How I test and rate power stations

I'm Scott Caine, and I test every power station on this site myself. Each unit is bought at retail, put through the same bench tests, and then used for a month or more in the places these units actually end up: a kitchen during an outage, a van, a campsite, a bedroom with a CPAP running.

Portable power stations are sold on a rated capacity printed on the box. What reaches the AC outlet is less, and manufacturers rarely say how much less. Measuring that gap on every unit, under the same conditions, is the core of what this site does.

So far I've tested eleven units from four brands, using each one for between 27 and 44 days:

• EcoFlow: Delta 3, Delta Pro 3, River 2 Pro
• Bluetti: Elite 100 V2, Elite 200 V2, AC70
• Anker: SOLIX C1000 Gen 2, SOLIX F3800 Plus, SOLIX C300
• Jackery: Explorer 1000 v2, Explorer 2000 v2

All eleven were bought at retail. No manufacturer supplied, loaned or discounted any of them. See how this site is funded.

Every unit goes through the same tests, so a result on one review can be compared directly with a result on another.

Because the capacity test uses one load for every unit, the usable percentages on this site are directly comparable with each other. Figures published by different laboratories often use different loads, so they can't be lined up against each other in the same way.

Bench figures don't tell you whether a handle digs into your fingers, whether a fan wakes you at 3 a.m., or whether the app forgets a setting after a firmware update. So after the bench, each unit moves into daily use for between 27 and 44 days.

That use is organised around seven real-world jobs. I rate each unit out of 5 on every job it was put to, and the reviews say what happened.

Outage drills. For the two large units — the EcoFlow Delta Pro 3 and Anker SOLIX F3800 Plus — I ran real household circuits through a transfer switch fitted by a licensed electrician. For the smaller units I ran a fridge and router off extension leads, which is how most people use them. The big and small units ran different fridges, so the observed runtimes compare within a size class, not across classes.

CPAP nights. My machine averaged about 23 W with the humidifier off, which is lighter than most. The CPAP page gives my observed nights alongside calculated nights at a typical 35 W and 70 W, so you can work from the figure that matches your machine.

These use-case ratings drive the picks on the use-case pages. They are not part of the overall score.

Every scored unit gets one overall score out of ten. It's built from two halves that are kept separate on every review, so you can see where a unit gains and loses.

The measured score uses five criteria with fixed weights, applied by formula to my bench measurements and the manufacturers' specifications. The score box on each review prints the full sum.

• Usable capacity and output (30%) — how much energy reaches the socket, and how much power the unit can deliver.
• Recharge and solar input (20%) — how fast it refills from the wall and how much solar it can take.
• Build, ports and durability (20%) — the ports, and the rated cycle life, including whether cycles are rated to 70% or 80% of original capacity.
• Portability (15%) — weight on my scale.
• Value (15%) — the manufacturer's list price divided by my measured usable watt-hours.

The weights are deliberate. Usable capacity and output carries 30% because it's the figure most often misrepresented, and it decides whether the unit does its job. Value carries only 15% because the cheapest unit per watt-hour is often the wrong unit for the use case.

After living with each unit, I rate it out of 5 in nine categories. They carry equal weight and are averaged.

The overall score is half measured, half hands-on. The hands-on average is out of 5, so it's doubled to put it on the same ten-point scale:

Overall = (measured score + hands-on average × 2) ÷ 2

For the EcoFlow Delta 3, the measured score is 7.94. Its nine hands-on ratings total 39 out of 45, an average of 4.33, which doubles to 8.67. (7.94 + 8.67) ÷ 2 = 8.30.

Reviews show the hands-on average rounded to one decimal place (4.3 for the Delta 3). The formula uses the unrounded figure, which is why doubling the rounded number doesn't always reproduce the overall score exactly.

Runtimes on this site are calculated unless a page says they were observed. I take my measured usable capacity and divide it by a stated load. These loads are fixed across every page so the comparisons stay honest:

So the EcoFlow Delta 3, measured at 862 Wh usable, runs the camp fridge for 862 ÷ 45 = 19.16 hours, and I print 19 h 09.

These are estimates with the working shown. Your fridge, your sun and your cable losses will differ. Where I also observed a runtime, the review says which is which. The Delta Pro 3 ran a fridge for 80 hours in my outage drill, beating its calculated household-fridge figure, because my fridge drew a little less than the assumed average.

• Street prices as figures to rely on. They move weekly and I don't track them. Every price in a table, a score or a cost-per-watt-hour figure is the manufacturer's list price on the date I checked. Where a unit is routinely sold below list, I say so in the text.
• Calculated runtimes dressed up as observations. Every runtime is labelled as either calculated or observed.
• Long-term durability findings I don't have. I've used each unit for weeks, not years. Cycle life and warranty terms are the manufacturer's ratings, reported as claims.
• Scores for units I haven't tested. They can appear in spec tables but not in rankings.

If a figure here is wrong, I want to fix it and say that I did. Every page carries a modified date, and material corrections are noted on the page itself rather than silently edited. Tell me what I got wrong.