Key takeaways
- Most rejections are completeness failures, not design failures. The system was fine; the drawing didn’t prove it.
- Nine causes account for the overwhelming majority: site plan mismatch, wrong code cycle, busbar violations, thin structural detail, sizing that doesn’t reconcile, missing labels, incomplete equipment docs, undimensioned setbacks, and equipment that isn’t on the approved list.
- A rejection doesn’t cost you the correction — it costs a second position in the review queue. Fixing the drawing takes an hour. Getting back in front of a plan checker takes as long as the first review did.
- Measure your first-pass rate before trying to improve it. Most installers track close rate and install volume but have no number for how often a submittal clears on the first try.
- Avila Solar Drafting builds plan sets in 2–3 business days, 1–2 for Fast Roof, to the code cycle your jurisdiction has actually adopted.
Why do solar permits get rejected?
Most solar permits are rejected for completeness rather than design. The array would have worked, the wiring was sound, and the plan checker still sent it back — because something on the sheet was missing, contradicted something else, or was written to a code cycle the jurisdiction hasn’t adopted.
None of the nine causes below are exotic. That’s the point. A plan checker reviews dozens of PV submittals a week and develops a short list of things to look for first. Knowing that list is most of the job — and the second half of this piece covers what it costs you when you don’t.
Part one: the nine causes
1. The site plan doesn’t match the field
The drawing shows the AC disconnect on one wall and the installer mounts it on another. The inspector fails it, even when the as-built location was the better engineering call.
This is the most common correction there is, and the most frustrating, because nothing was wrong with the installation. The plan is the agreement. If the field deviates, the plan gets revised — not the other way around.
It happens because the survey and the drafting happen at different times, by different people, with different information. The tech knows the meter is on the east wall and the only sensible disconnect location is beside it. The drafter, working from photos that don’t show that wall, puts it where the layout is cleanest.
The fix is upstream. Decide equipment locations on site, with the crew’s input, photograph each one, and dimension them. Then draw what will be built rather than what would be tidy. That upstream discipline is the backbone of the workflow we lay out in residential solar design: intake to AHJ-ready plan set.
Related failures in the same family: undimensioned working space in front of electrical equipment, missing distance from meter to utility disconnect, and property line setbacks left for the plan checker to scale off the drawing. Most of the upstream design factors that generate corrections start right here.
2. Code citations written to the wrong cycle
A submittal citing NEC 2020 into a jurisdiction on NEC 2023 gets kicked back even when the underlying design satisfies both. The plan checker isn’t evaluating your reasoning — they’re checking your sheet against their adopted code.
Adoption is uneven and slow. Multiple NEC cycles are in force across different jurisdictions at any given time, and plenty of states adopt with amendments that change the answer again. NFPA publishes NFPA 70, the National Electrical Code on a three-year cycle; jurisdictions adopt on their own schedule, and some run two cycles behind.
This is the most avoidable cause on the list, and the most common one to get wrong at scale. A shop operating across four counties can easily be filing into three different cycles, using one template.
Where it bites hardest is rapid shutdown. The requirements at NEC 2023 690.12 changed materially between cycles, so a compliant 2017-era detail is not automatically a compliant 2023 detail. Our NEC solar and storage rules breakdown treats the adopted cycle as the first thing to establish, not the last.
Confirm the cycle with the AHJ before drafting. Our solar plan sets are built to the cycle the specific jurisdiction has adopted for exactly this reason.
3. Interconnection that doesn’t satisfy the busbar rule
Backfeeding more into an existing panel than its busbar allows is a hard rejection, and it’s the one that most often forces a redesign rather than a redline.
Under NEC 2023 705.12(B)(2) — what installers call the 120% rule — the main breaker plus the backfeed breaker can’t exceed 120% of the busbar rating. The numbering moved between cycles: the same rule sits at 705.12(B)(3)(2) in NEC 2020, so cite the address that matches the edition your AHJ has adopted.
Work it through. A 200 A busbar has a 240 A allowance. With a 200 A main breaker, that leaves 40 A for backfeed. A 7.6 kW inverter at 240 V draws 31.7 A continuous; at the required 125%, that’s 39.6 A, so a 40 A breaker. It fits — exactly, with zero headroom.
Two separate things get submittals rejected here. The first is exceeding the allowance outright, which means a different interconnection method: a line-side tap, a main breaker downsize, or a service upgrade. None of those are drafting fixes.
The second is satisfying the rule but never showing the arithmetic. A plan checker who has to derive your numbers will find a reason to send it back. Put the calculation on the sheet.
Worth noting for the customer conversation: a panel sitting at exactly 240 A has no room for a battery, an EV charger, or a heat pump later. Flagging that during the sale is worth more than discovering it on the next job.
4. Structural details that reference nothing real
“2×6 rafters at 24 inches on center” with no span, no attachment spacing, no fastener spec and no load figures is a placeholder, and plan checkers read it as one.
The structural sheet has to establish a complete load path: existing framing dimensions and clear span, the specific standoff or mount being used, fastener type and embedment depth, attachment spacing, and dead, live, wind and snow loads for the site.
Span is the field most often missing and the one that actually governs. A 2×6 spanning eight feet and a 2×6 spanning fourteen feet are different structures, and only one of them takes an array without reinforcement.
Load figures come from the provisions referenced by the building code edition the jurisdiction has adopted — and different editions produce different numbers for the same roof, which is why the edition gets named on the sheet rather than implied.
Where the framing won’t carry it, or the jurisdiction requires it regardless, the package needs a stamped structural review from a PE licensed in that state. Guessing which projects need a stamp is its own reliable source of rejections, and structural corrections are the expensive kind — they usually mean a site revisit rather than a drafting fix.
5. Conductor and overcurrent sizing that doesn’t reconcile
Three numbers have to agree: the calculated continuous current, the conductor ampacity, and the overcurrent device rating. If any one contradicts the others, all three read as wrong.
NEC 2023 690.8(B) requires conductors and overcurrent devices sized at 125% of continuous current. That part is simple arithmetic and rarely wrong on its own.
What gets missed is derating. Temperature correction for conduit on a hot roof, and ampacity adjustment for conductor fill, can both pull a conductor’s usable ampacity below the breaker protecting it. The wire gauge looks right in isolation and fails once the corrections are applied.
Show the derated ampacity on the sheet, not just the gauge. Show the ambient temperature and the fill count you derated from. A submittal that documents its own corrections gives the plan checker nothing to question; one that shows a bare wire size invites them to run the numbers themselves — and they will run them conservatively.
6. Missing labels and placards
Label schedules are the least interesting sheet in the package and a reliable source of corrections.
Rapid shutdown labeling, disconnect identification, warning placards at the service equipment, and directory markings all have prescribed content, wording and placement. Under NEC 2023 the rapid shutdown marking sits at 690.12(D), inside the rapid shutdown section itself — it was 690.56(C) under the 2017 and 2020 cycles. A label schedule still citing 690.56(C) into a 2023 jurisdiction points the reviewer at an address that moved.
These get missed because they feel like an afterthought next to the electrical design. To a plan checker they’re a completeness check — quick to scan, unambiguous when absent, and easy to write up.
Build the label schedule into the template rather than adding it per project, and tie the wording to the adopted cycle rather than to whatever the last package used.
7. Incomplete equipment documentation
A plan set without manufacturer cut sheets is an incomplete submittal, regardless of how good the drawings are.
The City of San Diego’s Information Bulletin 301, a representative residential PV checklist, requires manufacturer specifications for modules, racking, inverters and any energy storage alongside the drawings. Most jurisdictions ask for the same.
The subtler failure is a mismatch. A spec sheet for a different inverter model than the one on the single line diagram, or racking documentation that doesn’t cover the actual attachment being used, reads as a discrepancy rather than an omission — and discrepancies get scrutinised harder than gaps.
If the equipment changes after drafting, the cut sheets change with it. That sounds obvious and it is the single most common version-control failure in a permit package.
Storage and racking are the two components most often left out entirely. A battery gets specified in the proposal and never reaches the drawings, so the submittal carries no mounting location, no clearances and no listing documentation for it, and NEC 2023 Article 706 applies whether or not the sheet acknowledges it. Racking fails the same way: a system picked for a roof type or ground condition it was never certified for, or a brand named on the structural sheet with no model, no attachment hardware and no manufacturer certification behind it. Both read to a plan checker as equipment you have not actually chosen. Specify the storage unit and the mounting system by model at intake, and submit their cut sheets with the modules and inverter.
8. Fire setbacks and roof access pathways
Setback and pathway requirements vary by jurisdiction more than almost anything else in a residential submittal, and they’re checked against the roof plan directly.
Ridge setbacks, pathway widths and clearance to obstructions are dimensioned items, not visual ones. An array that reads as compliant by eye but isn’t dimensioned on the sheet will draw a correction asking you to prove it.
Because these vary so widely, they’re also the requirements most likely to catch out a shop expanding into a new market with a template that worked next door. The AHJ quirks worth knowing in your specific jurisdictions are mostly in this category, and they’re worth documenting per-AHJ as you learn them.
9. Equipment that isn’t on the approved list
A module or inverter that is entirely legitimate — certified, and installed by the thousand elsewhere — still gets a submittal rejected when it isn’t on the equipment list the local utility or AHJ works from.
An approved-equipment list is an enumerated set of models a reviewer will accept without further evidence. Utilities publish them for interconnection, because a grid-tied inverter has to have demonstrated the grid-support behaviour that utility requires. Some states go further and maintain a central list that utilities and incentive programmes both draw from, California’s Energy Commission being the best-known example. A handful of AHJs keep their own for racking and fire-rated components.
Territory matters more than product quality here. The same inverter can sit on one utility’s list and be absent from the neighbouring one, either because the manufacturer never submitted it there or because the listing is still pending. Nothing about the hardware changed. The reviewer simply has no entry to check it against, and “it’s approved next door” is not a category they can accept.
It bites hardest on substitutions. Supply gets tight, the warehouse swaps in an equivalent module, and nobody re-checks the list — so the drawing, the cut sheets and the interconnection application all name a model the reviewer cannot clear. Check the list before you specify rather than after: the interconnecting utility’s current list for the inverter and any storage, the state list where one governs, and the AHJ’s own requirements for racking and fire-rated components. Then check it again if the equipment changes after drafting.
Part two: what a rejection actually costs
Why does a rejection cost more than the correction itself?
Because plan review is a queue, not a conversation. When a plan checker issues corrections, your submittal leaves their desk. When you resubmit, you re-enter the queue behind everything filed since — you don’t resume where you left off.
That’s the insight most cost estimates miss. The drafting fix is trivial. The re-queue is not.
It’s also why two installers with identical drafting quality can have wildly different cash conversion cycles. One filing into a jurisdiction with a five-day median review absorbs a rejection as an inconvenience. One filing into a six-week jurisdiction absorbs the same rejection as a lost quarter.
NREL’s SolarTRACE tool exists precisely because that variation is so wide and so poorly understood. It reports median permitting, inspection and interconnection cycle times across more than 1,500 jurisdictions with at least ten installs in the dataset, spanning 140 utilities and 26 states. If you operate across multiple AHJs, that data tells you which of your markets punish a rejection hardest — and where a clean first submittal is worth the most.
Understanding the permitting workflow as a queue rather than a review changes how you prioritise. It isn’t about drafting faster. It’s about not going around twice.
What does a permit rejection actually delay?
Everything downstream of it, in sequence: permit issuance, crew scheduling, the install, the inspection, utility interconnection and permission to operate, and the final customer payment milestone. Each one waits on the one before it.
Here’s the chain most residential shops run:
- Contract signed
- Site survey and design
- Plan set produced
- Permit submitted → plan review
- Permit issued
- Install scheduled into an available crew slot
- Install completed
- AHJ inspection
- Utility interconnection / PTO
- Final payment
A rejection at step 4 doesn’t just add days at step 4. It pushes step 6 past the crew slot you’d already allocated — and crew slots don’t wait. That slot gets filled by another job or it goes empty. Either way the delayed project lands in a later month, and the final payment lands with it.
Then there’s the part nobody schedules: coordination overhead. Someone reads the correction notice, interprets it, routes it to the drafter, checks the revision, and resubmits. That work is real, it’s unplanned, and it comes out of the same person who was supposed to be selling or scheduling.
How do you calculate the cost of a permit rejection?
Multiply your rejected submittals per year by the added calendar days per rejection, then convert those days into the operational unit that actually constrains you — usually install slots per month, not hours.
You need four numbers, and three of them are yours:
| Input | Where it comes from |
|---|---|
| A. Submittals per year | Your project tracker |
| B. First-pass rate | Your permit log — if you don’t have it, this is the number to start capturing |
| C. Median plan review time in your AHJs | SolarTRACE, or your own filing history |
| D. Your correction turnaround | Time from correction notice to resubmittal |
Rejections per year = A × (1 − B)
Added lag per rejection ≈ C + D
The trap is stopping there and multiplying by an hourly rate. That understates it badly, because the binding constraint on most installers isn’t labour hours — it’s how many jobs the crew can physically complete in a month. Express the answer in displaced install slots and it becomes a scheduling problem you can act on.
A worked example
Take a residential shop filing 60 submittals a year. Suppose their permit log shows a 70% first-pass rate, their AHJs run a 12 business day median plan review, and their correction turnaround averages 4 business days. (Use your own figures — these are illustrative inputs, not benchmarks.)
- Rejections per year: 60 × (1 − 0.70) = 18
- Added lag per rejection: 12 + 4 = 16 business days ≈ 3+ calendar weeks
Those 18 rejections don’t consume 288 sequential days — submittals run in parallel. What they do is push 18 projects out of their intended install month. For a shop running roughly five installs a month, that’s the equivalent of losing more than three months of scheduled throughput to re-queuing, spread across the year.
Now change one input. Take the first-pass rate from 70% to 90% and rejections drop from 18 to 6. Nothing about the crew, the sales team, or the AHJ changed. Twelve install slots came back.
Set that against what a plan set costs. PV plan set pricing depends on system size and complexity — current pricing is on the solar plan sets page. Twelve recovered install slots is not a close comparison.
That’s the actual return on plan set accuracy, and it’s why a pre-submittal check pays for itself faster than almost any other process change available to an installer.
How do you measure your own first-pass rate?
Log every submittal with three fields: date filed, date of first AHJ response, and whether that response was an approval or a correction notice. First-pass rate is approvals divided by total submittals. Two months of data is enough to act on.
Segment it by jurisdiction as soon as you have volume. Aggregate rates hide the problem — one difficult AHJ can drag a company-wide number down while every other market runs clean, and the fix for that is jurisdiction-specific rather than systemic.
Then segment by correction type, mapped against the nine causes above. If most of your corrections are structural, the answer may be a stamped PE review on more projects rather than better drafting. If they’re electrical, it’s usually a code-cycle mismatch. If they’re site plan corrections, it’s site survey data quality. If they’re production or shade related, it’s the modelling feeding your layout.
Each of those has a different owner and a different fix. An undifferentiated “we get a lot of corrections” has neither.
Where the delay is shrinking
Standardised and automated permitting is compressing review times in jurisdictions that adopt it, which raises the relative cost of a rejection everywhere else.
The U.S. Department of Energy describes SolarAPP+ as a platform that automates solar permitting for local governments, letting adopting jurisdictions approve compliant permits instantly rather than queueing them for manual review. Where it’s in place, the queue penalty largely disappears — but only for submittals that pass the automated compliance check on the first attempt. Automated checks are unforgiving about missing or inconsistent inputs in a way a human plan checker sometimes isn’t.
The broader pressure is real. DOE reports that residential solar soft costs fell by roughly 50% between 2010 and 2020 and need to fall a further 60–70% to hit its cost targets, with permitting and interconnection named components. Installers who can’t file clean packages will carry a soft-cost disadvantage the rest of the market is actively engineering away.
Knowing what belongs in a plan set in the first place is the foundation, and judging plan set quality is the operational skill on top of it. For shops trying to lift volume without adding drafting headcount, outsourcing the drafting is the lever with the shortest payback.
FAQ
Why do solar permits get rejected?
Most commonly for completeness rather than design: site plans that don’t match the installed equipment locations, code citations written to a cycle the jurisdiction hasn’t adopted, interconnection that doesn’t satisfy the busbar rule, structural details without real framing dimensions and span, conductor and overcurrent sizing that doesn’t reconcile after derating, missing labels, incomplete or mismatched manufacturer documentation, undimensioned fire setbacks, and equipment that isn’t on the utility’s or AHJ’s approved list.
What does a solar permit rejection actually cost?
One full additional review cycle, not the hour it takes to fix the drawing. If your jurisdiction’s median plan review is two weeks, a correction on day 14 sets you back to the end of the queue for another two weeks plus your correction turnaround — and pushes the install past the crew slot you’d allocated.
What are the most common solar plan set errors?
Equipment locations that differ from what will be installed, and calculations that are correct but not shown. Plan checkers verify rather than derive — if the arithmetic isn’t on the sheet, it reads as missing.
How many revisions does a typical solar permit take?
It varies widely by jurisdiction and submittal quality. Rather than relying on an industry average, log your own submittals and track first-pass rate per AHJ — that number is actionable in a way an average isn’t.
How do I speed up solar permitting?
Raise the first-pass rate before the drafting speed. A package that clears on the first submittal beats a package drafted a day faster and rejected. Confirm the adopted code cycle before drafting, dimension everything, show every calculation including derating, and check whether your AHJ accepts SolarAPP+ submittals.
What does an AHJ look for in a solar permit submittal?
Internal consistency, chiefly. That conductor sizing, overcurrent protection and interconnection agree with each other and with the adopted code cycle, that equipment locations match what will be built, and that structural details reference real framing dimensions and span.
What NEC code cycle does my jurisdiction use?
It varies by state and often by municipality, and some states amend the cycle they adopt. Multiple NEC cycles are in force across different jurisdictions at any given time. Confirm with the AHJ before drafting, because rapid shutdown and interconnection language differ meaningfully between them.
Stop paying for the second review
Nine causes, and eight of them are caught before filing by someone checking the sheet against the adopted code and the equipment against the local approved list.
Avila Solar Drafting builds permit-ready solar plan sets in 2–3 business days — 1–2 for Fast Roof — to the code cycle your jurisdiction has actually adopted, backed by our guarantee of accurate solar plan sets and six months of free revisions. PV plan set pricing depends on system size and complexity — current pricing is on the solar plan sets page.
Commercial, utility-scale, or a jurisdiction that keeps sending things back? Call 971-410-0655.
