diy solar

diy solar

SolArk Install Plan

Lots of us have legacy systems from earlier years without the same version of NEC, to the fireman just has to take his chances/use his judgement.
If he puts a fire ax through a PV panel or steps on it, more likely to be a problem.

Even the walkway to access roof is interpreted differently in different places.
If he walks on the eves which overhang beyond the wall of a 2 story house, might that collapse?
One location incudes that in 3' walkway, another say 3' inside the supporting wall.

Newest code only requires a walkway on the side fireman would climb, the North East side in my case closest to the street. That lets me panel over to the eves on South West side.
Also, 3' wide area straddling the ridge 18" on either side, so he can choose which side to vent (depending on wind direction) rather than 3' clear on both sides like previous revision NEC.

I think the code should have allowed no RSD so long as wires outside the array are run in metal conduit.

There are some model RSD good for 4 panels per unit. Depending on voltage/current limits that might be the most economical.
If I relocate some 120W 12V panels from ground mounts onto roof, I'm thinking 3 in series for 360W 67 Voc, and 4 of those strings into one RSD.
Then I'll only need 2 RSD per 24 panel 2700W string.

If you can do without RSD, so much the better.
Arc-fault is probably a more useful feature to protect your house from having a fire in the first place.
 
Arc-fault for PV strings is what I mean in particular. If wires get damaged or MC connector corrodes, hundreds to thousands of watts can be dumped into an arc at that location. Arc-fault opens both PV+ and PV- so less likely a complete circuit is available to sustain the arc.

In the house AC wiring, latest code requires arc fault for most 120VAC outlets, maybe other circuits. If a connection is bad or someone puts a picture hanger screw through a wire, that is supposed to disconnect and avoid starting a fire.

In house wiring, brush-type motors (like your vacuum cleaner) may cause nuisance trips.

In PV strings, radio interference may cause nuisance trips. How the wires are arranged may help by making a poorer antenna, less RF pickup. But I think the PV panel cell arrangement is an unavoidable antenna.

I've read that SMA string inverters were particularly bad at being confused by RF, leading to some installers abandoning them for other brands.

The system will carry large signals at 60 Hz, at the inverter and MPPT switching frequencies (perhaps 10 kHz), and periodic multiples of those (e.g. 120 Hz, 180 Hz, 20 kHz, 30 kHz) all the way into MHz. Those are to be ignored.
Arc-fault produces broad-band noise rather than periodic frequencies, but is a thousand times lower in amplitude than those.
So the trick for arc fault is see if there is random noise buried in the weeds among strong periodic signals.
It does that with a digital signal processor. So much complexity, and not all programmers are competent at their craft.

Probably Sol-Ark has arc fault, to meet requirements for rooftop panels.
 
I am planning to use this Reliance R310A transfer switch instead of the critical loads panel. I think it is the one SolArk sells and looks like the pictures in the manual.
It is the outdoor version. My SolArk is the outdoor version. They'll be in my garage but it's damp and salty here.

The is also the 50A version. I'll have to decide whether I need that - https://www.electricgeneratorsdirect.com/Reliance-Controls-R510A/p61744.html

For some reason these transfer switches are hard to find. It seems like there is a shortage.
 
Last edited:
"Factory-Direct Ships in 4-6 Weeks"
"Manufacturing Lead Times are fluid and changing. Get current order status and support now."

Apparently built when ordered.
Perhaps components like the enclosure are ordered from suppliers.

Don't know the cycles of components like switches and breakers, but the electronic components we ordered from DigiKey were typically stocked at the distributor, sometimes had to wait for next factory run (or scrounge up from other outlets.) With so many different resistors, capacitors, ICs, they aren't continuously produced. The fab or other factory probably rotates through making various part numbers to fill large orders. "Lead time" was typically shown as 12 to 26 weeks, sometimes longer.

The inexperienced guy ordering builds of our product would ask me for alternative parts when the PCB assembly he ordered didn't have entire BOM available with 3-days lead time. I told him to order the entire BOM for however many boards he anticipated having assembled the next quarter. He said that was too much money ($60k). I told him that was just 2 days burn rate for the company, and if he didn't order the parts we could slip deliveries out 3 months. Never listened to me.
 
I got mine, same part number, from an eBay seller mid October.
I forget if that was one of the orders that was actually fulfilled by Home Depo; some have been.

There's probably been an increase in demand by preppers from all walks of life.

The absurdity of interlock prices is clearly shown here:



A sheetmetal interlock costs more than the metal box with busbars.

(the $10 difference is reversed for my preferred QO, which has copper busbar for 125A and larger breaker panels.)
 
I got mine, same part number, from an eBay seller mid October.
I forget if that was one of the orders that was actually fulfilled by Home Depo; some have been.

There's probably been an increase in demand by preppers from all walks of life.

The absurdity of interlock prices is clearly shown here:



A sheetmetal interlock costs more than the metal box with busbars.

(the $10 difference is reversed for my preferred QO, which has copper busbar for 125A and larger breaker panels.)
Now I'm thinking about Amazon's free returns.
Thanks Hedges
 
I paid $120 when I got my QO1DM10030TRBR. Today I see them on eBay for $160 to $200. Home Depo is $165.
Originally I got a 63A DIN transfer switch (and circuit breaker, I think) out of China from an eBay vendor.
It made buzzing noises when power was present on the switched off branch, so I didn't trust it, replaced with the above Square D generator switch.

I have a couple of the sheetmetal interlocks as well. One is on an outdoor 225A panel so I can manually feed garage from inverter. At this time same panel feeds the inverter, so I manually turn that off. Once I move inverter to a line-side tap (actually tap between main disconnect and 225A panel) it won't try to chase its tail. Another will go on 125A panel in house after I pull wire for it.
 
I searched hard for those transfer switches. There were a bunch of suspicious links and they all had the same $850 price.
Most of the links that claimed they had the switchboxes looks so suspicious I wouldn't order from them.

I ordered the Square D switchbox from Amazon because it was only $20 more than Home Depot or Lowes and it has free returns until Jan 31st.
Amazon returns are the smoothest thing ever. I could probably return that thing even if I knocked out the breaker openings.
 
I previously got something like this - two DPST breakers, interlocked. Think it was a different brand. Didn't trust it when it made noises.


I saw different one on other sites, I think from a German or other European vendor. Difference was just switches, not breakers. The two DPST sections faced opposite directions, implemented On-Off-On. Those are probably good.

I feel safer with Square D. But rather than breakers, I'd prefer knife switch, either DPDT or two interlocked DPST. Square D has those, but larger and more expensive.
 
This is my updated plan. Sorry it's blurry but I don't know of a convenient way to make it better.

This is the AC disconnect - https://www.altestore.com/store/enc...b-60a-240vac-unfused-disconnect-2-pole-p8729/
I guess that would be considered a DPST knife switch.

I think the Transmitters should say RSD Transmitters.

I am probably going to submit this soon.

edit: the differences are that the critical loads panel has been replaced with the Reliance R510A transfer switch and RSD has been changed to string level.

1608244985600.png
 
Last edited:
Looks like critical loads panel gets only hot red but not hot black from main panel, even though it gets both red and black from SolArk.

I like to have hot and neutral follow the same path, not split and wrap around ferrous metal. It appears hot red from main panel completes circuit through neutral routed a different way, through SolArk.

RSD transmitter would have a switch located somewhere the fire department can reach it? I suppose it is really a keep-alive, not a shutdown, so if everything is off the panels are isolated. Powered off your critical loads panel, so PV works during power failures.

Ground would bond inside the visible-blade disconnect. That isn't explicitly shown as in other boxes.

Does critical loads panel show a bond between neutral and ground, same as in main load panel? Should be only at main.

1" EMT ... oversize conduit would make it easer to stuff wires, and could add more if you want to route an extra circuit so another panel later.
 
Looks like critical loads panel gets only hot red but not hot black from main panel, even though it gets both red and black from SolArk.
I noticed that. Is the extra hot to make 240V?
I like to have hot and neutral follow the same path, not split and wrap around ferrous metal. It appears hot red from main panel completes circuit through neutral routed a different way, through SolArk.
I agree with you but the Solark pictures show similar. I don't know what is typical regarding neutrals and grounds. I see a lot of variation in pictures and my own main panel has neutrals and grounds going to each other like they are both the same.
RSD transmitter would have a switch located somewhere the fire department can reach it? I suppose it is really a keep-alive, not a shutdown, so if everything is off the panels are isolated. Powered off your critical loads panel, so PV works during power failures.
Yes, there will be a switch. I will try to get that clarified.
Ground would bond inside the visible-blade disconnect. That isn't explicitly shown as in other boxes.

I will look into that. Are you saying the green wire should have what appears to be a blue breaker like the red and black wires have?
1608253038148.png
Does critical loads panel show a bond between neutral and ground, same as in main load panel? Should be only at main.
I think the diagram only shows bond in main panel. And the manual says to only connect them together in the main panel.

1" EMT ... oversize conduit would make it easer to stuff wires, and could add more if you want to route an extra circuit so another panel later.
I will look into bigger conduit. There will be a lot of wires going in there. The diagram only shows one of each wire.

Thanks so much for your input Hedges. I don't know how much information is supposed to be on this plan. They are flexible from what I gather.
But I need to understand everything so it is be best for me if more is shown on the plan. I am fairly sure AltE will fix it up for me.
 
Yes, two hots from opposite legs would make 240, so 120/240 available in protected loads panel from main if selected, not just from SolArk

Neutral and ground will go to the same busbar in main panel. Different busbar in all other panels.

Even if SolArk documentation authors don't know hot and neutral should be adjacent (even twisted), I would route the wires that way.

No not a "blue" breaker on ground wire, but a screw into the enclosure connecting ground wire. My switch had a busbar meant for neutral, also a green box bonding screw. I didn't connect the white neutral but rather the ground, and I bonded it to box.

I saw separate busbars for ground and neutral in both main panel and critical loads panel, but also a line between the two busbars in both panels. Thought that represented a connection, which doesn't belong in critical loads panel.

My 2", 1 1/4", 1", 3/4" conduit are all getting hard to put wires through. Even well below what the conduit fill charts say is allowed.
If you can run a 2", much more will fit. If perfectly straight, no bends, much easier. I got a Slug Buster and punched holes in the sides of the boxes for a straight conduit.

Sometimes plan can be "one-line", doesn't always break out L1/L2/N/G.
 
Yes, two hots from opposite legs would make 240, so 120/240 available in protected loads panel from main if selected, not just from SolArk

Neutral and ground will go to the same busbar in main panel. Different busbar in all other panels.

Even if SolArk documentation authors don't know hot and neutral should be adjacent (even twisted), I would route the wires that way.

No not a "blue" breaker on ground wire, but a screw into the enclosure connecting ground wire. My switch had a busbar meant for neutral, also a green box bonding screw. I didn't connect the white neutral but rather the ground, and I bonded it to box.
I understand now
I saw separate busbars for ground and neutral in both main panel and critical loads panel, but also a line between the two busbars in both panels. Thought that represented a connection, which doesn't belong in critical loads panel.
You are right, I will ask them to remove the red line between the ground and neutral in the critical loads transfer switch.
My 2", 1 1/4", 1", 3/4" conduit are all getting hard to put wires through. Even well below what the conduit fill charts say is allowed.
If you can run a 2", much more will fit. If perfectly straight, no bends, much easier. I got a Slug Buster and punched holes in the sides of the boxes for a straight conduit.

Sometimes plan can be "one-line", doesn't always break out L1/L2/N/G.
I will see if I can mount the panels and inverter next to each other and if I can use big conduit from side to side.
 
My AC disconnect back in 2003 was required to be visible-blade. That was for PG&E to be certain the system was disconnected from grid.
Later, the requirement was dropped but utility could unplug meter to accomplish the same.

A generator transfer switch or interlocked breaker doesn't have that large a gap between conductors. Some models, the interlock is defeated if cover removed from breaker panel.

At work when "Lock-out/Tag-out" was discussed, I asked, Why not "Lock-out/Short-out"?

I see panel setbacks from roof edges in my area now, but older installations didn't have the restriction.
Not likely to be local rules against firemen doing what is necessary to make it safer to enter the building.
 
"Factory-Direct Ships in 4-6 Weeks"
"Manufacturing Lead Times are fluid and changing. Get current order status and support now."

Apparently built when ordered.
Perhaps components like the enclosure are ordered from suppliers.

Don't know the cycles of components like switches and breakers, but the electronic components we ordered from DigiKey were typically stocked at the distributor, sometimes had to wait for next factory run (or scrounge up from other outlets.) With so many different resistors, capacitors, ICs, they aren't continuously produced. The fab or other factory probably rotates through making various part numbers to fill large orders. "Lead time" was typically shown as 12 to 26 weeks, sometimes longer.

The inexperienced guy ordering builds of our product would ask me for alternative parts when the PCB assembly he ordered didn't have entire BOM available with 3-days lead time. I told him to order the entire BOM for however many boards he anticipated having assembled the next quarter. He said that was too much money ($60k). I told him that was just 2 days burn rate for the company, and if he didn't order the parts we could slip deliveries out 3 months. Never listened to me.
This "Factory Direct" seller has awful reviews online and with BBB
If availability doesn't change after the holidays I will have to find a different transfer switch.

I see panel setbacks from roof edges in my area now, but older installations didn't have the restriction.
Not likely to be local rules against firemen doing what is necessary to make it safer to enter the building.
My plan doesn't have the setbacks dimensioned. I'd hate for the inspector to make me relocate the panels and racking.
 
I emailed the county about the setback from the ridge to the panels. They replied 36 inches.
It's probably better for me to have some space so I don't fall on the panels. ?
1608576871995.png
 
Back
Top