safety when connecting panels into strings

[RShackleford]

When panels are connected into strings (by series connection), very high voltages can accrue. When connecting the panels to each other and to the inverter (and possibly junction boxes or combiner boxes along the way), what is the best way to insure safety ?

In my specific case: a ground-mount system. The ends of each of my two strings go into a junction box. There they are connected by a terminal block to THWN wire which runs underground in conduit to the SMA inverter.

I can think of a couple of good ideas:

1. Wire the strings up when the sun's not shining. However, when wiring the strings, the voltages you might come into contact with are probably the sums of the open-circuit voltages of the panels which are connected at any given time. If I'm interpreting these I-V curves (below) for my panel correctly, it looks like the open-circuit voltage for the panel doesn't drop much with lower irradiance: note how the various curves tend to come back together on the right, where the current goes to zero (thus, open-circuit). So I dunno if low sunlight will offer all that much protection. Might need to be so low that seeing to work could be an issue.

2. Do the connections, that involve handling bare wire, before the strings are wired. In my case, connect the THWN to the inverter. Then connect the THWN, and the cables which join the strings to the THWN (MC4 at one end, bare wire at the other end), to the terminal block in the junction box. After that all wiring is done by plugging MC4 connectors together.

3. Don't do it when the ground is wet. Wear rubber-sole shoes. Wear some kind of gloves ?

That's all I got. Maybe I worry too much.


Screen Shot 2020-06-21 at 1.56.57 AM.png

[peakbagger]

I have always done it in the dark or with the panels completely covered with cardboard. During on testing event during the daytime I got to see a very impressive arc that I do not want to see again.

[Mike90250]

Leave all breakers and switches OFF

This prevents any current (amps) flow and prevents arcs. You still have voltage potential but the MC4 connectors are well insulated. Don't perform electrical operations when it's wet or damp

Any arc generated with MC connectors is fatal to the thin gold plate internally on the contacts and the connector will overheat and fail. Let your circuit breaker do it's job of disconnecting, that's why it's got a handle on it, and why I recommend breakers over fuses (fuse holders are not rated for disconnect)

Wire all the screw terminal connections first, open the breakers and then, plug the MC connectors together. You can do a voltage test across the breaker terminals before you energize it, you should get your expected open circuit voltage (of your series string)

[RShackleford]

I have always done it in the dark or with the panels completely covered with cardboard.

I don't have enough cardboard, or a tarp big enough, so I guess I'll just do it in the dark with a headlight.

During on testing event during the daytime I got to see a very impressive arc that I do not want to see again.

What do you mean by "on testing event" ? You'd have to disconnect/reconnect the DC wires to see an arc, right ?

[RShackleford]

Leave all breakers and switches OFF

This prevents any current (amps) flow and prevents arcs. You still have voltage potential but the MC4 connectors are well insulated. Don't perform electrical operations when it's wet or damp

Any arc generated with MC connectors is fatal to the thin gold plate internally on the contacts and the connector will overheat and fail. Let your circuit breaker do it's job of disconnecting, that's why it's got a handle on it, and why I recommend breakers over fuses (fuse holders are not rated for disconnect)

Wire all the screw terminal connections first, open the breakers and then, plug the MC connectors together. You can do a voltage test across the breaker terminals before you energize it, you should get your expected open circuit voltage (of your series string)

I'm not quite following. Breakers are on the AC side - I'm connecting the inverter's AC output to the panel via an AC disconnect and a breaker in the main panel, and one or both of them will be disabled. I guess then hopefully the inverter will not sink any current from its DC inputs, so no current can ever flow (unless you accidentally connect the panels into a ring), which should prevent any arc'ing - definitely worthwhile (to prevent connector damage as you say, among other things).

However, the inverter has a DC load/break switch, and when that is switched off, the inverter can't sink any current from the strings anyhow, so it shouldn't really matter whether the AC is connected or not.

You can do a voltage test across the breaker terminals before you energize it, you should get your expected open circuit voltage (of your series string)

Don't follow at all. Not gonna see any voltage on the breaker terminals at all, without it being connected to the grid.

Are you talking about a breaker on the DC inputs ? That would explain my confusion. I don't believe I have that though, except maybe it's built into the aforementioned DC load/break switch somehow.

[bob-n]

In your case, I think Mike means having the DC load/break switch open. You need something to open-circuit the panel loop. Fronius, SolarEdge, and SMA have DC disconnect built into the inverter. That needs to be open while you do the plugging.

[RShackleford]

In your case, I think Mike means having the DC load/break switch open. You need something to open-circuit the panel loop. Fronius, SolarEdge, and SMA have DC disconnect built into the inverter. That needs to be open while you do the plugging.

Ok, thanks for clearing that up.

So the main risk is open-circuit voltages, adding (in series) in partially-connected strings. Does my interpretation of the panel's I-V curves, that Voc can still be pretty high even with very low irradiance, make sense ? As mentioned above, best mitigated by doing the potentially-shocking connections early-on and then the MC4 ones. After that, lineman's gloves are probably overkill, but maybe some sort of hardware-store glove can provide a little insurance.

[Mike90250]

Switch the inverter's AC off, that shuts it all down, with a cheap easy breaker. Then turn off the DC disconnect, which should have no power flowing anyway (through the expensive DC switch)

So now the wiring connectors are safe from a spark on connect/disconnect. But you still have the high DC voltage potential on the string as you start connecting panels in series. The MC connectors are "supposed" to be Touch Safe, but you can never tell.....

[bob-n]

You're right. Open-circuit panel output voltage can be significant, even on a dreary day, and multiple panels add up. In this plot, at no load, even with 1/5th of STC, you still get 88% of the maximum output voltage from the panel.
solarcells2.png

[bcroe]

I reserve most panel electrical work till complete darkness, when they are no
longer dangerous. One method is break a string at several points so voltages
do not add up.

Curves above look like 0.53 to 0.59 volts per cell OC variation. Bruce Roe

[solarix]

SEIA once did a test in their solar lab of covering an array with one of those blue tarps and then measuring the open circuit voltage. Still plenty of DC voltage left to kill you....
Just leave one MC4 connection on an array to be the last thing connected and you'll be just fine.

[RShackleford]

SEIA once did a test in their solar lab of covering an array with one of those blue tarps and then measuring the open circuit voltage. Still plenty of DC voltage left to kill you....

Damn, as I suspected (from I-V curves above).

Just leave one MC4 connection on an array to be the last thing connected and you'll be just fine.

Not sure how that helps. We're talking open-circuit voltages.

[J.P.M.]

SEIA once did a test in their solar lab of covering an array with one of those blue tarps and then measuring the open circuit voltage. Still plenty of DC voltage left to kill you....
Just leave one MC4 connection on an array to be the last thing connected and you'll be just fine.

I had my house tented about 4 yrs. ago. Aside from the adventures of keeping the tenters off the array (honest !), when the tent was over the array/house, the array measured zero output current and a fraction of a volt potential. That tarp was completely opaque.

At approx. 1200 - 1215 hrs. P.D.T. (11:09:26 to 11:24:26 solar time) today, under cloudless skies, I covered the solar sensor on my Davis weather station for 5 minute periods, first for 5 minutes with a sample of brown tarp, then uncovered for 5 minutes, and then covered for 5 minutes with a sample of blue tarp, both tarps are the std. stuff of the type you'll find at Big Box.

Both tarps are fairly new and not used much.

The uncovered GHI at the time varied from 1003 W/m^2 at 1200 hrs. PDT to 1013 W/m^2 at 1215 hrs., both readings uncorrected for instrument temp.

The brown tarp transmitted 0 W/m^2. The blue tarp transmitted 105 w/m^2. The tarps appear to be of similar weight and material.

Take the above FWIW.

[Mike90250]

post was held in moderation, released now

[RShackleford]

The brown tarp transmitted 0 W/m^2. The blue tarp transmitted 105 w/m^2. The tarps appear to be of similar weight and material.

So you're saying that unless I acquire a (presumably expensive) fully-opague tarp like the tenters used, I'm better off with the cheap brown tarps than the cheap blue ones ? I think even that will be a pain though; I think I'll just do it in the dark; a good floodlight out there seems easier and cheaper.