Thursday's purchase order, and a clock you do not own
The quote came back signed on Tuesday. Scaffold is booked for the fourteenth, the hybrid inverter is on a four-week lead, and the wholesaler wants the order placed today to hold that slot. The drawing shows a single hybrid unit and ten kilowatt-hours of battery in the garage. Somewhere between the survey and that purchase order, somebody has to decide whether this installation is one you switch on and then tell the network about, or one you ask permission for and then wait.
Those are not two versions of the same form. The first is a notification: you commission, you register the installation, and the job is finished on the day the scaffold comes down. The second is an application against a design, answered on the distribution operator's published timetable rather than on yours, and it can come back with conditions attached — an export limit, a curtailment requirement, or a contribution towards reinforcing a bit of network four streets away. Any of those changes the equipment list, and an equipment list that changes after the order has been placed is a restocking charge and an apologetic phone call.
So the decision sits upstream of the order, because what gets declared is a specific inverter model with a specific rating and a specific type-test reference. That is also why the shortcut everyone reaches for — drop the inverter a size and stay under the threshold — needs working through properly rather than in the van. It is often the right answer. It is sometimes an expensive way of throwing away a fifth of the yield on a roof that will never be re-scaffolded.
What is being assessed is alternating current, at the boundary, per premises
The network operator has no interest in kilowatt-peak. It cannot see the array, and it never sees direct current at all. What it assesses is the maximum the installation can deliver into its network at the point of connection, which is the inverter's alternating-current output rating and nothing else. This single fact resolves most of the confusion on site: a 4.8 kWp array behind a 3.68 kW inverter is a 3.68 kW connection, and always was.
In Great Britain the split is set by two documents from the Energy Networks Association. Engineering Recommendation G98 covers fully type-tested micro-generators up to and including 16 A per phase, which at nominal 230 V is 3.68 kW on a single-phase supply and around 11 kW spread across three phases. Inside that limit, a single micro-generator at a domestic premises can be installed and commissioned first and notified to the distribution network operator afterwards, within the period G98 sets. Above it, Engineering Recommendation G99 applies and the sequence reverses: you apply, the operator assesses, and you connect on the terms of the offer. The trap is that G98 also carries an apply-before-you-connect route — for premises that already have generation, and for cases outside the straightforward domestic one — so being under 16 A per phase does not automatically mean the notify-after path is open to you.
The threshold is a property of the premises, not of the box you are installing. An existing 3.68 kW array on the same supply does not stop counting when you add a hybrid; the two aggregate, and the site is assessed on the total. Battery discharge aggregates too, and this is where hybrid inverters catch people out, because plenty of them are rated to discharge the battery at a higher power than they are rated to convert from the panels. If the unit can put 5 kW onto the network at nine in the evening from the battery alone, then 5 kW is what it can put onto the network, and the fact that the sun set three hours ago is not a mitigation the assessment recognises.
Outside Great Britain the shape of the decision is identical and the numbers are not. In the United States the interconnection procedure is set state by state, with the equipment side governed by IEEE 1547 and inverters certified against it under UL 1741 including its grid-support supplement, while NFPA 70 Article 705 governs how the source is landed on a busbar inside the building. In Australia and New Zealand, AS/NZS 4777.1 covers the installation and AS/NZS 4777.2 the inverter, but the capacity you are permitted to connect and export is a matter for the local distribution business and differs between them. Find the governing procedure for the actual address before you size anything, and get the answer in writing.
| Line on the quote | Counted? | Reason |
|---|---|---|
| Array size in kWp | No | Direct current never reaches the network; the assessment stops at the inverter terminals |
| Inverter continuous AC output rating | Yes | This is the figure on the type-test record and the figure the notification or application asks for |
| Battery discharge power | Yes, wherever it can be exported | A unit that discharges at 5 kW can deliver 5 kW to the network irrespective of irradiance |
| Battery capacity in kWh | No | Energy is not power — the kWh figure drives fire separation, siting and runtime, not the connection class |
| Generation already on the supply | Yes, aggregated | The limit belongs to the premises, so an earlier array is added to the new one |
| Backup output on an islanded circuit | Only if it can reach the network | Depends on the transfer arrangement being genuinely incapable of parallel operation; the network operator rules on that, not the installer |
Fixing the AC rating first, then letting the roof follow
Every other design decision on this job now hangs off one number, so choose it deliberately rather than letting the inverter be picked by whichever hybrid the wholesaler has in stock. Decide the connection route, take the AC rating that route allows, and only then work out how much array that rating will carry. Doing it the other way round — array first, inverter to suit — is how a job that would have gone in under notification ends up in an eight-week queue for the sake of half a kilowatt.
The rating you want is rarely the array's nameplate, because an inverter matched to nameplate DC spends almost its whole life loafing. Deliberate over-sizing of the array relative to the inverter is standard practice, and it converts hours of partial output into energy at the cost of trimming a handful of bright, cold midday peaks. That trade is what makes the notification route survivable: a smaller inverter does not cost you a proportional share of the annual yield, it costs you the top slice of a few hundred hours.
Run it in reverse for this decision — the answer it returns is the AC rating, so raise the array DC figure until that answer stops at the rating your chosen route permits.
Total nameplate DC watts of the panels.
How much larger the array is than the inverter.
Soiling, wiring, mismatch, temperature and inverter efficiency.
Inverter AC rating
5,000 W AC
Capacity sizing only. String voltage limits, MPPT window, temperature coefficients and the local wiring rules all constrain the final selection.
- Inverter rating
- 5 kW AC
- Array DC rating
- 6 kW DC
- Realistic peak AC after losses
- 5.16 kW
- Headroom over realistic peak
- -3.1 %
They open the calculator with your figures already in it
Solar Inverter Sizing Calculator: 5,000 W AC — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 5,000 W AC — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- Does not check string voltage against the inverter's MPPT window or its maximum input voltage — cold-weather open-circuit voltage is what breaks inverters and it depends on the panel's temperature coefficient.
- Does not consider grid export limits, which in several markets cap the inverter rather than the array.
- Shading is not modelled. Partial shading is better addressed with optimisers or microinverters than with sizing.
How much roof a capped inverter will still take
With the AC rating fixed, the array becomes an optimisation rather than a constraint. Ratios in the region of 1.1 to 1.3 are common practice, so a 3.68 kW inverter comfortably carries anything from just over four kilowatt-peak to about four and three-quarters depending on the plane, and on a shaded or poorly oriented roof it will carry more, because the array's real output curve is flatter than its nameplate suggests. A north-European east-west split behind a capped inverter loses very little to clipping; a clear south plane at low latitude loses a good deal more.
Two hard limits sit on top of that reasoning and neither is negotiable. The first is the manufacturer's maximum DC input power and maximum input current per tracker, printed on the datasheet for the exact model, and exceeding it voids the warranty whatever the ratio arithmetic says. The second is the roof: the usable rectangle after setbacks, obstructions and row spacing is what it is, and the array sizing exercise has to be reconciled against a measured area rather than a plane area lifted off imagery.
There is a commercial argument for filling the roof even under a cap, and it deserves stating to the customer explicitly. Scaffold, access, labour and the fixed costs of the day are spent whether you install fourteen modules or nineteen. The marginal modules are close to material cost. If the roof will never be economically re-scaffolded, the array you install now is the array the building has, and a deliberately generous DC side behind a modest inverter is often the version of this job that ages best.
It works in daily kilowatt-hours, so bring a daily figure and the site's sun hours, then check the module count it returns against both ceilings: the inverter's maximum DC input and the rectangle the survey measured.
The amount of energy (in kWh) you want the array to produce per day, on average.
The rated output of a single panel under standard test conditions.
Your location's average 'peak sun hours' per day, not total daylight hours.
Accounts for inverter losses, wiring losses, panel soiling, and temperature derating.
Solar panels needed
7 panels
Actual output varies with roof orientation, tilt, shading, and seasonal sun-hour changes — a solar installer's site-specific production estimate (often via satellite/aerial modeling) is significantly more accurate than this general planning figure.
- Estimated daily output per panel
- 1,600 Wh
They open the calculator with your figures already in it
Solar Panel Array Sizing Calculator: 7 panels — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Estimated cost — your price
This site holds no price list for this material — local prices vary too much to publish honestly. Enter your supplier's price and the result is costed with it.
What this calculation does not cover
- This is an energy-balance estimate, not an electrical design. It does not check module voltage and current against an inverter's operating window, conductor and overcurrent sizing, disconnect or rapid-shutdown requirements, or whether your service panel can accept the array's back-fed breaker. Grid-tied work normally needs a permit and a utility interconnection approval before anything is ordered.
- Nothing here checks that the panels fit. The count takes no account of module dimensions, the usable area of each roof plane, row spacing, walkways and fire-access setbacks, or obstructions such as vents, chimneys and dormers - on most roofs the space left after those subtractions is the real limit, not the energy target.
- The sun-hours figure is taken at face value with no correction for orientation, tilt or shade. Published peak-sun-hour averages assume a well-oriented, unobstructed plane, so an east/west, shallow or shaded roof produces less than this count implies. A single efficiency percentage also cannot represent shade that moves across the array through the day and the year, or the difference between a string inverter, where one shaded module drags its whole string, and microinverters or optimisers.
- One average sun-hours number hides the seasonal swing. An array sized to a yearly average falls well short in midwinter and overshoots in summer; off-grid systems are normally sized on the worst month instead. Batteries, days of autonomy and depth of discharge are outside this calculation entirely.
- No allowance is made for panel output declining over the system's life, for load growth such as an EV charger or a heat pump, or for utility rules that cap system size or restrict export. Any of those can change the number of panels you should actually buy.
A battery is a load on the way in and a generator on the way out
On a grid-tied hybrid the battery is not there to survive the night in the off-grid sense. It is there to move the middle of the day into the evening, and that is a much smaller job than autonomy. Size it from the load that actually falls between sunset and sunrise on a shoulder-season day, not from a total household consumption figure and not from the number of days anybody wants to be independent for, because the grid is still there and the seasonal shortfall is not something a domestic battery has ever addressed.
Under an export cap the sizing question changes shape again, and usefully. Storage becomes the mechanism that stops clipped or curtailed energy going to waste: output that the cap would have thrown away in the middle of the day goes into the battery instead and comes back out in the evening at full retail value. That makes a battery more valuable behind a limit than without one, which is worth putting in front of a customer who is weighing an approval wait against a capped install.
Capacity carries its own regulatory weight, separate from anything the network operator does. Where a battery may be sited, what it must be separated from, and what aggregate capacity is permitted in a dwelling are governed by NFPA 855 and the International Residential Code's stationary storage battery provisions in North America, and by BS 7671 together with the IET Code of Practice for Electrical Energy Storage Systems in the United Kingdom. A garage installation that satisfies the network operator and fails the fire separation requirement is still a failed job.
Put the evening load in where it asks for the daily one and set autonomy to a single day — a grid-tied battery is covering one night, not a run of them. The nameplate it returns is capacity; the discharge rating that goes on the connection paperwork is a separate number off the inverter datasheet.
Energy used per day.
Days the bank must run with no charging.
Chemistry sets how much of the nameplate you can actually use.
Nameplate capacity required
11.7 kWh nameplate
Nameplate needed to deliver the stated usable energy. Real capacity falls with age and with temperature — a bank at 0 °C (32 °F) delivers materially less than its rating.
- Energy required at the load
- 10 kWh
- Usable share of nameplate
- 85.5 %
- Depth of discharge
- 90 %
- Round-trip efficiency
- 95 %
- At 48 V nominal
- 243.66 Ah
They open the calculator with your figures already in it
Battery Bank Sizing Calculator: 11.7 kWh nameplate — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 11.7 kWh nameplate — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- Capacity fades over life. Most warranties guarantee 70-80% of nameplate at end of term, so a bank sized exactly today is undersized in year eight.
- Cold reduces available capacity substantially, and lithium cannot be charged below freezing without a heater.
- Says nothing about power. A bank with enough energy can still be unable to start a motor — that is a continuous and surge kW question, not a kWh one.
Hybrids close the string window earlier than string inverters do
Having settled the AC side, the DC side has to fit an inverter that was chosen for a reason other than its trackers. Hybrid units commonly offer fewer maximum power point trackers than a comparable string inverter, a lower maximum input current per tracker, and a higher start voltage, because the DC bus has a battery on it as well as an array. On a roof split across three planes that combination can leave you with no legal string arrangement at all, and it is far better to find that out with a datasheet in front of you than with modules on the scaffold.
Both ends of the voltage window still have to hold simultaneously, and both are set by temperature. The upper bound is fixed by the coldest condition the array will ever see with light on it, using the manufacturer's temperature coefficients for the exact module and the design minimum temperature from the climatic data the adopted wiring rules reference. The lower bound is the hot-afternoon case, where the tracking voltage sags and a string sized only for the cold limit can fall out of the tracker's range in the months it should be earning most. Where the resulting band is one or two modules wide, that is a signal to change the inverter or add module-level electronics, not a tolerance to work inside.
One consequence is specific to the approval route and easy to miss. An application is assessed against a declared model with a declared type-test record, so a substitution made later because the string count did not work is not a paperwork detail — it is a different machine with a different certificate, and depending on the operator it can mean going round the assessment again. Close the string design before the application is submitted, not before the modules are delivered.
This settles the cold end only — the most modules a string may hold before the hybrid's maximum input voltage is passed. The start voltage and the per-tracker current limit come off that model's datasheet by hand.
The maximum allowable DC voltage for the inverter or system.
The panel's rated open-circuit voltage at Standard Test Conditions.
Accounts for Voc increasing as ambient temperature drops. NEC Table 690.7(A) gives 1.12 for a record low of -1 to -5 C (31 to 23 F), and 1.16 by -11 to -15 C (13 to 5 F).
Maximum panels per string
13 panels
The correction factor depends on your site's record low ambient temperature and the panel manufacturer's temperature coefficient — use NEC Table 690.7(A) or the manufacturer's datasheet coefficient for your specific location, not an assumed value.
- Temperature-corrected Voc per panel
- 44.8 V
They open the calculator with your figures already in it
Solar PV String Sizing Calculator: 13 panels — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Estimated cost — your price
This site holds no price list for this material — local prices vary too much to publish honestly. Enter your supplier's price and the result is costed with it.
What this calculation does not cover
- This is the cold-weather upper voltage limit only. Nothing here checks the minimum string length: whether the string's operating voltage at the hottest expected cell temperature still sits above the inverter's MPPT window and start-up voltage. A string that passes this check can still be too short to make power on a hot afternoon.
- The calculation is voltage only. It does not look at short-circuit current, the module's maximum series fuse rating, the inverter's per-MPPT input current limit, conductor ampacity, or combiner fusing, so a string length that passes here has had only half of a string design checked.
- It applies whichever single voltage limit you type. It does not reconcile the inverter's maximum DC input against the module nameplate's own maximum system voltage rating or against the limit your code and occupancy impose - enter the lowest of the three, because the calculator cannot tell you which one governs. The Voc field also stops at 80 V, which covers crystalline silicon modules; thin-film modules run to several times that and will be clamped down to 80 V, returning a string far longer than the module can take.
- The calculator does no temperature arithmetic of its own. It never sees your site's record low ambient temperature or the module's Voc temperature coefficient - it multiplies by whatever factor you type, so a factor taken from the wrong row produces a confident, wrong answer. The factor field also stops at 1.25, the coldest row of NEC Table 690.7(A).
- This is a sizing check, not a design. It assumes every module in the string is identical and does not substitute for the inverter manufacturer's own string-sizing tool or for a design reviewed by the authority having jurisdiction.
The AC run is what decides whether the inverter stays online
Direct-current cable sizing on the roof is the familiar part: hot conduit, derated ampacity, a long run and a voltage drop you decide to accept rather than discover. The alternating-current run from the inverter to the board deserves the same treatment for a different reason. An inverter exporting into a supply raises the voltage at its own terminals, and the thinner and longer that run is, the higher the rise. Push it far enough and the unit trips on over-voltage on exactly the bright afternoons it should be earning, which presents to the customer as an intermittent fault and to you as a return visit with a clamp meter.
The protective settings that cause that trip are not yours to adjust. They are part of the type-test compliance under the applicable engineering recommendation or the inverter's certification to IEEE 1547 through UL 1741, and the region setting on an AS/NZS 4777.2 inverter has to match the one the local distribution business requires. The correct response to nuisance over-voltage tripping is a heavier AC conductor and a shorter route, not a loosened setting — and where the rise persists with the cable properly sized, it is evidence of a network voltage issue to report rather than to engineer around. Where a limitation scheme is being used, plan the CT position and its cable route in the same exercise, because it has to sit upstream of everything and that is often the most awkward run on the job.
Set the DC conductor by how much drop you will accept across that run rather than by a rule of thumb. It is a two-wire calculation, so a single-phase AC run to the board takes the same arithmetic at 230 V — a three-phase run does not.
The expected current carried by the PV circuit conductor.
The nominal DC system voltage of the PV circuit.
The one-way distance from the source (e.g. array/combiner box) to the load (e.g. inverter/battery).
The target maximum percentage of system voltage lost to conductor resistance.
The resistivity of the conductor material at its expected operating temperature.
Minimum conductor cross-sectional area
0.00844 in²
Convert the resulting minimum cross-sectional area to the next larger standard wire gauge (AWG) or metric conductor size — do not round down. Also verify the selected conductor's ampacity rating (NEC Table 310.16, with appropriate temperature/conduit fill derating) independently meets or exceeds the circuit current.
They open the calculator with your figures already in it
Solar Panel PV Wire Sizing Calculator: 0.0084 in² — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 0.00844 in² — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- Voltage drop is not ampacity, and none of the code side of the job is done here. The current is used exactly as typed, with no PV multiplier applied to it, no fuse or breaker sized, and no account taken of the temperature rating of the terminals the conductor lands in. A size that meets this drop target can still be thermally inadequate and non-compliant.
- Copper only. The resistivity field spans copper from cold to a conductor at its rated temperature; aluminium is about 1.6 times as resistive and does not fit inside that range, so an aluminium run has to be sized against the aluminium figure and comes out roughly two AWG sizes larger for the same drop.
- Nothing raises the resistivity for how hot the conductor actually gets. The figure is whatever you enter, so leaving the 20 °C (68 °F) table value in place understates the drop of a loaded circuit sitting in conduit on a sunlit roof by around a fifth against the 75 °C (167 °F) figure.
- Conductor metal only. MC4 connectors, fuse holders, breakers, combiner busbars and every termination in the run add resistance the formula never sees, so an assembled circuit drops more than the copper alone accounts for.
- This models a two-wire DC circuit carrying its full current over the whole run. It does not describe the inverter's AC output circuit, where reactance, power factor and the three-phase multiplier replace the factor of two, and it does not describe a home run that picks up further strings along its length.
Pricing the wait against pricing the cap
Both routes cost the customer money and the honest comparison puts them in the same units. The approval route costs a delay, and a delay has a season attached to it: an eight-week wait entered in March is most of a spring, while the same wait entered in September costs comparatively little because the months it eats were never going to generate much. The capped route costs the energy above the cap, which is a summer-weighted loss concentrated in a few hundred midday hours — and which a battery recovers a substantial part of, since that energy is going somewhere useful rather than nowhere.
Two things have to be true before any export figure enters the arithmetic. The tariff has to exist and be signed: what exported energy is worth is a matter of the supplier's offer, and in Great Britain the Smart Export Guarantee sets the framework, with its own conditions about installation certification under standards such as MCS MIS 3002 and about the meter being capable of half-hourly readings. And the meter has to be right. An export payment against a meter that cannot record export half-hourly is a conversation that ends badly some months after handover, when the scaffold is long gone.
Treat the payback figure as what it is. A simple payback is one division — total outlay over annual saving — with no financing, no degradation, no inverter replacement and no tariff movement in it. Its value here is comparative rather than absolute: run it twice, once for the capped system installed next month and once for the full system installed after approval, using the same assumptions in both, and read the gap rather than either number. If the gap is small, take the route that gets the scaffold down sooner.
Run the capped design and the approved design through the same simple division and compare the two answers, rather than treating either one as a forecast.
The total installed cost, after any rebates or tax credits already applied.
How much you expect to save on electricity bills per year.
Estimated payback period
10 years
This is a simple payback estimate that ignores financing costs, electricity rate changes over time, panel degradation, and maintenance — a full financial analysis from your installer will be more precise.
They open the calculator with your figures already in it
Solar Payback Period Calculator: 10 years — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
What this calculation does not cover
- There is no time value of money in this result. It counts nominal currency recovered, with no discount rate, no inflation applied to the savings, and no escalation of electricity rates over the period — a discounted break-even measured in today's money will land somewhere else.
- Financing is absent entirely: no interest rate, no loan term, no payment schedule. If you are borrowing, enter the sum of every payment you will actually make rather than the cash price, and read the answer as the point where cumulative savings match total outlay, not as a cash-flow date.
- The annual saving is held flat for the whole period. Panel output degrades year on year, soiling and new shading cut it further, and nothing is deducted for inverter replacement, cleaning, monitoring, insurance or repairs — there is no ongoing-cost input to put them in.
- Your tariff structure is not modelled. The calculation cannot separate output you consume as it is generated (worth your retail rate) from output you export (credited at whatever your utility pays), and it does not strip out fixed supply or connection charges, demand charges, or time-of-use pricing. It also assumes your current export or net-metering arrangement runs unchanged for the full payback period.
- This is a screening division, not a production estimate, a system design, or financial advice. Nothing here checks that your roof, orientation or array size can deliver the savings figure you entered, that you qualify for the incentives already netted out of the cost, or that the payback falls inside the equipment's service life, its warranty term, or your remaining time in the property — a result of 40 years is printed as plainly as one of 8.
The checks that belong before the purchase order, not after it
Everything above collapses into a short sequence that has to be finished while the order is still cancellable. It takes a morning. It is the difference between a job that closes on the day the scaffold comes down and one that stalls with modules in a customer's garage.
Do it in this order, because each answer constrains the next, and stop at the first one that changes the equipment list.
- Establish which procedure governs the address — the distribution network operator, state interconnection rules, or distribution business — and get their current threshold and published response timescale from them rather than from a previous job.
- Total the generation already on the supply, including any existing array and any inverter that can discharge a battery to the network, and add the proposed unit to it.
- Confirm the specific inverter model is on the relevant type-test or certification register for the route you intend to use; if it is not listed, the notify-after route is closed whatever its rating.
- Read the continuous AC output rating and the maximum battery discharge power off the datasheet as separate numbers, and declare the higher of the two.
- Settle the string arrangement against that model's tracker count, start voltage and per-tracker current limit, at both temperature extremes, before anything is declared or ordered.
- If a limitation scheme is in the design, ask the operator in writing how they will treat it and price the sensing hardware, its cable route and its commissioning verification into the quote.
- Check the fire separation, siting and aggregate capacity rules for the battery against the room it is going in, under the codes adopted locally.
- Confirm the export tariff and the metering arrangement exist before any export income appears in the customer's numbers.
What has to be true after the switch closes
If the installation went in under the notify-after route, the notification is not optional and it is not open-ended: it carries the type-test reference, the rating and the installation details, and it has to be filed inside the period the recommendation sets. An unnotified system is invisible to the operator, which means the next application on that supply — an EV charger, a heat pump, a second array — is assessed against a picture of the premises that is wrong, and the correction lands on whoever is standing there at the time.
The file that leaves site should let a stranger reconstruct the decision. Record the declared rating and which of the two numbers on the datasheet it came from, the region or grid-code setting the inverter was commissioned with, the string arrangement and the measured open-circuit voltages against the calculated ones, the export limit and the verification that the limitation scheme actually reduces output when tested, and the meter arrangement the export tariff depends on. In five years someone will want to add to this system, and the single thing that will make that easy or expensive is whether the capacity already declared to the network is written down where they can find it.
Settle these before the order is placed
Each line closes a question that becomes a restocking charge or a second application once the equipment has been bought.
- The governing procedure for this address — Named operator, current threshold, published response timescale — obtained from them, not carried over from the last job.
- Total existing generation on the supply — Every inverter already connected, including anything that can discharge a battery outward. The limit belongs to the premises.
- Both ratings from the inverter datasheet — Continuous AC output and maximum battery discharge power, read separately. The higher one is what gets declared.
- Type-test or certification listing for the exact model — Not the range, the model. An unlisted unit closes the notify-after route regardless of its rating.
- String arrangement inside the hybrid's own window — Tracker count, start voltage and per-tracker current, checked at the design low and design high temperatures.
- Export limitation scope, if used — CT position upstream of all sources, monitored comms, fail-safe behaviour, and the operator's written view of how they will treat it.
- Battery siting against the adopted fire code — Separation, room and aggregate capacity rules decided before the garage wall is committed to.
Opens the calculators above on one screen with the dimensions from this article already filled in. Quantities only — this site publishes no price list, because local prices vary too much to publish honestly.