The cap usually binds before the ports do
Run the defaults and the yard has twelve ports, but only eleven of them can run together. A 19.2 kW port draws 20.9 kW at the meter once the 92 percent efficiency is applied, the 250 kW cap has 235 kW left after the 15 kW of site load, and 235 divided by 20.9 is eleven and a bit. The twelfth port is a fitting on a wall.
That pattern is the normal one, and it is why counting chargers gives the wrong answer. The port count is what a depot buys and the cap is what actually decides how many run, and the two are set by different people at different times.
Reading the two finish times
The page prints a floor and a schedule. With the defaults the floor is 5 h 16 min — the page reports 2.4 days of plug time shared across eleven ports — and the schedule finishes at 6 h 34 min. Both are correct and the gap between them is real: vehicles do not split across ports, so the last few land badly and one port ends up carrying two long jobs.
Neither number is the interesting one on its own. What matters is that both sit inside the 9 hour window, leaving nearly two and a half hours of slack. Slack is what absorbs a vehicle coming back late, and a yard planned to finish at exactly the window length has none.
When it does not fit, it says so
Drop the cap to 120 kW and the same fleet gets five ports at once. The schedule needs 12 h 18 min against a 9 hour window and only sixteen of the twenty-four vehicles are done in time. The page prints the shortfall and the count rather than truncating the list and reporting a tidy finish time, because a schedule that quietly drops six vehicles is worse than no schedule.
There is a harder version of the same failure. If a single vehicle needs more hours than the window has, no number of ports helps — it already has one to itself. The page checks for that case separately and says outright that buying chargers will not fix it. The only levers left are a longer window, more output per port, or that vehicle coming home with more in the pack.
What is not in here
Vehicles arrive over an evening rather than all at once, and a return spread over two hours behaves differently from twenty-four vehicles landing together. Charging power tapers near the top of a pack, so a vehicle topping up the last few points averages well under the port rating. And where load management hardware is fitted, it shares a power budget continuously between whatever is plugged in rather than switching whole ports on and off, which usually beats this schedule slightly.
The finish time is therefore something to test against a real night with a stopwatch and a submeter, then adjust. The part that generalises is the structure: find how many ports the cap allows, share the plug hours across them, and compare against the deadline.
Questions people ask
How many chargers does a fleet depot need?
Fewer than the vehicle count, and often fewer than the ports already installed can run. With the defaults, twenty-four vehicles finish in six and a half hours on eleven simultaneous ports, and the eleven is set by the site power cap rather than by the twelve ports on the wall. Work the plug hours, divide by the ports the cap allows, and compare against the window.
What limits how many EV chargers can run at once?
On this page, the site power cap you supply, less whatever else the site is drawing, divided by what one port pulls at the meter. That is arithmetic. What your service, feeder or transformer can actually carry is a load calculation performed by a licensed electrician against your adopted code, with the utility involved where the supply is in question, and nothing here answers it.
What happens if the charging window is too short?
The page says so and reports how many vehicles finish rather than quietly truncating. It also separates two different failures: not enough ports running together, which more capacity fixes, and a single vehicle needing longer than the window, which it does not. In the second case adding chargers buys nothing, because that vehicle already had a port to itself.
Why does the schedule take longer than the total divided by the ports?
Because vehicles do not split across ports. With the defaults the arithmetic floor is 5 h 16 min and the achievable schedule is 6 h 34 min, a 24.6 percent gap that comes entirely from whole jobs landing unevenly at the end. The floor is printed alongside so you can see how much of the finish time is the workload and how much is the packing.
Does this include the cost of charging the fleet?
Only the energy, and only if you enter a rate. The larger number on most commercial bills is the demand charge, which is billed on the highest short interval of the month rather than on total energy, and it is worked out separately. The peak draw figure on this page is the input that page needs.