Electrical panel upgrade

[CraziFuzzy]

I think most inspectors will open the main and take a peak. The plan checker will look at the drawings and they ought to match what's at your home. I would question somebody taking the type of risks you are suggesting. What other corners will they cut if they think they won't get caught? Check the neighbors main electrical panels then start plotting and planning.

Not sure I'm taking any actual 'risks' with this setup, it was a question of code, not safety (don't you love when those don't agree), and unfortunately, it would likely be fully allowed under the 2014 NEC 705.12(D)(3)(d). The act of having lighter load on the side of the main opposite the solar feeder would meet the 'engineering' qualification - unfortunately, most jurisdictions won't even entertain the 2014 NEC for a few more years.

Obviously, this won't even be an issue if I can identify that it's a 125A bus, just trying to ascertain my options if I cannot (as well as spark some discussion on the matter).

[sensij]

Not sure I'm taking any actual 'risks' with this setup, it was a question of code, not safety (don't you love when those don't agree), and unfortunately, it would likely be fully allowed under the 2014 NEC 705.12(D)(3)(e). The act of having lighter load on the side of the main opposite the solar feeder would meet the 'engineering' qualification - unfortunately, most jurisdictions won't even entertain the 2014 NEC for a few more years.

Obviously, this won't even be an issue if I can identify that it's a 125A bus, just trying to ascertain my options if I cannot (as well as spark some discussion on the matter).

I'm not sure about your code citation. Did you mean 2014 NEC 705.12(D)(2)(3)(d)?

(d) Connections shall be permitted on multiple-ampacity busbars or center-fed panelboards where designed under engineering supervision that includes fault studies and busbar load calculations.

Even 2011 NEC 705.12(D)(7) potentially allows center-fed, if you can come up with a Art. 220 calculation that supports it.

(7) Inverter Output Connection.
Unless the panelboard is rated not less than the sum of the ampere ratings of all over-current devices supplying it, a connection in a panelboard shall be positioned at the opposite (load) end from the input feeder location or main circuit location. The bus or conductor rating shall be sized for the loads connected in accordance with Article 220.

[CraziFuzzy]

I'm not sure about your code citation. Did you mean 2014 NEC 705.12(D)(2)(3)(d)?

Ha... yeah, that was bad type-fu.

Even 2011 NEC 705.12(D)(7) potentially allows center-fed, if you can come up with a Art. 220 calculation that supports it.

The problem with the 2011 D.7 section, is that the to logically get to the 220 article calculation, the inverter "shall be positioned at the opposite (load) end from the input feeder location or main circuit location", making use of that article in center-fed panels technically impossible. That is why they added the provisions in the 2014 NEC for it. If I could just move my house to Colorado, I'd be good!

[sensij]

The problem with the 2011 D.7 section, is that the to logically get to the 220 article calculation, the inverter "shall be positioned at the opposite (load) end from the input feeder location or main circuit location", making use of that article in center-fed panels technically impossible. That is why they added the provisions in the 2014 NEC for it. If I could just move my house to Colorado, I'd be good!

Do you mean the logic of this statement?

"Unless the panelboard is rated not less than the sum of the ampere ratings of all over-current devices supplying it"

Read the 2011 version a few more times. I think it means if your panelboard rating is greater than or equal to that of the supplying OCPD's, you can consider positioning the inverter input at someplace other than the opposite end from the feeder. My understanding is that the 2014 wording wasn't meant to change the interpretation of center-fed installations, but just make it it clearer. However, getting your AHJ to agree might be challenging, either way.

[CraziFuzzy]

Do you mean the logic of this statement?

"Unless the panelboard is rated not less than the sum of the ampere ratings of all over-current devices supplying it"

Read the 2011 version a few more times. I think it means if your panelboard rating is greater than or equal to that of the supplying OCPD's, you can consider positioning the inverter input at someplace other than the opposite end from the feeder. My understanding is that the 2014 wording wasn't meant to change the interpretation of center-fed installations, but just make it it clearer. However, getting your AHJ to agree might be challenging, either way.

That clause in (7) is an additional restriction that is somewhat based on (2). (2) says that the sum of the supplying OCPDs must be <120% of the bus rating. (7) says (parsing out the triple negative), that if the bus isn't rated for more than the sum of the supplying OCPD's, you have to install the inverter feed at the opposite end. In my case, I've got a (possibly) 100A bus, fed with a 100A main, and am wanting to add a 20A inverter supply to it. With that, the sum will be 120A, which means the panel rating will NOT be less than the sum of the supplying OCPD's, which means the first sentence of (7) will not be met. This clause is what allows you to derate the main to allow capacity for an inverter install on a center-fed panel, in my case, potentially downrating the 100A main to an 80A main would allow the installation of the 20A inverter. in which case, any location on the panel would suffice. What is likely preventing me from downsizing to such a small main feeder, however, is the article 220 requirement.

[sensij]

Yes, with a 100 A busbar you will not be able to meet (7). With a 125 A center-fed busbar you might be able to, but that will be true under 2011 or 2014. The relief that 2014 really offers is with respect to that fact you can use 125% of the inverter output circuit current, instead of using the size of the OCPD. With string inverters, it won't matter much, but if you use micros, you can get a few more in before the service panel upgrade would be needed.

[CraziFuzzy]

Yes, with a 100 A busbar you will not be able to meet (7). With a 125 A center-fed busbar you might be able to, but that will be true under 2011 or 2014. The relief that 2014 really offers is with respect to that fact you can use 125% of the inverter output circuit current, instead of using the size of the OCPD. With string inverters, it won't matter much, but if you use micros, you can get a few more in before the service panel upgrade would be needed.

That is not entirely true. With a proper engineering design, the limit of the bus has less to do with the actual limit on the source sizing. By doing a proper worst case busbar load calculation, you'll find that you can have a 100A center-fed bus, with a 20A feeder on one extreme end, and as long as the max load on the opposite side of the main is less than 100A, it will not exceed 100A on any point of the bus. That is the key thing that changed in 2014, is that the 125% is not a hard limit, and neither is the 100%, as long as a proper engineering calculation can show no overload possible.

In fact, you could actually have a 100A center-fed panel, and a 100A end-connected inverter, and as long as the sum of the OCPD's on either side of the center feed do not exceed 100A, and no individual load exceeds the 100A, you will not have over 100A at any single point on the bus, which could justifiably meet 2014 NEC 705.12(D)(2)(3)(d)'s requirements. In reality, few will end up using (d), due to the extra step of the load calc, and will try to stay under the 120% rule for end fed or, if center fed, simply tell the homeowner that a new panel is required - but for those companies that are willing to do it, it opens up a lot more options - once 2014 is accepted in each jurisdiction.

Sadly, I won't be able to wait that long, so I'll have to go an alternate route.

[inetdog]

That is not entirely true. With a proper engineering design, the limit of the bus has little to do with the actual limit on the source sizing. By doing a proper worst case load study, you'll find that you can have a 100A center-fed bus, with a 20A feeder on one extreme end, and as long as the max load on the opposite side of the main is less than 100A, it will not exceed 100A on any point of the bus. That is the key thing that changed in 2014, is that the 125% is not a hard limit, and neither is the 100%, as long as a proper engineering calculation can show no overload possible.

In fact, you could actually have a 100A center-fed panel, and a 100A end-connected inverter, and as long as the sum of the OCPD's on either side of the center feed do not exceed 100A, and no individual load exceeds the 100A, you will not have over 100A at any single point on the bus, which could justifiably meet 2014 NEC 705.12(D)(2)(3)(d)'s requirements. In reality, few will end up using (d), due to the extra step of the load calc, and will try to stay under the 125% rule or if center fed, simply tell the homeowner that a new panel is required - but for those companies that are willing to do it, it opens up a lot more options - once 2014 is accepted in each jurisdiction.

Sadly, I won't be able to wait that long, so I'll have to go an alternate route.

Your reasoning applies perfectly well to the physics of heating in the panel buses. But it does not meant that the same arrangement satisfies the conditions written in the NEC, which do not handle center fed panels well at all.

[CraziFuzzy]

What's most upsetting, is that my "center-fed" panel is lopsided, with only 2 positions on each bus above the main, and the rest below. I'd love to be able to blank off and remove those positions making it functionally an end-fed panel with fewer positions, but I doubt that would be accepted as an option. The inspector is likely to not accept a label as the only thing preventing those spots from being filled, and actually removing them would void the UL listing (that I can't even find on my panel, with the label completely gone).

[CraziFuzzy]

Your reasoning applies perfectly well to the physics of heating in the panel buses. But it does not meant that the same arrangement satisfies the conditions written in the NEC, which do not handle center fed panels well at all.

The 2014 version does:

705.12(D)(2)(3)(d) Connections shall be permitted on multiple-ampacity busbars or center-fed panelboards where designed under engineering supervision that includes fault studies and busbar load calculations.

That is what I was discussing - 2014 brought in the ability to get a usable sized solar connection onto an existing center-fed panel, but as it is 'complicated' in how the loads are arranged, they left it to the engineers to handed in each case instead of a blanket 120% of bus rating. I would agree, that the 2011 and older codes make center-fed much more difficult.

In reality, on a single-fed panel, a center-fed design is thermally superior, and why it was a common design for quite some time, keeping peak spot bus current lower in all but extreme cases than the same loads on an end-fed arrangement. Adding a solar feeder can still maintain this greater safety margin, but requires potential loads to be spread properly. Earlier codes didn't allow this flexibility, 2014 does - it's just not adopted in CA (or most of the country) yet.

[inetdog]

The 2014 version does:


That is what I was discussing - 2014 brought in the ability to get a usable sized solar connection onto an existing center-fed panel, but as it is 'complicated' in how the loads are arranged, they left it to the engineers to handed in each case instead of a blanket 120% of bus rating. I would agree, that the 2011 and older codes make center-fed much more difficult.

In reality, on a single-fed panel, a center-fed design is thermally superior, and why it was a common design for quite some time, keeping peak spot bus current lower in all but extreme cases than the same loads on an end-fed arrangement. Adding a solar feeder can still maintain this greater safety margin, but requires potential loads to be spread properly. Earlier codes didn't allow this flexibility, 2014 does - it's just not adopted in CA (or most of the country) yet.

Good point.
Most AHJs will still interpret that to require a signed and sealed drawing and/or statement from a PE, not just the opinion of the installer with a contractor's license or a licensed electrician.

[CraziFuzzy]

Good point.
Most AHJs will still interpret that to require a signed and sealed drawing and/or statement from a PE, not just the opinion of the installer with a contractor's license or a licensed electrician.

Agreed a PE stamp may be required, but that's likely far cheaper than an unnecessary new panel. I have read a few articles that seem to expect many AHJ's to issue some guideline in this regard. As I mentioned above, it is just a matter of ensuring the total OCPD sizing opposite the main doesn't exceed the bus rating. That is a relatively easy guideline to prove and follow, so it might be possible by 2020 to do it without the stamp.

[SoCalsolar]

Would also need

Would also need an 80 amp main breaker usually $150+ labor,engineers stamp $200+ Likely at $4-500 vs. Main panel Upgrade to 125 amp $1200-$1600. Could make sense.

[CraziFuzzy]

Would also need an 80 amp main breaker usually $150+ labor,engineers stamp $200+ Likely at $4-500 vs. Main panel Upgrade to 125 amp $1200-$1600. Could make sense.

Derating to 80A main would be the easiest part, and in that case, wouldn't require an engineer at all. But that may not be a possibility. Technically, you can't derate the service disconnect to less that 100A.

One-Family Dwelling. For a one-family dwelling, the service disconnecting means shall have a rating of not less than 100 amperes, 3-wire.

Now, I'm not saying this is often enforced, as there are an awful lot of solar installs in my area without upgraded panels, but by the letter of the code, 100A is the smallest you can go, which means on my center-fed panel, technically I couldn't even add a single amp of power. In fact, line-side would technically be out, as I'm not sure there's a connection point between the meter and the main breaker (direct buss bar from meter socket to main breaker). About the only options, if it IS a 100A buss (still undetermined, as three neighborhood houses I've checked so far, none have a sticker), would be separately metered solar tied in before my main service meter, or panel replacement.

My guess is the local AHJ is assuming 125A busses (replaceable main breaker plus construction in the 90's probably backs that assumption up - I just hate not seeing it in writing), and that is how the other installs have been done in the neighborhood, but I haven't gotten a chance to chat up any of the neighbors with the actual solar.

[inetdog]

Even in residential, you can have more than one service disconnecting means (up to six motions of the hand worth) grouped together, all of which can be smaller than or up to the service size, with no limit on the sum of their ratings. That does not help the OP's situation, but it is an interesting option for some other situations, such as supply side connections.
Not that the PV disconnect is a service disconnect anyway.