neutral wire end and move to AC2.
If you are using AC2, install a Polaris multi-tap (can clamp on middle of wire with a gap cut in the insulation) and have both AC1 and AC2 go to same screw terminal.

That way you won't have to repair anything.

These screw types aren't as convenient for torque wrench. Maybe adapted to screw bit. Use a socket extension; that helps minimize tilt of screw bit.

Stranded wire doesn't get crushed in screw terminal like it would with hydraulic crimper, and keeps settling.
Bend wire at an angle so you can wiggle it causing rotating motion in terminal. Wiggle, screw, wiggle, screw, repeat until screw doesn't turn more.
It may settle more by next day, so check and repeat.

 

[Skypower]

Bend wire at an angle so you can wiggle it causing rotating motion in terminal. Wiggle, screw, wiggle, screw, repeat until screw doesn't turn more.
It may settle more by next day, so check and repeat.

Very wise advice. An old electrician taught me this exact thing. And he was right.

 

[Disturbedfuel15]

On Sunny Island, the two neutral connections (AC1 and AC2) are simply in parallel. Same for ground.

If you're not using AC2, simply cut off damaged neutral wire end and move to AC2.
If you are using AC2, install a Polaris multi-tap (can clamp on middle of wire with a gap cut in the insulation) and have both AC1 and AC2 go to same screw terminal.

That way you won't have to repair anything.

These screw types aren't as convenient for torque wrench. Maybe adapted to screw bit. Use a socket extension; that helps minimize tilt of screw bit.

Stranded wire doesn't get crushed in screw terminal like it would with hydraulic crimper, and keeps settling.
Bend wire at an angle so you can wiggle it causing rotating motion in terminal. Wiggle, screw, wiggle, screw, repeat until screw doesn't turn more.
It may settle more by next day, so check and repeat.

Ohhhh you are so wise and this is why I took the chance of posting in the forum! I have many UL-listed polaris taps on hand. I completely forgot AC1 and AC2 (neutral) are parallel. Obviously--they eventually connect to the same bussbar. I am currently using it, but the polaris tap method is brilliant.

Thank you for saving me time and money.

 

[wpns]

I had a similar failure on an SMA grid-tie inverter, and left myself a note to tighten the connections yearly. And then I decommissioned tin favor of an EG4 ESS solution. Still need to check the torque yearly...

 

[dougbert]

wow, glad something I shared has helped out

this web site has so many others who share and we all learn along the way

 

[Disturbedfuel15]

Now that I am outside, ready to repair this, I realized that I forgot to ask something. Of course, I'll also google after this.

1. Should I replace the entire length of neutral? (I assume yes, it is compromised)

Google was quick. The answer is a clear, yes. Not sure about if it was 200ft, but that's not my scenario. It's heading about 3 feet down into the midnite solar e-panel.

 

[Hedges]

No, the wire wasn't carrying current above its ampacity. There was high contact resistance, so only the end got overheated. Over distance, as more energy was dissipated, temperature drops lower.

There could be concern about PCB overheating. Avoid high mechanical stress, see if screw backs out easily, maybe cut wire if not.
Then I would just smoke test it. Worst case would be damaged PC proceeding to short L-N. But I'll assume not.

 

[wpns]

Now that I am outside, ready to repair this, I realized that I forgot to ask something. Of course, I'll also google after this.

1. Should I replace the entire length of neutral? (I assume yes, it is compromised)

Google was quick. The answer is a clear, yes. Not sure about if it was 200ft, but that's not my scenario. It's heading about 3 feet down into the midnite solar e-panel.

Mine burned the length of the red wire to where it looked like another black wire into the conduit, so I swapped the whole thing as it was only about 6 feet long.

 

[Calvin98]

Wow. Thanks for sharing. I’m gonna put a yearly check on my list also. Do we know the official size of the screwdriver that fits into that Phillips type head? I ask because I’ve used the normal Phillips head in the past and when tightening down the head sometimes slips out of the screw head. So I’m assuming it’s not 100% correct fit for the screw and then of course, not sure of what the actual torque is. Plus that plastic case on a PCB board doesn’t really lend itself to super high torque values. But in the manual, being German made, they probably list a torque value. Need to check.

Over the years, I’ve noticed that regardless of how tight I make a screw contact like that on a breaker, if I come back a day or a hours later, I can always tighten it down a little bit more. I think Hedges idea of wiggling and tightening and wiggling and tightening is a good idea. I also like the idea of Ferrules. Those are fantastic. Especially on finer gauge wire. You can make a nice square and on the wire that fits into a connector like that. I noticed that the SMA DC disconnects do have ferrules already crimped on them and they fit perfectly inside the round holes on the inverter side.I’ve been using bare copper wire in the DC side, but I think I will be putting ferrules on them from now on. Even with the heavier gauge wire, I think you end up with more surface contact area than without a ferrule. A little bit more flat surface area for that screw to tighten down on.

 

[wpns]

Wow. Thanks for sharing. I’m gonna put a yearly check on my list also. Do we know the official size of the screwdriver that fits into that Phillips type head? I ask because I’ve used the normal Phillips head in the past and when tightening down the head sometimes slips out of the screw head. So I’m assuming it’s not 100% correct fit for the screw and then of course, not sure of what the actual torque is. Plus that plastic case on a PCB board doesn’t really lend itself to super high torque values. But in the manual, being German made, they probably list a torque value. Need to check.

Over the years, I’ve noticed that regardless of how tight I make a screw contact like that on a breaker, if I come back a day or a hours later, I can always tighten it down a little bit more. I think Hedges idea of wiggling and tightening and wiggling and tightening is a good idea. I also like the idea of Ferrules. Those are fantastic. Especially on finer gauge wire. You can make a nice square and on the wire that fits into a connector like that. I noticed that the SMA DC disconnects do have ferrules already crimped on them and they fit perfectly inside the round holes on the inverter side.I’ve been using bare copper wire in the DC side, but I think I will be putting ferrules on them from now on. Even with the heavier gauge wire, I think you end up with more surface contact area than without a ferrule. A little bit more flat surface area for that screw to tighten down on.

SMA might very well be using something Phillips-adjacent like PosiDrive.

 

[Calvin98]

Will need to check. I have a Phillips bit set that has several different types of bits with different shapes. PZ, PH, 1,2,3 etc. We’ll just need to make note of the correct type to use on the SI case.

Normally, I just use a normal Phillips bit screwdriver and don’t worry about it, but probably not a good idea.

 

[kommando]

Put your reading glasses on and shine a torch on the screw head.

The difference between Phillips and Pozi screw heads

The difference between Phillips and Pozi screw heads. We talk all about this subject. Blog post by BS Fixings.

bsfixings.uk

The SI5048 for Europe uses a Wago 745 connector or similar so no screws to tighten.

 

[robbob2112]

Funny you should mention this - I was out installing a 240 breaker for the new furnace humidifier and while I was there I checked the torque on all the breakers in that sub panel - all were loose. Now I need to go through the whole main panel and check them.

The tip on the screwdriver for all of my breakers is SQ02 - it is square - I have the insulated set from Milwaukee for this work. And SQ02 will do so long as the tip doesn't come out and then you have to get it loose... Looks like I could use a flat blade if it was exactly the right width but the square drive fit perfect and lets me torque it down.

Anyways, my panel and sub panel are both Siemens. (full and 1/2 width breakers)
Square D appears to be a flat head
Midnite solar appears to be the SQ01 or flathead (1/2 width)

And, when I tried to look it up I was lead astray - the square tip is also called a Robertson tip depending on the MFG

 

[Calvin98]

Right. Robertson square drive tip. After many years of just using a flat blade screwdriver I put a Robertson screw drive in my kit just for those breakers that have the square drive. Makes it a lot easier to tighten well. If your flat blade screwdriver is too small it won’t reach both ends. The square drive is a much better option. But now I have to check which tip to use for the Sunny Island’s and put that in the Kit too.

 

[Calvin98]

The SI5048 for Europe uses a Wago 745 connector or similar so no screws to tighten.

Wago 745. Right. That’s the connector I saw on that SMA dc disconnect that I have. It has around sleeve so a round ferrule will fit in perfectly. That’s a really nice connector. Tight fit and easy to assemble and disassemble. No torque specs needed.

 

[robbob2112]

How does the Wago hold the wire in? - looking at the designs I can't quite see the mechanism. I assume they must be good or they wouldn't be used but not sure how I feel about spring loaded.

 

[kommando]

Better to look on YouTube to see how it operates but by inserting a screwdriver in the top slot and levering down the lower spring steel opens to accept the bare ended cable, once inserted you remove the screwdriver and the lower spring clamps on the cable.

 

[Calvin98]

Yes, you have to insert a small screwdriver to loosen the spring. On one of my inverters on the DC side there’s a built-in lever that you can use to take the tension off the spring. I guess for the AC side you need to use the screwdriver. But I seem to remember that on mine there’s also a lever that you can use.

 

[jonathanpatr]

Well, now I know. I'll say that, it was a stupid mistake.... But I feel very lucky I caught it before big damage occurred.

My dual SI-6048s have been running my whole home great for a couple/few months now. The other day, I noticed lights flickering. Then, my wife was using the microwave when it restarted...

Went out and checked the inverters. I could hear 'crackling' inside the slave. Immediately shut everything off, running on grid power (passthrough is setup for maintenance). Removed the cover and oh sh*t...

View attachment 258078

Now I have to figure out if I have to replace:
1. Connector (best case)
2. Motherboard
3. Entire Sunny Island (worst case)

I found this thread https://diysolarforum.com/threads/schneider-wiring-failure.77976/#post-996967 which is nearly identical. I particularly wanted to highlight this person's helpful post:



I have some research and work to do. I hope this helps others to remember to triple-check their torque values!

This is key to torqueing. My electrician learned this the hard way on my 240 plug. He torqued once...evse didn't work properly so I called him and he did the 3 time torqueing described here. Perfect operation ever since.