F—01
Roof sizing from satellite and photo
Trace the roof plane from imagery or a site photo. Setbacks, obstructions and walkways are subtracted before modules are placed.
Design · Financing · Monitoring
Know the payback before the panels go up.
Amperfield turns a roof photo, an address and a utility bill into a bankable proposal — real irradiance data, honest ROI, and a live monitoring record that holds the installed system to what was sold.
We sell the financial model, not the panels. Every figure on this page that describes a system is labelled sample output from our engine — not a customer result.
Cumulative net cash position — 59.5 kWp rooftop
Sample model output
- Simple payback
- 5.5 YR
- Array sized to roof
- 59.5 kWp
- Modelled specific yield
- 1,842 kWh/kWp
- Project IRR, 25 yr
- 16.8 %
Sample site used consistently across every figure on this page. Amperfield has not yet published measured fleet data — see what we will publish
01 — The problem
Rooftop solar is sold on vague promises.
High electricity costs and an unstable grid have made self-generation an economic necessity here, not an environmental preference. That should make solar an easy decision. It isn't — because nothing in the sales process is verifiable.
01
Buyers cannot see a credible payback model before committing
A one-page quote with a headline "saves you 60%" is not a financial model. There is no irradiance source, no degradation curve, no tariff assumption, and no way to interrogate the arithmetic before signing.
02
Installers quote the same roof inconsistently
Two contractors walking the same building produce different array sizes, different yield claims and different payback periods — because each is estimating by hand, from different rules of thumb.
03
Once the panels are up, nobody checks
The promise made at the point of sale is never reconciled against what the system actually produces. Soiling, string faults and shading go unnoticed for months. That accountability gap is why buyers distrust the next quote.
| Line item | Typical quote | Amperfield |
|---|---|---|
| Irradiance source | Unstated | Cited dataset + year |
| Array layout | Panel count only | Roof plan, setbacks, keep-outs |
| Loss breakdown | None | Line-by-line to specific yield |
| Tariff assumption | Today's rate | Escalation band, stated |
| Degradation | Ignored | Modelled to year 25 |
| Post-install proof | None | Metered vs. modelled, monthly |
02 — The solution
A sizing and payback engine, then a witness.
Three inputs — a roof photo or satellite trace, an address, and twelve months of utility bills — produce a proposal a bank would accept. After commissioning, the same model becomes the benchmark the system is measured against.
- Geometry first. The roof is traced, not guessed: obstructions, setbacks, tilt and azimuth all resolve to a usable-area number before a single module is placed.
- Irradiance, not rules of thumb. Plane-of-array irradiance is derived for the specific tilt and azimuth, then reduced by a stated loss stack to specific yield.
- Cash, not kilowatt-hours. Yield meets the actual tariff structure, self-consumption profile and financing terms to produce cumulative cash position by year.
- The model survives commissioning. The proposal's monthly production curve becomes the expectation line on the owner's monitoring dashboard.
Roof sizing — array layout plan
Sample model output
03 — Capabilities
Seven things the platform does.
Built for the people who quote solar and the people who own it. Nothing here is a dashboard for its own sake — every module exists because somebody has to defend a number.
F—02
Irradiance-based yield modelling
Plane-of-array irradiance for the actual tilt and azimuth, reduced through a published loss stack to specific yield in kWh/kWp.
F—03
Bankable payback and ROI
Cumulative cash position, IRR, NPV and break-even year — each traceable back to the assumption that produced it.
F—04
Installer quoting and proposals
Branded, versioned proposals generated from the model. Every quote your team sends uses the same arithmetic.
F—05
Live monitoring and alerts
Inverter data compared against the modelled curve. Sustained deviation raises an alert with a probable cause attached.
F—06
Financing and lease scenarios
Compare cash purchase, term loan and lease side by side on one cash-flow chart, with the cost of capital stated.
F—07
Portfolio view for multi-site owners
Municipalities, cooperatives and estates see every site's performance ratio in one table, ranked by deviation from model.
WHO IT'S FOR
Installers, owners, municipalities
Solar installers and electrical contractors quote with it. Factories, hospitals, schools and commercial landlords own the monitoring record. Municipalities and cooperatives run shared generation across it.
04 — How it works
Roof to proposal in five steps.
Each step produces an artefact the next step consumes. Nothing is entered twice, and every assumption stays attached to the number it produced.
-
STEP 01
Capture the roof
Drop a pin or upload a site photo. Trace the roof plane, mark obstructions, set tilt and azimuth.
OutputRoof plane area, usable area after setbacks, orientation.
-
STEP 02
Lay out the array
Modules are packed into the usable area at the specified pitch and row spacing, cut around every keep-out zone.
OutputModule count, DC rating, string plan, inverter sizing ratio.
-
STEP 03
Model the yield
Monthly plane-of-array irradiance is converted to AC energy through a stated loss stack — soiling, temperature, mismatch, wiring, inverter.
OutputMonthly and annual specific yield, performance ratio.
-
STEP 04
Build the financial case
Twelve months of bills define the tariff and consumption profile. Yield meets tariff; financing terms meet capex.
OutputPayback year, IRR, NPV, financing comparison, exportable proposal.
-
STEP 05
Commission and watch
The inverter feed is connected and the modelled curve becomes the expectation line. Deviation beyond tolerance raises an alert.
OutputLive performance ratio, alerts, monthly metered-vs-modelled report.
Step 03 — irradiance to specific yield
Sample model output
Plane-of-array irradiance (kWh/m²/month)
Specific yield after losses (kWh/kWp/month)
| Roof plane | 788 m², flat concrete deck |
|---|---|
| Usable after setbacks | 268 m² |
| Modules | 96 × 620 W monofacial |
| DC rating | 59.5 kWp |
| Tilt / azimuth | 12° / 186° |
| Inverter | 50 kW, DC:AC 1.19 |
| Installed cost | $785 / kWp |
| Tariff, year 1 | $0.148 / kWh, +3%/yr |
| Self-consumption | 84% of production |
| Degradation | 0.5% / yr after year 1 |
Every proposal ships with this table attached. If a number moves, the chart above moves with it.
05 — Monitoring
The promise, kept under observation.
A proposal is a claim. Monitoring is what turns it into a record. The modelled curve stays on the chart for the life of the asset, and the platform argues with reality on the owner's behalf.
amperfield / portfolio / Site 004 — light industrial, 59.5 kWp
- Produced today
- 315 kWh
- Modelled expectation
- 362 kWh
- Performance ratio
- 0.87
- Open alerts
- 1
Metered AC output
Modelled expectation from the proposal
Flagged deviation window
ALERT · STRING B UNDERPERFORMANCE. Output held 24–28% below model for four consecutive hours around solar noon while irradiance stayed nominal. Probable cause: partial shading or soiling on strings B3–B5. Assigned to the installer of record; owner notified.
Illustrative dashboard populated with demo data for the sample site. Interface shown is representative of the monitoring product.
06 — Pricing
Priced per seat, then per site.
Installers pay for the engine by the seat. Owners pay for the monitoring record by the site. Nobody pays a percentage of the system cost, because that would give us a reason to oversize the array.
Installer
Solar installers, electrical contractors, EPC teams
$99
PER SEAT / MONTH
- Unlimited roof models and proposals
- Branded, versioned proposal export
- Financing and lease scenario comparison
- Shared assumption library across the team
- Hand-off into monitoring at commissioning
Site monitoring
Building owners — factories, hospitals, schools, landlords
$8
PER SITE / MONTH
- Live metered-vs-modelled production
- Underperformance alerts with probable cause
- Monthly performance-ratio report
- The original proposal kept on file as the benchmark
- Installer of record notified on every alert
Portfolio
Municipalities, cooperatives, multi-site estates
On request
ANNUAL, BY FLEET SIZE
- Fleet table ranked by deviation from model
- Shared-generation and allocation views
- Role-based access for finance and facilities
- Data export and API access
- Named onboarding contact
Is there a charge per kilowatt installed?
No. Seats and sites are the only two meters. Pricing that scales with system size would reward us for recommending a bigger array than the roof or the tariff justifies.
Who pays for monitoring — the installer or the owner?
The asset owner, at $8 per site per month. Installers frequently bundle the first year into the system price; the subscription stays in the owner's name either way, so the record survives a change of contractor.
What happens if the system underperforms the proposal?
The deviation is logged, the probable cause is attached, and both the owner and the installer of record are notified. Amperfield does not underwrite performance — we make the gap visible and attributable, which is the part that is currently missing.
Can we bring our own irradiance dataset?
Yes, on the Portfolio plan. The dataset and vintage used in each model are printed on the proposal regardless of the source, so a lender can check them.
Pilot sites
We are onboarding a first cohort of installers and owner-operated sites. Pilot participants get the engine at no cost during the pilot window in exchange for permission to publish anonymised metered-vs-modelled results.
That published data is what will fill the empty figures below — not a marketing estimate.
07 — Credibility
We will not quote a number we cannot show you.
A platform whose whole argument is "verify the claim" cannot open with unverifiable claims of its own. Below are the three figures we intend to publish, left deliberately empty until there is metered data behind them.
Awaiting data
Installed capacity modelled
Awaiting data
Mean forecast accuracy
Awaiting data
Installer partners live
How each figure will be produced
- Capacity modelled — counted at commissioning from the installer's sign-off, not from proposals sent.
- Forecast accuracy — twelve full months of metered generation divided by the modelled figure in the original signed proposal, reported as a distribution rather than a single flattering average.
- Partner logos — displayed only for firms with a signed agreement who have approved the use of their mark. No stock logos, no "as used by" implications.
- Method note — published alongside the figures, including sites excluded and why.
Where the engineering comes from
Amperfield is built in a market where grid instability and high tariffs have already made self-generation an economic decision rather than an environmental one — which is exactly the market that punishes an over-optimistic model hardest.
An-Najah National University's renewable energy engineering programme sits next door to that problem, and is our direct pipeline for both engineering talent and instrumented pilot rooftops.
What we will hand a lender
| Irradiance source + vintage | Printed |
|---|---|
| Full loss stack | Printed |
| Tariff & escalation basis | Printed |
| Sensitivity band | P50 / P90 |
| Post-install reconciliation | Monthly |
Model a roof this week. Argue with the numbers, not the salesman.
Send us an address and a recent utility bill and we will run the model with you — installers on a demo call, owners on their own building.