This page is for solar design engineers, site and project engineers, O&M engineers and solar consultants. Most solar interviews mix a few questions on why you chose the field, a solid block of PV knowledge such as modules, inverters, string sizing, shading, earthing and commissioning tests, and then stories and scenarios from real sites: a roof that could not take the load, an inverter that trips every morning, a customer holding a cheaper quote. Each question shows what the interviewer is checking, a shape for your answer and a short answer you could say out loud. Grid rules differ by country and utility, so learn the local ones before the day.
Search all questions by round, difficulty and level, or save the ones you want to practise.
Start: what first pulled you in, a course, a project or a job.
What you have done: the hands-on work so far, in one or two lines.
Direction: the part of solar you want to go deeper in and why.
"I studied electrical engineering, and my final-year project was a small rooftop system for our hostel, which we sized, simulated and then helped a local installer put up. Seeing the meter run backwards on the first sunny day hooked me. After graduating I joined an installer as a site engineer, so I've spent two years on roofs doing layouts, cabling, earthing and commissioning. What I've learned is that most problems I fix on site were decided at the design stage, like a string that was too long for a cold morning or an inverter placed in full sun. So I want to grow into design, where I can use what I've seen on site to get it right on paper first."
Saying solar is simply the future or a booming sector, with nothing personal or hands-on behind it.
Their segment: residential, commercial and industrial, or utility-scale, and roughly what size.
What that means: the typical challenges in that segment.
Your fit: the experience or interest that matches it.
"From your website and project photos, most of your work is commercial rooftops, factories and warehouses, somewhere in the range of a few hundred kilowatts. That tells me the hard parts are structural checks on old sheet roofs, getting the work done while the building stays in use, and dealing with the utility for export approvals. It also means the customer is a business that cares about payback and downtime, not just the idea of going green. I've done two warehouse projects at my current company, including one where we had to redesign the layout around skylights and roof vents, so I know the kind of problems that come up. I'd like to do more of that work at a larger scale."
Talking only about solar in general, with no idea whether the company builds home systems or large plants.
What you bring: training, internships or projects with real hardware.
Respect for risk: heights and DC that cannot simply be switched off.
How you learn: shadowing, checklists and asking before acting.
"I won't claim I know everything a site needs, but I think I'm ready to be useful safely. During my internship I spent six weeks with an installation crew, so I've worn a harness, made up and checked DC connectors, and helped with string voltage checks under supervision. The thing that stayed with me is that a module produces voltage whenever light hits it, so you can't just switch the array off like a house circuit. That changed how I treat every connector. My plan for the first months is to follow the site's method statements exactly, ask before I do anything I haven't done before, and learn the commissioning checklist from the senior engineer until I can run it on my own."
Treating the roof and the DC side as low risk, or saying you'll figure things out on your own as you go.
Sources: datasheets, manufacturer training, standards, utility notices.
Practice: trying new tools or kit on real projects.
Sharing: passing it on to the team.
"I split it into three things. For products, I read the datasheets and installation manuals of new modules and inverters properly, especially the parts about warranties and installation limits, and I do manufacturer training when it's offered. For rules, I keep an eye on notices from the local utility and regulator, because a change in export limits or metering rules changes how we size systems overnight. I also go back to the relevant standards when a question comes up on site instead of relying on memory. And for skills, I try new design or monitoring tools on a real project rather than just reading about them. When I learn something useful, I share a short note with the team, because a rule change missed by one person becomes a problem for everyone."
Saying you keep up by watching videos, with no mention of datasheets, standards or local rules.
The effect: light frees charge carriers in a silicon PN junction; the junction's field separates them into a DC current.
Three points: Voc with no load, Isc with the terminals shorted, and the maximum power point where voltage times current peaks.
What moves them: irradiance mostly moves current; temperature mostly moves voltage.
"A solar cell is a silicon PN junction. When light with enough energy hits it, it frees electrons and leaves holes behind, and the electric field at the junction pushes them apart, so if you connect a load, a DC current flows. Cells are wired in series inside a module to build up a useful voltage. On the IV curve, open-circuit voltage is the voltage with nothing connected, short-circuit current is the current with the terminals shorted, and the maximum power point is the spot on the curve where voltage times current is highest. The inverter's MPPT keeps hunting for that point. Current rises roughly in step with irradiance, while voltage falls as the cell heats up, which is why modules make less power on hot days."
Saying a module produces its nameplate watts all day, or mixing up which of voltage and current depends on temperature.
Mono versus poly: single crystal versus many grains, and what that does to efficiency and space.
Half-cut: lower current per cell section, so lower resistive losses and better shade behaviour.
Bifacial: rear-side gain that depends on the ground, the height and the row spacing.
"Monocrystalline cells are cut from a single silicon crystal, so they're more efficient and give you more watts per square metre, which matters on a small roof. Polycrystalline cells are cast from many crystals, they're a bit less efficient, and mono has largely taken over the market, with newer cell designs like TOPCon built on it. Half-cut modules split each cell in two, so each half carries half the current and the resistive losses drop, and the module is wired so that shade on one half hurts less. Bifacial modules collect light on the back as well. The gain is real on a ground mount with light-coloured ground, decent height and wide rows, but on a flush-mounted roof with little light reaching the back, it's small. So I pick based on the site, not the label."
Claiming bifacial always adds a big fixed gain regardless of how and where it is mounted.
Grid-tied: no battery, exports surplus, shuts down in an outage.
Off-grid: batteries sized for days without sun, often with a generator backup.
Hybrid: grid plus battery, can back up chosen loads during an outage.
Fit: reliable grid, no grid, or an unreliable grid.
"A grid-tied system has no battery. It feeds your loads and sends any surplus to the grid, and when the grid fails it shuts itself down for the safety of line workers, so there's no backup. It's the cheapest option and right where the grid is reliable and export is credited. An off-grid system has no grid at all, so batteries and the inverter are sized for the load and for a few cloudy days, and there's often a generator as backup. That suits remote sites. A hybrid system is connected to the grid but also has a battery, so it can keep essential loads running during a power cut and store daytime solar for the evening. It fits where the grid is unreliable or export pays poorly."
Saying a standard grid-tied system keeps the lights on during a power cut.
String: simple and cost-effective for clean, unshaded arrays.
Module-level: micro-inverters or optimizers for complex, shaded roofs or where rules need module-level shutdown.
Central: large plants where fewer, bigger units make sense.
Trade-off: cost, failure points, access for repair.
"For a clean roof with one or two orientations and no real shade, I'd use a string inverter with enough MPPT inputs, because it's cheaper, simple and easy to service at ground level. If the roof is broken up into many small faces, or there's shade from a tank or chimney, micro-inverters or optimizers make sense, because each module is tracked on its own and one shaded module doesn't drag the string down. They also help where the local code asks for module-level rapid shutdown. The downside is many more electronic parts on the roof to fail. Central inverters are for large ground plants, where you have big uniform blocks and want fewer units to maintain. On commercial rooftops, several string inverters are usually the sweet spot."
Recommending micro-inverters or central inverters for every job without asking about shading, size or service access.
Energy need: total yearly units, and the seasonal and daytime pattern.
Yield: divide by the local specific yield per kilowatt peak.
Reality checks: roof area, connection and export limits, metering rules, future loads.
"First I add up the yearly units and look at the pattern, whether summer is heavy because of cooling, and how much is used in the daytime. Say it's 6,000 units a year. Then I need the local specific yield, how many units one kilowatt peak makes in a year at that tilt and orientation, from a simulation tool or solar resource data. If it's about 1,500, that points to roughly 4 kilowatts peak. With modules of around 450 watts, that's nine modules. Then I check it fits the usable roof after setbacks and shade, that it's within the connection and export limits the utility allows, and how the local metering rule treats surplus. Finally I ask about plans like an electric car or a heat pump, which might justify going bigger."
import math
annual_kwh = 6000 # from twelve months of bills
specific_yield = 1500 # kWh per kWp per year at this site
module_kw = 0.45
system_kwp = annual_kwh / specific_yield # 4.0 kWp
modules = math.ceil(system_kwp / module_kw) # 9 modules
print(system_kwp, modules, modules * module_kw) # 4.0 9 4.05Picking a size from the roof area or the budget alone, without looking at how much energy the home actually uses.
Maximum: Voc corrected to the coldest expected temperature must stay under the inverter's and the module's maximum voltage.
Minimum: Vmp at the hottest cell temperature must stay above the MPPT minimum.
Current: string current within the MPPT input limit.
"The voltage of a module goes up when it's cold and down when it's hot, so I check both ends. For the maximum, I take the datasheet Voc and correct it to the lowest ambient temperature at the site, because at sunrise on a cold day the cells sit near air temperature with almost no load. That cold Voc times the number of modules must stay below the inverter's maximum DC input and the module's maximum system voltage. For the minimum, I correct Vmp to a hot cell temperature, which can be far above air temperature, and make sure the string still sits above the MPPT's lower limit, ideally inside its full-power range. Then I check the string current against the MPPT input current rating."
import math
voc, vmp = 49.5, 41.8 # datasheet values at STC, volts
beta_voc, beta_vmp = -0.0028, -0.0034 # per degree C, as a fraction; no Vmp figure? use Pmax
t_min, t_cell_hot = -5, 70 # coldest ambient, hottest cell temperature
inv_max_v, mppt_min_v = 1000, 200
voc_cold = voc * (1 + beta_voc * (t_min - 25))
vmp_hot = vmp * (1 + beta_vmp * (t_cell_hot - 25))
print(math.floor(inv_max_v / voc_cold)) # 18 modules at most
print(math.ceil(mppt_min_v / vmp_hot)) # 6 modules at leastSizing strings on the datasheet Voc at 25 degrees with no temperature correction.
Definition: array kilowatts peak divided by inverter AC kilowatts.
Why oversize: real conditions rarely reach standard test conditions, so the inverter runs nearer its rating for more hours.
Cost: clipping at peak times, which you model and accept only when it pays.
"The DC to AC ratio is the array's peak DC rating divided by the inverter's AC output rating. Modules rarely deliver their nameplate, because heat, dust, cable losses and lower sunlight pull them down, so an array matched one to one would leave the inverter underused most of the day. If I put a bit more DC on it, the inverter works closer to its rating in the mornings, evenings and on cloudy days, and the energy per unit of inverter cost goes up. The price is clipping, where on bright, cool middays the array could give more than the inverter can pass, so the top gets cut off. Ratios a bit above one are common. I model the clipping loss and check the inverter's maximum DC input before deciding."
Saying the array must never be larger than the inverter, or oversizing without checking the inverter's DC input limits.
Net metering: exports offset imports, usually at a similar rate.
Gross metering: all generation sold at a set tariff, all use bought separately.
Net billing: exports credited at a lower rate, so self-use matters more.
Design effect: how big to go and whether a battery pays.
"Under net metering, a two-way meter records what you import and export, and your exports offset your imports over the billing period, often close to the retail rate. That lets you size to cover your yearly use, because daytime surplus effectively pays for night use. Under gross metering, everything the system makes is sold to the utility at a set tariff and you buy all your consumption as normal, so the design is about maximum yield. Under net billing, exports are credited at a lower rate than what you pay for imports, so every unit you use yourself is worth more than one you export. Then I size closer to the daytime load and look harder at a battery. The rules differ by country and by utility, so I always check the current local scheme first."
Promising savings based on exporting everything at the retail rate without checking the local metering rules.
The mistake: what went wrong and your part in it.
Impact: what it cost in yield, time or rework.
Fix: how you put it right and who you told.
Change: the check or habit you added so it doesn't repeat.
"On one of my first commercial designs I put the string inverters on an outside wall that took the full afternoon sun, because it kept the cable runs short. In the first summer the monitoring showed them derating most afternoons, right when the array was producing the most. Their internal temperatures were at the limit, and I realised I'd made that call on the drawing without thinking about heat. It was my design, so I told my manager straight away and worked out the lost energy honestly. We fitted a simple shade canopy and fixed the airflow clearance to match the manual, and the afternoon derating stopped. Since then, inverter location, sun, ventilation and clearance are a line on my design checklist, and I check them on the site visit, not from photos."
Choosing a mistake that was really someone else's fault, or one with no technical detail and nothing changed afterwards.
Rooftop checks: roof age, structural capacity, sheet type, orientation, access.
Ground checks: land ownership and future use, soil, drainage, fencing, cable run.
Yield and O&M: optimal tilt and easier cleaning on the ground.
Advice: cost and risk side by side, often a mix.
"I'd start by asking what the plot is for, because if they plan to expand the factory in five years, covering it with panels is a bad idea. For the roof, I'd check its age and remaining life, since re-roofing under an array means taking it all off, and get a structural check of the purlins and sheets. For the ground, I'd look at ownership, soil for the foundations, drainage, fencing and how far the cable has to run to the connection point. The ground mount can use the best tilt and even bifacial modules, and it's easier to clean and maintain, but it adds civil work and land use. Often the best answer is the roof first if it's sound, then ground for the balance."
Recommending the roof without asking about its age and strength, or the plot without asking about the owner's plans for it.
Series limit: cells in series share one current, so a shaded cell holds the rest back.
Hot spot: a shaded cell can be driven into reverse and heat up.
Bypass diodes: let current flow around a shaded group of cells.
Design response: layout, separate MPPTs, module-level electronics.
"In a string everything is in series, so every cell has to carry the same current. If one cell is shaded, it can't make that current, and without protection it gets forced into reverse bias and starts dissipating power as heat, which is how you get hot spots and damaged cells. Bypass diodes, usually three per module, each across a group of cells, switch on when their group is shaded, so the current flows around it. You lose that section's output, but not the whole string. The catch is that the IV curve now has more than one peak, and an inverter that doesn't scan the full curve can sit on the wrong one. So I avoid shaded spots, keep shaded areas on their own MPPT, or use module-level electronics."
Saying shading one module only reduces output by that one module's share, with no idea of the series effect or hot spots.
Situation: the site and the plan before you visited.
What you found: the specific thing, structural, shade or electrical.
Change: what you redesigned and why.
Result: what that saved or avoided.
"We had a school building quoted from satellite images for about 50 kilowatts on a flat concrete roof. When I went up, I found two things the photos didn't show. There was a row of water tanks on a raised platform that would shade a big part of the roof in the winter afternoons, and the old parapet had cracks, so I didn't want to anchor anything near it. I took shade readings and photos and did a quick sun-path check. We moved the array away from the tanks, split it across two inverter inputs so the partly shaded rows didn't pull the others down, and used a ballasted structure. The system came out a bit smaller, but the yield estimate held up in the first year."
A survey story where nothing was measured and the design was simply copied from the quote.
Show the effect: a shade analysis with the expected yield from that face.
Options: other faces, fewer modules, separate MPPT, module-level electronics.
Let them decide: in writing, with a realistic estimate.
"I'd first find out why they want that face. Often it's the one they see from the street, or it looks like the biggest. Then I'd run a shade analysis and show them in plain terms what those panels would make compared with the same panels on another face. If they still want that roof, I'd design so the shaded panels can't drag the rest down, with their own MPPT input or with optimizers or micro-inverters, and I'd give an estimate that includes the shade loss. I'd also mention that trimming the tree is a conversation with the neighbour, not something we can promise. The main thing is they decide knowing the real numbers, and it's all written into the proposal."
Installing on the shaded face with the standard yield estimate, or refusing without explaining any options.
Earthing: module frames, structures and enclosures bonded to earth, tested for resistance.
DC side: string fuses where parallel strings need them, DC isolator, DC surge protection.
AC side: breaker, isolator, residual current device as the inverter maker specifies, AC surge protection.
Lightning: a protection system where the risk assessment calls for it.
"All exposed metal, the module frames, the mounting structure and the inverter body, gets bonded together and taken to earth, so a fault can't leave a frame live, and I test the earth resistance against the local requirement. On the DC side, I fit string fuses when enough strings are in parallel that the reverse current from the others could exceed what a module can take, which with typical module ratings starts at about three strings in parallel. There's a DC isolator rated for DC at the inverter, and DC surge protection devices rated for the array voltage. On the AC side there's a breaker, an AC isolator, a residual current device of whatever type the inverter maker specifies, and AC surge protection. Whether I add a lightning protection system depends on the risk assessment, and I follow the local wiring code, such as IEC 60364-7-712 or the NEC."
Treating the DC side like household AC, such as fitting AC-rated switches or skipping surge protection.
Read the pattern: moisture makes a small insulation weakness show up.
Data first: the inverter's logged insulation readings.
Isolate: test each string's insulation resistance while it is still damp.
Find and fix: connectors, pinched cable, junction boxes; retest after.
"That pattern usually means there's a weak spot in the DC insulation somewhere that only leaks enough to earth when it's wet. The inverter checks the array's insulation before connecting, sees it's too low, and waits until the sun dries things out. I'd pull the inverter's logs to see the readings, then go on site early, while it's still damp. With the strings isolated and following the safe procedure, I'd test the insulation resistance of each string to earth to find the one that's low. Then I'd split that string down and look at the usual suspects: a connector not fully mated or with water inside, a cable pinched under a rail or clamp, rodent damage, a cracked junction box or backsheet. After the repair, I'd retest on another damp morning."
Suggesting the threshold be lowered or the setting disabled so the inverter starts.
Visual and earthing: inspection, protective earth continuity, labels.
DC tests: polarity, string Voc, string current, insulation resistance.
System tests: inverter start-up, anti-islanding, monitoring online.
Records: results and documents handed over.
"I follow a standard sequence, such as the one in IEC 62446-1. First a visual inspection: fixings, cable routing, connectors fully mated, labels in place. Then continuity of the protective earth and bonding. On the DC side I check the polarity of every string, measure open-circuit voltage and compare strings with the same module count, because a low one usually means a wrong module count, a shorted bypass diode or a damaged module. Then string current, and insulation resistance of the DC circuits to earth. After that I start the inverter, check it syncs, and test anti-islanding by opening the AC breaker and confirming it stops exporting. Finally I confirm monitoring is online, and hand over the test sheets, the drawings and the datasheets."
Describing commissioning as switching the inverter on and seeing that power appears on the screen.
Formula: actual AC energy divided by what the DC capacity would make under the measured irradiation at standard conditions.
Inputs: metered energy, installed kilowatts peak, plane-of-array irradiation.
Losses: temperature, soiling, shading, mismatch, cables, inverter, downtime.
"Performance ratio compares what the plant actually delivered with what it would have delivered if it turned all the sunlight on its modules into energy at its rated efficiency. I take the AC energy from the meter for the period and divide it by the installed kilowatts peak times the plane-of-array irradiation in kilowatt-hours per square metre, divided by the reference irradiance of one kilowatt per square metre. Because it's normalised for sunlight, it lets me compare months and plants fairly. A well-built grid-tied plant often sits somewhere around 0.75 to 0.85. What drags it down is heat, soiling, shading, mismatch, cable and inverter losses, clipping and downtime. In hot months PR drops naturally, so I also look at a temperature-corrected PR before blaming the plant."
e_ac = 12400 # kWh exported in the month
p_dc = 100 # installed kWp
h_poa = 155 # kWh per m2 on the module plane in the month
g_stc = 1.0 # kW per m2 at standard test conditions
pr = e_ac / (p_dc * h_poa / g_stc)
print(round(pr, 2)) # 0.8Calculating PR from nameplate and hours of sunshine, or ignoring that the irradiance must be measured on the module plane.
Symptom: what the monitoring or the customer showed.
Narrowing down: compare, isolate, measure, one step at a time.
Cause and fix: the real root cause and what you changed.
Prevention: what you changed so it doesn't happen again.
"At my last company a 30-kilowatt rooftop was making noticeably less than its twin next door. The monitoring showed one MPPT input lower than the others, so I started there. With the DC isolator open, I measured Voc on the two strings on that input. One was short by roughly a third of a module's voltage. I scanned that string with a thermal camera around midday in strong, steady sun and found a module where one whole cell group showed up warmer than the rest, the classic sign of a bypassed section. Its bypass diode had failed short, so that third of the module was being skipped all the time. We replaced it under warranty. After that I added a string-by-string Voc comparison to our commissioning sheet, because the fault had probably been there since installation."
Swapping parts until the problem went away, with no measurements to show why.
Trust the data: irradiance sensor, meter and communication first.
Localise: compare inverters and strings to find where the loss is.
Gradual causes: soiling, vegetation, a slowly failing string.
Confirm and fix: site visit, test, then check PR recovers.
"First I'd make sure the drop is real. A dirty or tilted irradiance sensor, or a meter or data logger problem, can fake a PR change. If the sensor checks out, I'd compare inverters and strings to see whether the loss is spread across the plant or sits in one place. A steady drop across everything over two weeks usually points to soiling, especially after dry, dusty weather or bird activity, and I'd compare a cleaned reference row against a dirty one. If it's localised, I'd look for tripped strings, blown fuses, an inverter derating from heat or blocked air filters, or new shade from growing vegetation. Once I find it, I fix it, then watch the PR for a week to confirm it's back."
Sending a cleaning crew straight away without checking the data or finding where the loss actually is.
Worry: the real concern, which is often not the first one said.
How you answered: plain explanation, their own bill, honest estimate.
Outcome: what they decided, and how it held up.
"A shop owner told me solar was too expensive, but after a few questions the real worry was that he'd been burnt by a cheap installer before and didn't trust generation promises. So instead of a brochure, I took his last year of bills and showed him how I got the size, what the estimated yearly units were and what assumptions sat behind them, including losses from heat and dust. I gave him a conservative number and told him what would make it lower. I also walked him through the monitoring app he'd get, so he could check our claim himself. He signed a few weeks later, and in the first year the output landed a little above the estimate, which is how he ended up referring his brother."
A story where you won by promising the bill would go to zero, or by talking over the customer's concern.
Cause: what slipped and why.
Communication: when and how you told the customer.
Recovery: what you did to limit the delay.
Lesson: what you do differently now.
"We installed a 20-kilowatt system for a clinic, but the utility's meter change and export approval took weeks longer than expected, so the system sat finished but not allowed to export. As soon as I saw the application had stalled, I called the owner instead of waiting for him to ask. I explained exactly which step we were waiting on, what we'd submitted, and that we could run the system for self-use in the meantime within the rules, so he'd still get some savings. I followed up with the utility weekly and sent him a short update every Friday. When approval came through, he wasn't delighted with the wait, but he said the updates were why he trusted us. Now I start grid paperwork before materials are even ordered."
Going quiet until the customer chases you, or blaming the utility without doing anything to move things forward.
Don't attack: thank them and ask to look at it together.
Compare like for like: modules, inverter, structure, cabling, protection, warranties, service.
Show the difference: what the gap buys over the system's life.
Their call: let them choose with clear facts.
"I'd thank them for showing me and ask if we can go through both quotes side by side, because the kilowatt number alone doesn't tell you much. I'd compare the module and inverter makes and warranties, whether the size is DC or AC, the structure material and coating, the cable specification, whether surge protection, earthing and monitoring are included, and what service they get after handover. Often the gap is in exactly those lines. If the other quote is genuinely equivalent, I'd say so and see what I can do. If it isn't, I'd show them what that difference means over twenty-five years, like a rusting structure or no one answering the phone when the inverter fails. And I'd never run the other company down."
Dismissing the competitor as cheap and unreliable, or matching the price without checking what was actually quoted.
Acknowledge: the frustration, and whether it should have been explained earlier.
Explain: the system shuts down on purpose to protect people working on the lines.
Options: a battery or hybrid upgrade if backup matters.
"I'd start by saying I understand it's frustrating to have solar on the roof and still sit in the dark. Then I'd explain that a grid-tied system is designed to switch off when the grid goes down. It's a safety rule, because otherwise it could push power into lines that repair crews think are dead. So the system did exactly what it should, and it'll restart on its own when the grid is back. If nobody told them this when they bought it, I'd own that and make sure our sales notes cover it in future. If backup is important to them, I'd offer to look at a battery or hybrid setup that can keep chosen circuits running during an outage, and give them a clear quote."
Telling the customer it's a fault and sending someone out, or blaming them for not knowing.
The hazard: what you saw, concretely.
Action: how you stopped it and what you said.
Fix: what had to be in place before work restarted.
After: reporting, and what changed for later jobs.
"On a factory roof with fibre cement sheets, I saw two installers walking directly on the sheets to save time instead of using the crawl boards, and one of the skylights along their path wasn't covered. Those sheets can give way under a person. I called them back to the walkway calmly, stopped that part of the job and spoke to the supervisor. We covered the skylights, laid extra crawl boards along the cable route and reran the morning briefing so everyone knew where they could and couldn't step. We lost about an hour. I logged it as a near miss, and after that we added skylight covers and fragile-roof marking to the pre-start checklist for every sheet roof."
Saying you've never seen anything unsafe on a site, or that you let it go to keep the schedule.
The shortcut: what they wanted to do and why it mattered.
How you handled it: facts, not rank.
Outcome: the fix and the relationship afterwards.
"On a commercial rooftop the cabling subcontractor was running short on the DC connectors we'd specified and wanted to join some strings with a different make. Mating connectors from different makers is a common cause of heating and failure years later, because they aren't tested to work together. I didn't make it about who's in charge. I showed the foreman the module and connector manuals that call for matching pairs and explained that a burnt connector on a roof is our callback, not the supplier's. We paused those joints for a day, got the right connectors, and used the day to finish cable tray work. He wasn't happy about the delay, but later he started flagging material shortages to me before they became a problem."
Accepting the shortcut to keep the schedule, or handling it by pulling rank and damaging the relationship.
Stop: get people back from the edge right away.
Fix: edge protection or anchors and harnesses as the method statement says.
Restart: only when protection is in place and checked.
Report: log it and fix the planning gap.
"I'd stop the work straight away and ask everyone to move back from the edge, calmly, without making a scene. A fall from a roof is one of the worst things that can happen on a solar job, so this isn't something to discuss while they carry on. Then I'd find out why the protection wasn't set up, whether the equipment didn't arrive, or the crew thought it was a short job. Work near the edge only restarts once there's edge protection or proper anchor points and everyone is clipped on, as the method statement says. I'd log it as a near miss and check with the supervisor that the fall protection is on the delivery list and the pre-start check for every job."
Letting the work continue because it's almost finished or because the crew is experienced.
Non-negotiables: safety, connectors, earthing, torque, testing.
Where you flex: sequencing, extra hands, order of work.
How you work with others: early warnings, not last-minute surprises.
"For me, a solar system has to be right for twenty-five years, not just on handover day. So a few things never get rushed: fall protection, connectors crimped and mated properly, earthing, torque on the structure, and the full commissioning tests. Most failures I've seen years later came from one of those. Where I'll flex is everything around them. I'll resequence the work, bring cable pulling forward while we wait for a delivery, or ask for extra hands on a critical day. And I tell the project manager early if something threatens the date, so we can plan around it instead of cutting a corner at the end. I like teams where saying 'this needs another day' is treated as doing the job, not slowing it down."
Saying you'll do whatever it takes to hit the date, with no line you won't cross.
ClapAssist is an AI interview assistant for Mac and Windows. It listens to the interview on your computer and shows you what to say, in short lines you can read while you talk. Your resume and notes are never stored on our servers. It stays out of screen share on every plan; only you can see it.