I've been on the receiving end of that call more times than I can count: the crew is scheduled to arrive on site in two days and there's no power.
I coordinate emergency off-grid energy deployments for a company that handles construction sites, temporary offices, and event venues. Over the past three years I've handled roughly 40 rush deployments — maybe closer to 60 if you count the standard projects that got compressed into half their normal timeline. When you're working at that pace, things get missed unless you have a system.
This is the checklist I've settled on. Not the theoretical version — the one I'd actually walk through if the call came in tomorrow morning.
Here are the five steps:
- Audit the actual load (not what you assume it'll be)
- Size the battery bank by usable capacity, not nameplate
- Spec and configure the MPPT controller correctly
- Don't skip surge protection — especially for comms gear
- Load test, not just power-on test
Step 1: Audit the Actual Load
This is the step most people skip. I've watched contractors guess "probably 5 kilowatts," order based on that number, and then trip the system when a single 5 kW-rated motor pulls 8 kW on startup.
Don't guess. Measure, or work from nameplate data if you can't measure.
For inductive loads — motors, compressors, HVAC — multiply continuous running watts by 2 to 3 to get startup peak. For resistive loads, add a 10–20% safety margin.
One question I get a lot: how many kWh does a level 2 charger use?
A level 2 EV charger typically draws between 6.2 kW and 7.7 kW continuously. At roughly 7.4 kWh per hour, charging a typical EV from near-empty to full (roughly 50 kWh needed) takes about 6.8 hours.
But here's the realistic take: that kind of charging load off-grid is a different order of magnitude. Most field deployments I've done either exclude EV charging from the off-grid circuit entirely or treat it as a non-critical load that gets shed when battery reserves drop. If it has to run off-grid, size for it early — don't discover the requirement after the system design is finalized.
For a construction site or event venue, start with a critical-load list: lighting, communications equipment, office devices, charging stations. Rank by priority. Then design with headroom.
Step 2: Size the Battery Bank by Usable Capacity
For field deployments, LiFePO4 is the default choice now. Cycle life, safety profile, and temperature tolerance beat the alternatives for this use case.
On a project last year, I used LiTime 100Ah LiFePO4 batteries in the bank. Nameplate capacity: 100Ah × 12.8V = 1.28 kWh.
But that's not what you can actually draw.
Usable capacity gets eaten by a few layers: depth of discharge (LiFePO4 is usually designed to run to about 90%), inverter efficiency (85–90%), and the battery management system's own overhead.
The realistic number is closer to 1.0–1.1 kWh usable per battery.
So if the site needs 10 kWh usable to run overnight, you're looking at a minimum of 10 batteries. That's a significant footprint — plan transport and placement before you're on site, not after.
I learned this one the hard way. In March 2024, on an event deployment, I sized the bank by nameplate. The BMS cut off early when the evening load spiked, and we lost power to part of the venue for about 40 minutes. Not catastrophic, but embarrassing — and it cost me two hours of explaining to the client that I shouldn't have needed to spend.
Now I always add a 15–20% buffer to battery capacity. Carrying two extra batteries is cheaper than apologizing on site.
Step 3: Spec and Configure the MPPT Controller Correctly
The MPPT controller determines how much power you can actually pull from the solar array. Undersize it and you're leaving generation on the table. Oversize it and you're burning budget.
Victron Energy's MPPT 100/30 is one I've used on a lot of these deployments. The specs are straightforward: 100V max PV input, 30A max charge current.
What that means in practice:
- 12V system: up to 30A × 12V ≈ 360W of PV input
- 24V system: up to 30A × 24V ≈ 720W
- 48V system: up to 30A × 48V ≈ 1440W
(Actual output varies with temperature and panel operating point — use these as planning numbers, not hard limits.)
If you need more than a single 100/30 can handle at your chosen system voltage, either parallel multiple MPPT controllers or step up to a larger model.
For configuration, read the Victron Energy MPPT control manual for your specific model. Don't skim it. The wiring conventions and DIP switch settings vary between models in ways that will bite you if you assume.
Use the VictronConnect app for initial setup. Set battery type, charge voltages, and enable DVCC if applicable. I once skipped setting the charge voltage on a field install — the batteries never got above 80% and I spent an hour troubleshooting the array before realizing it was a config issue, not hardware. Design wasn't wrong. My setup was.
Checklist before you connect anything:
- PV array open-circuit voltage must stay under 100V. That's a hard limit, not a suggestion.
- Battery polarity reversed will kill the controller. Don't ask me how I know.
- Connect battery first, then PV. Reverse the order on disconnect. The manual says this. People still get it wrong.
Step 4: Don't Skip Surge Protection (Especially for Comms Gear)
This is the one I'm most stubborn about.
On a deployment last year, the operations manager told me "surge protectors are an extra cost we can skip — just get the system running." I pushed back and installed them anyway. It wasn't a popular decision in that meeting.
Three months later, a thunderstorm took out the inverter and two batteries on another system at the same site. Around $3,000 in damage, plus downtime. The system I'd specified had surge protection on the AC output, DC side, and the modem/router line. It came through without a scratch.
That operations manager never brought up skipping surge protection again.
For field deployments, surge protection should cover:
- AC output side: protects against back-feed from grid or generator
- DC/PV side: solar arrays pick up transients, especially when runs exceed about 10 meters
- Communications equipment: cellular modems, routers, satellite terminals
That last one is the most overlooked. A modem surge protector is an inexpensive part, but it's the difference between losing connectivity when a surge hits and keeping your remote monitoring, automated alerts, and site communications online.
Part of me understands why people see it as an optional add-on — it's a line item that doesn't seem to do anything until the day it does. But I've been on the wrong side of that trade-off once. I'd rather have the part and never need it.
I first realized how critical this was after a deployment where an unprotected modem got fried in the first week. The system itself kept running fine, but I lost all remote telemetry and alarm capability. That meant sending someone to site — 80 kilometers round trip — every time we needed a status check.
Now I don't sign off on a deployment if the comms equipment isn't surge-protected. It's not negotiable.
Step 5: Load Test, Not Just Power-On Test
Powering on and seeing the lights come on doesn't mean the system is ready to hand over.
My final commissioning process:
- Apply a known resistive load (a heater or halogen work light) at about 50% of system capacity for 30 minutes. Watch voltage drop and temperature rise.
- Apply the actual load and run it for at least 2 hours. Monitor the battery voltage curve — if it sags too fast under real load, something's wrong with the sizing or the cells.
- Test transition (if the system has grid/generator failover) — confirm the inverter switches cleanly.
- Verify MPPT behavior with load switching — go from full load to near-zero and confirm the controller re-tracks without faulting.
When you hand off to the site supervisor, document the essentials. I know it feels like extra work in the moment, but a setup sheet that answers "I flipped the switch and nothing happened" at 2 a.m. is worth its weight.
Building this checklist has actually been one of the better operational decisions I've made — and I say that as someone who resisted formalizing it at first. I tried winging it once, missed Step 4 on a communications circuit, and had to ship a replacement modem overnight. After that, I printed the checklist and taped it to my desk.
Things That Go Wrong (Or Almost Do)
A few lessons worth pinning somewhere visible:
Underestimating inverter idle draw. Inverters consume power even when they're not powering anything — usually 10–30W. If a system needs to run for three days, that overhead adds up. Factor it into the bank sizing.
Confusing nameplate capacity with usable capacity. This is the most common mistake I see. Batteries are sold by nameplate specs. Sites run on usable energy.
Sizing PV for rated output. If the array is rated at 1000W but actual output is closer to 700W, that's not a defective panel — it's tilt angle, temperature, and dust. Design to the realistic output, not the spec sheet.
Undersizing the controller. One 100/30 on a 24V system looks fine until you're trying to push 800W of PV through it. You need two units or a larger model.
Skipping surge protection on comms equipment. Hard to overstate this one. It's the cheapest insurance on the whole system.
Cutting system capacity to hit a budget number. I get it — budgets are real. But adding capacity after the fact costs more than specifying it right the first time. If the project allows it, design for peak load, not average.
At the end of the day, speed alone doesn't win repeat work. It's speed without the callbacks. The checklist isn't glamorous, but it's the reason we can move fast and still get it right on the first pass.