Capacity, Duty Cycle and Dose

A stored quantity divided by a rate — and the rate is the hard part

How long a tank lasts, how long a battery runs, how long a cartridge treats: every one of these is a capacity divided by a rate of consumption. The division is trivial and everybody gets it right.

What gets the answer wrong is treating the rate as a constant. Consumption depends on load, on temperature, on how the equipment is being used and on how it degrades — and in most of these systems it varies more between operating conditions than any error in the capacity figure ever could.

So the honest form of every answer here is a duration AT A STATED RATE, and the stated rate is an assumption rather than a measurement. The pages say which rate they used, because a runtime quoted without its load is a number that will be wrong by a factor rather than by a percentage.

A generator is least efficient when it is barely working

A generator's fuel consumption is not proportional to its output. A substantial share is consumed simply keeping the engine turning, so at low load the fuel per unit of electricity produced is far worse than at three-quarters load — which is where most machines are near their best.

That inverts the intuition about sizing. An oversized generator running a small load does not sip fuel; it burns a disproportionate amount of it for the work done, and the runtime from a given tank is much shorter than a linear scaling suggests.

It also does real damage. A diesel engine run lightly loaded for long periods does not reach its designed cylinder temperatures, so fuel and lubricating oil pass unburnt into the exhaust — WET STACKING — where they foul the turbocharger, glaze the bores and eventually reduce the engine's ability to take load at all. The remedy is a load bank or a correctly sized machine, and the failure is caused by under-use rather than over-use.

So a runtime calculation has a sizing consequence attached: the honest answer to a long required runtime is usually a machine matched to the load with a larger tank, not a larger machine.

A propane tank can be full and unable to deliver

This is the fact the page exists for. Liquefied petroleum gas leaves a tank as VAPOUR, and the vapour has to be boiled off the surface of the liquid. Boiling absorbs heat, which the liquid takes from the tank wall and ultimately from the surrounding air.

So a tank's ability to supply gas is a heat-transfer limit, not a volume one. It depends on the WETTED SURFACE AREA — the area of tank wall in contact with liquid — and on the temperature difference to the air outside. As the tank empties the wetted area shrinks, and as the weather cools the temperature difference shrinks, and the vaporisation rate falls with both.

The consequence is a tank that holds plenty of fuel and cannot meet a peak demand on a cold morning when it is a quarter full. The appliance starves, the regulator freezes, and everything downstream behaves as though the tank were empty.

That is why tank sizing is driven by peak demand and by the coldest expected temperature rather than by how often anyone wants a delivery, why multiple smaller tanks manifolded together can out-perform one large one of the same total volume, and why frost on the outside of a tank is a diagnostic rather than a curiosity: it marks the liquid level, and it means the tank is working hard at its limit.

Softener capacity is a trade against salt, not a fixed number

An ion exchange softener's capacity between regenerations is not a property of the resin alone. It depends on how much salt is used in each regeneration, and the relationship is strongly diminishing: a heavy regeneration restores more capacity but a smaller share of what the salt could theoretically have delivered.

So a softener can be run two ways. Regenerating hard gets more litres between cycles and uses more salt per litre treated; regenerating lightly is far more salt-efficient and means more frequent regenerations. Neither is wrong, and a manufacturer's headline capacity is almost always the high-salt figure while its efficiency rating is the low-salt one — which is why the two numbers in a brochure cannot both apply at once.

Sizing follows from the household's hardness and daily use: the grains removed per day is hardness times volume, and the resin volume is chosen to give a sensible interval. Oversizing hurts here too, for a reason particular to this equipment: resin that sits unregenerated for a long time between cycles fouls and channels, so a softener that regenerates rarely performs worse rather than better.

The other input people omit is that regeneration itself uses WATER, and a good deal of it. A system's true consumption is the softened water delivered plus the backwash and rinse discharged, and on a metered supply or a septic system that discharge is a real quantity rather than a rounding.

A dose is a property of both materials

Dosing calculations look like coverage rates and behave differently in one important way: the dose depends on what is being treated as much as on the treating agent.

Composting is the clearest case. The target is a CARBON-TO-NITROGEN ratio in the blended mixture, and the blend is a mass balance across two materials with different ratios and different moisture contents — so the quantity of one needed is a function of the other's composition, not a fixed proportion. Get it wrong toward nitrogen and the pile goes anaerobic and smells; toward carbon and it simply does not heat.

Moisture is the second balance running alongside the first and it is easy to lose. Materials are usually quoted wet, while the carbon-to-nitrogen ratio is a dry-mass relationship, so a blend computed on as-delivered masses is computing the wrong ratio. The conversion is the same discipline the soil-state paper describes: reference everything to the one phase that does not change.

Chemical solidification of washout slurry works the same way. The dose depends on the slurry's water content and on its chemistry, so a figure per unit volume is a starting point that is confirmed by trial on the actual material. And it is worth stating what solidifying achieves: it makes a liquid manageable as a solid, and it does NOT neutralise it. Concrete washout remains strongly alkaline after it is solidified, and its disposal route is determined by that rather than by its consistency.

What has to be measured

Every quantity on this page has an observation that beats the calculation. A generator's real consumption comes from its fuel gauge across a known period at the actual load; a tank's real vaporisation limit shows itself on the coldest morning of the year; a softener's real capacity is the point at which hardness appears at the tap, which is a test strip rather than a specification.

So the useful role of these calculators is to size before the system exists and to explain behaviour once it does. A generator that runs out earlier than predicted, a gas appliance that falters only in January, a softener that has stopped softening a week before it should — each of those has an explanation on this page, and the explanation is usually that a rate assumed constant is not.

And each page states the assumption it used, because a duration or a dose quoted without its conditions is not a wrong answer so much as an unfinished one.