# Sezibwa rooftop solar: panel layout and installer brief

**For:** the Green Volta installers and the NFE site team. **Owner:** Aaron Tushabe. **Last reviewed:** 21 Oct 2026. **Status:** proposed. The layout was worked out from satellite imagery, so do the site checks below before drilling.

## The plan

- **What:** 8 × Golden Sun GS600N-144M (600 W, 2278 × 1134 mm, 30 kg each), 4.8 kWp in total, from the [Muden quote](https://bookstack.nearlyfreeenergy.com/link/156).
- **Where:** the long north-west-facing face of the house at about **0.36236, 32.53602** (Nansana). The pin 0.36245, 32.53592 lands on the neighbour's roof about 15 m to the north-west.
- **How:** one row in portrait (long side up the slope), split into **two groups of 4 with a 1.5 m gap** around a vent or flue. The top edge sits about 0.4 m below the ridge. Flush-mounted at the roof pitch.
- **Wiring:** all 8 panels in **one series string into PV1** of the SRNE ASP48120SH3. PV2 stays free.

![Satellite layout: 8 panels on the NW roof face](https://bookstack.nearlyfreeenergy.com/attachments/73?open=true)

![Rails, clamps and string wiring drawing](https://bookstack.nearlyfreeenergy.com/attachments/74?open=true)

## Why this roof face

At 0.36° N the sun is close to overhead all year, so the direction a face points makes little difference:

<table id="bkmrk-roof-face-kwh-per-ye"><thead><tr><th>Roof face</th><th>kWh per year</th><th>Worst month (kWh/day)</th></tr></thead><tbody><tr><td>**North-west (chosen)**</td><td>**6,780**</td><td>**16.8 (Nov)**</td></tr><tr><td>South-west</td><td>6,770</td><td></td></tr><tr><td>South-east</td><td>6,660</td><td>15.7 (Jun)</td></tr><tr><td>North-east</td><td>6,640</td><td></td></tr></tbody></table>

PVGIS (SARAH3 data, 2005 to 2023), 20° pitch, 14% system losses, no tree shade.

The trees decide it. Tall trees stand against the north, north-east and east sides of the house, and big trees to the south and south-west. The north-west face looks away from most of them. All 8 panels are on one string, so shade on one panel pulls down all 8. The north-west face also gives the steadiest output through the year.

## Site checks before drilling

![Roof words used in the site checks](https://bookstack.nearlyfreeenergy.com/attachments/77?open=true)

**Words used in these checks**

<table id="bkmrk-word-what-it-means-h"><thead><tr><th>Word</th><th>What it means</th><th>How to check it on site</th></tr></thead><tbody><tr><td>Ridge</td><td>The top line of the roof, where the two roof faces meet.</td><td></td></tr><tr><td>Eave</td><td>The bottom edge of the roof, where the gutter is.</td><td></td></tr><tr><td>Roof face</td><td>One sloping side of the roof. Ours is the side facing the alley (north-west).</td><td></td></tr><tr><td>Pitch</td><td>How steep the roof is, in degrees.</td><td>Lay a phone level app on the sheet, pointing up the slope. Or use a 1 m level and the table in picture 4.</td></tr><tr><td>Face depth</td><td>The distance from the eave to the ridge, measured along the sheet (following the slope, not on the ground).</td><td>Run a tape up one rib from the gutter edge to the ridge. It must be at least 3.0 m.</td></tr><tr><td>Purlin</td><td>Wooden strips under the iron sheets, running the same way as the ridge. The sheets are nailed to them, and the L-feet must screw into them.</td><td>Each straight row of roofing nails across the roof is a purlin. Measure the gap between rows (picture 2).</td></tr><tr><td>Rafter</td><td>The sloping timbers under the purlins, running from eave to ridge.</td><td>Look inside the roof space if you can, and check for rot or termite damage.</td></tr><tr><td>Crest and valley</td><td>The high ribs and the low channels of the corrugated sheet.</td><td>Drill only through a crest. Rain runs down the valleys (picture 3).</td></tr></tbody></table>

- [ ]  **Vent or flue.** Satellite shows two dark marks about halfway along the face, roughly 8.7 m from the north-east end. If nothing is there, close the gap between the groups to about 0.3 m (keep the two rail pairs separate) and centre all 8 panels on the face.
- [ ]  **Pitch.** Measure it and write it on this page. Anything from 10° to 25° is fine.
- [ ]  **Face depth.** Eave to ridge along the slope must be at least 3.0 m (2.28 m panel, 0.4 m at the ridge, margin at the gutter). The satellite estimate is about 4.7 m.
- [ ]  **Purlins.** Find their positions and spacing. Each rail has to sit on a purlin line.
- [ ]  **Shade.**
    - Stand at the array position at about 8 am, noon and 4 pm, or use a sun-path app.
    - Trim the trees on the north, north-east and east sides and on the south and south-west sides to below ridge height.
    - Check that the neighbour's roof (north-west, across the alley) is not higher than our ridge.
- [ ]  **Roof condition.** Replace rusty or loose sheets and rotten timber under the array first.
- [ ]  **Cable route.** Measure from the drop point to the inverter so the 6 mm² cable can be ordered.

## Mounting

The hardware on the Muden quote matches this layout exactly: two 5.1 m rails per group of 4 portrait panels.

<table id="bkmrk-item-per-group-total"><thead><tr><th>Item</th><th>Per group</th><th>Total</th><th>On the quote</th></tr></thead><tbody><tr><td>Rail, 5100 mm</td><td>2</td><td>4</td><td>4</td></tr><tr><td>End clamp</td><td>4</td><td>8</td><td>8</td></tr><tr><td>Inter clamp</td><td>6</td><td>12</td><td>12</td></tr><tr><td>L-foot</td><td>6</td><td>12</td><td>12</td></tr></tbody></table>

- **Rail position:** rails run parallel to the ridge, about a quarter of the panel length (about 570 mm) in from the top and bottom edges, so about 1.14 m apart. 
    - The panel datasheet does not give the clamping zone, so check Golden Sun's installation manual.
    - Move the rails within that zone so they land on purlins.
- **L-feet:** 3 per 5.1 m rail gives spans of about 2.25 m. Check the rail's span table. 4 per rail (about 1.5 m spans) is safer: that's 4 more L-feet, about 48,000 UGX at the quoted price.
- **Fixing:** fix the L-feet into the purlins, not just the iron sheet. Go through the crest of the corrugation, with EPDM washers and sealant.
- **Airflow:** leave about 100 mm under the panels. Cooler panels produce more.
- **Load:** about 240 kg of panels plus rails, spread over about 21 m².

## Wiring: one string of 8 into PV1

<table id="bkmrk-string-of-8-srne-asp"><thead><tr><th></th><th>String of 8</th><th>SRNE ASP48120SH3, per PV input</th><th>OK</th></tr></thead><tbody><tr><td>Open-circuit voltage</td><td>422 V at 25 °C, about 440 V on a cold morning</td><td>800 V max</td><td>✓</td></tr><tr><td>Max-power voltage</td><td>359 V, about 310 V on a hot afternoon</td><td>200 to 650 V tracking range</td><td>✓</td></tr><tr><td>Current</td><td>13.4 A at max power, 14.0 A short-circuit</td><td>22 A max</td><td>✓</td></tr><tr><td>Power</td><td>4.8 kW</td><td>9 kW max</td><td>✓</td></tr></tbody></table>

The hot and cold figures assume typical n-type temperature coefficients (open-circuit voltage −0.27 %/°C, max-power voltage −0.30 %/°C), because the panel datasheet does not give them.

- **Two strings of 4 will not work.** They would run at about 155 to 180 V, below the inverter's 200 V minimum. The 4 + 4 split is physical only.
- **Panel to panel:** panels 1 to 4 connect directly with their 300 mm leads. A short jumper crosses the gap to panel 5, then panels 5 to 8 connect directly.
- **Return lead:** the string's + leaves panel 1 and its − leaves panel 8, so one lead runs back along the row (about 11 m). 
    - Tie it under the panels along the lower rail.
    - Drop both leads at whichever end is nearer the inverter.
    - Keep + and − together all the way to the inverter. A small loop picks up less lightning surge.
- **Cable:** 6 mm² UV-rated PV cable (the 4 mm² was out of stock). Voltage loss over a 30 m run is under 1%.
- **MC4 connectors:** crimp them with the proper tool and mate only with compatible connectors.
- **At the inverter:**
    - The 32 A 4-pole DC isolator, wired to its diagram for 500 V+ DC (usually two poles in series on each polarity).
    - Then the 1000 V DC surge protector, with a short lead to earth.
    - The inverter is IP20, so it must be indoors.
- **Earthing:** bond every rail and panel frame to the site earth with at least 6 mm² copper.

## Not on the Muden quote: bring these

- 6 mm² PV cable for the return lead (about 11 m), the jumper and the run to the inverter. Measure first; 30 to 40 m was budgeted.
- Earthing conductor, lugs or bonding clips, and the connection to the earth rod.
- Roof sealant, EPDM washers, UV-rated cable ties or clips.
- Optional: 4 extra L-feet.

## Commissioning

- [ ]  **Before closing the isolator:** check polarity and measure the string's open-circuit voltage. 
    - Expect about 370 to 425 V in daylight. It reads lower when the panels are hot.
    - About 50 V low means a panel is left out of the string.
    - About 100 V low means one panel is connected backwards.
- [ ]  **Insulation test:** if a tester is available, test + and − to earth.
- [ ]  **Close the isolator.** Check PV1 voltage, current and power on the inverter display and in OpenEMS.
- [ ]  **Photos:** the panel labels and serial numbers (for the warranty), the finished array, the isolator and the surge protector.
- [ ]  **Update this page** with the as-built positions, the measured pitch and the voltage reading.

**Commissioning log**

<table id="bkmrk-date-and-time-%28eat%29-"><thead><tr><th>Date and time (EAT)</th><th>Check</th><th>Result</th></tr></thead><tbody><tr><td>26 Oct 2026, about 16:45</td><td>String open-circuit voltage at the bottom of the 6 mm² drop cable, isolator open (INGCO DM360011 meter, 600 V DC range)</td><td>379 V, positive on the red probe. All 8 panels are in series, none reversed, and the drop cable is continuous.</td></tr></tbody></table>

![String open-circuit voltage, 26 Oct 2026: 379 V](https://bookstack.nearlyfreeenergy.com/attachments/79?open=true)

## Expected output

Average kWh per day by month, unshaded, north-west face at 20° (PVGIS):

<table id="bkmrk-jan-feb-mar-apr-may-"><thead><tr><th>Jan</th><th>Feb</th><th>Mar</th><th>Apr</th><th>May</th><th>Jun</th><th>Jul</th><th>Aug</th><th>Sep</th><th>Oct</th><th>Nov</th><th>Dec</th></tr></thead><tbody><tr><td>18.6</td><td>20.0</td><td>20.4</td><td>19.7</td><td>18.7</td><td>18.5</td><td>19.1</td><td>18.2</td><td>18.2</td><td>17.7</td><td>16.8</td><td>17.2</td></tr></tbody></table>

That is about 6,780 kWh a year. It agrees with the 5.2 peak-sun-hour assumption in the [sizing re-run](https://bookstack.nearlyfreeenergy.com/link/149) (about 19.5 kWh/day). Single days swing a lot with the weather. A month that comes in well below these figures points to shade or a wiring fault.

## Assumptions

- Roof dimensions come from Google and Esri satellite imagery, scaled from the Google Maps scale bar, and are good to about ±0.5 m. The pitch is assumed to be 20°.
- The PVGIS figures include no tree shading.
- Sources: Golden Sun GS600N-144M datasheet; SRNE ASP 8 to 12 kW three-phase datasheet V1.0.

## Related pages

- [Sezibwa solar purchase: final Muden quote (19 Oct 2026)](https://bookstack.nearlyfreeenergy.com/link/156)
- [Solar Sizing Re-run, Sept 2026 (600 Wp panels)](https://bookstack.nearlyfreeenergy.com/link/149)
- [SRNE ASP48120SH3 Safe Settings Write: design, registers, and commissioning](https://bookstack.nearlyfreeenergy.com/link/128)
- [Microgrid Cabinet Line Diagram](https://bookstack.nearlyfreeenergy.com/link/124)

## As built, 26 Oct 2026

- All 8 panels in **one row** on two continuous rail lines; the unused vent was removed (landlord agreed), so no gap.
- Clamps: the one-row layout needs 14 inter + 4 end clamps; 12 + 8 were supplied, so the 4 spare end clamps were used back to back at one joint.
- 6 mm² PV cable (negative 25 m, positive 15 m; ~0.4% drop). FMPV32 4-pole DC isolator, manual layout 4B; 1000 V DC surge protector across +/−; inverter **PV1** (PV2 unused).
- Readings: 379 V open circuit at 16:45 EAT; inverter PV1 152 V at sunset (18:39).
- Open: **earthing** (no earth point on site) for the surge protector and rails; possible **tree shading** at one end of the array.
- Operating plan and commissioning log: [Sezibwa Solar + Battery Operating Plan](https://bookstack.nearlyfreeenergy.com/books/product/page/sezibwa-solar-battery-operating-plan).