Hand-considered hypotheses, 3701–3800

Refrigeration and cryogenics, written one at a time. Vapour-compression cycles and their accounting, compressors with their oils and refrigerant pairs, expansion and two-phase flow in evaporators, regenerative cryocoolers, vacuum and cryopumping, the cryogenic insulation interface, materials at low temperature, and frost with defrost. Each entry carries the logical form chosen for it, printed above the claim, so the variety of argument can be audited against the text instead of taken on trust. None has been tested; no citation is given because none is being invoked, and every quantity offered as an estimate is marked as one.

Cycles, accounting and part load

Misattribution
H3701. Measured cycle COP is usually debited to the compressor's isentropic efficiency, but in a laboratory rig a comparable share of the shortfall can be heat picked up by the suction line between the evaporator outlet and the compressor inlet, which raises mass flow without raising refrigerating effect. The two separate because the line gain scales with the difference between line and ambient temperature while the efficiency term does not. Instrumenting both ends of the suction line at two ambient temperatures decides it.
Bookkeeping deficit
H3702. Subcooling is booked as recovered capacity, yet when it is produced by a suction-line heat exchanger the compressor's inlet density and its power rise along with it, so the two sides of the ledger are kept in different units. Restating both as exergy removes most of the apparent gain, and what survives is the exchanger's approach temperature read against the condensing temperature.
Threshold
H3703. There is a receiver charge below which an expansion valve stops feeding steadily and begins to hunt, and the threshold is set by its charge relative to the coil's internal volume rather than by the superheat setting. It appears first as a periodic suction pressure at constant load, a periodicity that is routinely read as a valve fault.
Inversion
H3704. Reducing refrigerant charge in a capillary-tube system can increase capacity over a range, because a less flooded condenser exposes more area and lowers the condensing pressure; the assumption that undercharge only costs capacity holds only above the charge at which the evaporator stops being fully wetted. A charge series at fixed ambient, measuring capacity and wetting together, shows where the sign changes.
Identity claim
H3705. The compressor inefficiency and the cycle irreversibility reported for one machine are a single entropy production assigned to two different control volumes, so with the boundaries drawn consistently the two numbers must agree. The disagreement between the two literatures is therefore about where the cut is made, not about the machine.
Crossover
H3706. Circuit-to-circuit imbalance in a multi-evaporator system switches from liquid-limited to pressure-limited as load falls, so a header that starves the far circuits at full load can balance them at half load. There is a load at which the imbalance vanishes whatever the header design, and locating it costs one afternoon of instrumentation.
Saturation
H3707. Superheat hunting in a thermostatic valve saturates rather than grows with disturbance size: beyond a certain amplitude the valve is fully open for part of every cycle, so the mean superheat stops rising while the cycling amplitude keeps growing. Mean superheat is therefore a poor indicator of instability, and the amplitude is the honest one.
Count
H3708. In a cascade plant there are at most as many independently trimmable charge quantities as there are temperature levels, so any commissioning sheet that trims more is adjusting one quantity twice. The count follows from the degrees of freedom, not from the hardware.
Lag
H3709. Capacity measured at start-up lags the steady value by the evaporator's own thermal mass, and for a coil whose metal outweighs its refrigerant charge the lag is minutes rather than seconds. Short-cycle tests therefore compare machines on their thermal mass while claiming to compare their capacity.
Null model
H3710. Most of the seasonal advantage claimed for variable-speed over fixed-speed equipment disappears when both are compared at the same average part-load ratio with cycling losses and standby power included. The residual advantage is the ratio of the two part-load curves, obtainable from a few days of run-time data rather than a season.
Signature
H3711. If condenser fouling is causing a slow capacity decline, the approach temperature rises while discharge superheat stays flat; if charge loss is causing it, superheat rises while approach stays flat. The pair of trends is diagnostic and needs no work on the circuit, only two sensors read together.
Sufficiency claim
H3712. One pull-down curve from ambient to setpoint is sufficient to identify both a cabinet's UA and its effective thermal mass, provided the compressor's capacity curve is known. The two are separable because they enter the curve with different time constants, so the fit is not the underdetermined exercise it appears to be.
Degeneracy
H3713. Charge and capillary setting are degenerate in a small system: many pairs give the same evaporating temperature at one ambient, and only a second ambient breaks the degeneracy. Commissioning at a single ambient therefore fixes nothing, so identical units shipped to different climates are not identical units.
Energy audit
H3714. Any claimed COP above the Carnot value for its own measured temperatures is an audit failure at the boundary, and the usual culprit is an air-side temperature used where a refrigerant temperature belongs. Recomputing the claim against its own stated boundary settles it without new experiments.
Non-monotonic window
H3715. Evaporator capacity does not rise monotonically with fan power: below the optimum the air side is film-limited, above it the fan's own heat and coil dry-out exceed the gain. The maximum is interior, and its position moves with the latent load, so a fan curve quoted at one condition misleads at another.

Compressors, oil and refrigerant pairs

Hidden variable
H3716. The capacity of a hermetic system at fixed pressures depends on the fraction of oil circulating, which changes with compressor speed and is almost never measured. Two nominally identical units can therefore differ by more than the tolerance on their compressors, and the difference is invisible to a refrigerant-side test.
Reversal
H3717. Small oil fractions can raise capacity by sealing clearance gaps, and larger ones lower it by raising pressure drop and fouling the evaporator, with the sign change at a low fraction that many designs cross. This is why oil-free is not automatically better, and why oil-return tests at one charge disagree.
Scale argument
H3718. At small displacements the re-expansion loss scales with pressure ratio but leakage scales with clearance cubed, so a small compressor's efficiency is set by manufacturing tolerance rather than by port design. There is a displacement below which tighter tolerances beat better porting, and it is larger than current miniaturisation assumes.
Hidden variable
H3719. The viscosity that governs start-up is not the grade on the datasheet but the viscosity index, because the bearing sees the oil at the coldest point it reaches and the film collapses there first. A compressor can pass a running test and fail on its first cold start with no measurable difference at steady state.
Artefact
H3720. The efficiency gain reported for a new refrigerant in a drop-in test is often the effect of the lubricant it was paired with, since a drop-in changes both fluids at once. Running the new refrigerant in the old oil, and the old refrigerant in the new oil, separates them at the cost of two extra charges.
Identity claim
H3721. Two-phase pressure drop and boiling heat transfer in a tube are governed by the same interfacial structure, so a correlation set that fits one should fit the other through the same flow map. Disagreement between the two literatures is then disagreement about the map, which is a smaller and more tractable dispute.
Selection effect
H3722. Published refrigerant-oil solubility data are biased toward miscible pairs, because immiscible pairs are hard to measure and rarely submitted. Design tools therefore extrapolate into the region where the interesting failures occur, and the bias is invisible because the database looks large.
Competing rates
H3723. Refrigerant absorption into oil is diffusion-limited at short times and convection-limited at long times, and the crossover is why an oil charge appears to equilibrate in minutes on a bench and in hours in a field unit. Bench solubility data taken before the crossover are measurements of a transient.
Impossibility
H3724. No simple single-refrigerant circuit can match a zeotropic blend's glide to a counterflow load without a composition shift somewhere in the loop, so glide matching and composition stability cannot both be had in one circuit. Designs claiming both have moved the shift into the accumulator, where it is not measured.
Ratio invariant
H3725. For a given pressure ratio the ratio of discharge temperature rise to isentropic temperature rise is nearly fixed by the gas's heat capacity ratio, so discharge temperature is a better indicator of pressure ratio than of efficiency. Using it as an efficiency indicator inverts the diagnosis in exactly the cases that matter.
Hysteresis claim
H3726. Hysteresis in a compressor map between increasing and decreasing speed is thermal rather than aerodynamic, arising from casing expansion changing tip clearance. The loop width should therefore scale with the casing's time constant and vanish on a water-cooled shell, a comparison no map currently reports.
Hidden term
H3727. The temperature at the compressor shell is not the suction temperature at the valve, and the difference is heat picked up in the motor windings. Using shell temperature as the evaporator outlet biases superheat and mass flow in opposite directions, so the two errors partly cancel in COP and do not cancel in capacity.
Two-regime claim
H3728. Below a few percent oil the mixture's viscosity is set by the refrigerant dissolved in it, above that by the oil's own grade, so oil-return tests at different charges disagree because they sit on different sides of the crossover (estimate for typical polyol-ester systems). No single test at one charge can locate the crossover.

Expansion, two-phase flow and evaporators

Misattribution
H3729. Frost on a distributor is credited to a starved circuit, but it is often the liquid line flashing at a restriction upstream of the distributor. The two differ because distributor frost appears on every circuit at once, while starvation frosts one.
Hidden term
H3730. Superheat measured at the evaporator outlet includes the sensor's contact resistance, which changes as oil deposits on it, so a slow seasonal drift in superheat can be a dirty sensor rather than a changed load. A second, cleaned sensor at the same station separates drift from load without touching the valve.
Bound
H3731. An evaporator cannot present a surface colder than the refrigerant's local saturation temperature, so a coil reported to run below dew point while its refrigerant is above it is reporting its sensor's conduction path. The bound is thermodynamic and needs no experiment, only a consistent pair of readings.
Threshold
H3732. Below a mass flux a horizontal evaporator stratifies and its upper surface dries out, so capacity falls off a step rather than a slope. The threshold follows the liquid-to-vapour density ratio rather than the heat flux, so different refrigerants reach it at different qualities.
Inversion
H3733. Increasing the number of parallel circuits can lower part-load capacity, because each circuit then carries a lower mass flux and stratifies; the rule that more circuits reduce pressure drop is a full-load rule. A circuit-count series run at half load shows the inversion and is cheap to run.
Identifiability limit
H3734. Evaporator UA and refrigerant maldistribution cannot both be identified from one air-side traverse, since a uniformly underfed coil and a badly distributed one produce similar profiles. A second traverse at a different airflow breaks the degeneracy, so installations reporting both from one test are reporting a fitted ratio.
Conservation audit
H3735. An evaporator's air-side balance omits fan power and the latent heat of condensate leaving the coil, and the omission grows as the coil's sensible heat ratio falls. Audits that close the balance by adjusting UA are adjusting the wrong term, and the error is largest in the humid conditions that decide the design.
Symmetry argument
H3736. Optimising a coil for cooling anti-optimises it for defrost, because both duties are set by fin spacing, so a design reporting improvement in both has changed its defrost termination criterion rather than its geometry. Running both tests on one coil, with the criterion held fixed, exposes the substitution.
Sign prediction
H3737. An oversized thermostatic valve oscillates with a nearly constant period, while an undersized one sits at steady superheat above setpoint and does not oscillate, so the two failures have different signatures at the same mean superheat. Reading the period as well as the mean separates them without opening the circuit.
Hidden variable
H3738. The valve's bulb charge sets maximum operating pressure far more than the superheat setting does, so a system that loses capacity at high ambient is usually reporting its bulb charge. Swapping two bulbs of different charge at fixed adjustment demonstrates it.
Scale argument
H3739. At tube diameters below the capillary length the liquid film stops being distributed by gravity and starts being distributed by surface tension, so stratified-flow correlations fail at small diameter for reasons that have nothing to do with the refrigerant. The failure is geometric and predictable from the Bond number alone.
Artefact
H3740. Reported evaporator fouling is often the air sensor sitting in the fan's recirculation zone, reading a colder and dirtier-looking stream than the coil sees. Moving one sensor, with nothing else changed, tests it. A coil cleaned while the sensor stays put keeps reporting the same fouling, which is the tell.
Lag
H3741. Coil response to a valve step lags by the refrigerant inventory's residence time, and the lag grows with charge, so a fast test underestimates the steady gain from any valve change. The underestimate is systematic and grows with the very charge that improves part-load behaviour.
Consistency requirement
H3742. A reversible unit's two coils must each be sized for the worse of their two duties, so a claim of symmetric performance implies one coil is oversized and its extra area is being paid for in fan power. Measuring both duties on the same unit shows which coil carries the penalty.

Cryocoolers and regenerative machines

Misattribution
H3743. Pulse-tube loss at low temperature is commonly debited to the regenerator's axial conduction, but at liquid-helium temperatures a comparable share can be the working gas's volumetric heat capacity falling relative to the matrix's, which changes phasing rather than conduction. The two separate because only the second improves when charge pressure is raised at fixed geometry.
Two-regime claim
H3744. A Stirling cooler has a frequency below which regenerator ineffectiveness dominates and above which gap-seal pressure-drop hysteresis dominates, and the two regimes want different charge pressures. One operating point therefore cannot be optimal for both an 80 K stage and a 4 K stage, and the two-stage machine is the compromise that follows.
Count
H3745. Pulse tubes have at most three independently tunable quantities, charge pressure, frequency, and one impedance, that set their cold-end temperature at fixed geometry. An optimisation reporting four parameters is fitting one of them twice, and its optimum will not reproduce on a second machine.
Reversal
H3746. Increasing regenerator mass helps a Gifford-McMahon cooler at low frequency and hurts it at high frequency, because the matrix's heat capacity must be read against the gas's per-cycle enthalpy swing rather than against the load. The optimum mass is therefore a function of frequency and moves whenever the machine is sped up.
Identity claim
H3747. Double-inlet and inertance tubes are the same phase-shifting mechanism presented at two impedances, so a machine fitted with both is over-parameterised and its optimum is a line rather than a point. Mapping the line, instead of hunting a point on it, is the cheaper experiment.
Crossover
H3748. Below a certain cold-end temperature a regenerative cooler's capacity becomes limited by the working gas's real-gas enthalpy rather than by the regenerator, so further matrix improvement buys nothing. The crossover is where the gas's isentropic exponent stops changing with temperature, and it can be found from property data before hardware is built.
Artefact
H3749. Quoted no-load temperatures include conduction down the cooler's own cold finger and radiation through its shroud, and both scale with the same cold area. The quoted base temperature is therefore a statement about the mounting as much as about the machine, and two laboratories can disagree by more than their thermometry.
Operator artefact
H3750. Cooldown curves taken with a heater clamped to the cold tip contain the heater's heat capacity and its lead conduction, and both masquerade as a slow machine. A curve with the load applied radiatively and no heater block is a different measurement, not a repeat.
Material limit
H3751. A single-stage regenerative cooler cannot pass the point where its regenerator's binder or the working gas itself changes state, and the limit is material rather than thermodynamic. It appears as a step in capacity rather than a gradual decline, and it is often mistaken for a vacuum problem.
Scale argument
H3752. In a pulse tube the ratio of wall conduction to gross refrigeration scales inversely with cold-tip diameter while regenerator effectiveness scales with length, so below a diameter the machine is a heat leak with a refrigerator attached (estimate). Miniaturisation therefore has a floor set by geometry, not by the drive electronics.
Lag
H3753. Cold-end temperature responds to a load step with two time constants, the cold mass and the regenerator, and the second is usually the longer. Load maps taken with short dwells therefore underestimate the steady temperature at every point, and the underestimate grows toward the coldest point.
Sign prediction
H3754. If a pulse tube's cold end warms when charge pressure is raised, the cause is regenerator pressure drop; if it cools, the cause is improved phasing. The sign of the response is diagnostic and costs one pressure change. Two charge pressures and one cold-end thermocouple are the whole experiment.
Impossibility
H3755. No regenerator can be optimal across two temperature decades, because matrix heat capacity and gas heat capacity vary differently with temperature, so a two-stage machine must accept a mismatch at one stage. The practical rule is to put the mismatch on the warm stage, where the exergy penalty is smaller.
Regime shift
H3756. There is a cold-end temperature below which helium's rising viscosity makes the orifice impedance dominate the phase shift, so an orifice setting that is optimal near 80 K is wrong near 20 K. A machine tuned at one temperature and run at another is an untuned machine.
Hidden variable
H3757. The vacuum in a cryocooler shroud is not the pressure its gauge reads, because the gauge sits at room temperature while the cold surfaces cryopump the residual gas. The effective pressure at the cold end can be orders of magnitude lower, which is why machines with apparently poor vacuum still reach base temperature.
Conservation audit
H3758. A cooler's electrical input minus the enthalpy leaving with the exhaust gas must equal the cold load plus every loss, and audits that omit the regenerator's enthalpy swing can produce efficiencies above the Carnot bound. Closing the balance needs no new hardware, only the exhaust's mass flow and temperature.

Vacuum, cryopumping and residual gas

Misattribution
H3759. The pressure rise in a sealed cryostat during warm-up is credited to wall desorption, but a comparable share can be the cold surfaces releasing what they cryopumped during cooldown. The two have different temperature signatures, so a warm-up with one surface held cold separates them.
Measurement artefact
H3760. Outgassing rates from the rate-of-rise method include the gauge's own outgassing and the chamber's virtual leaks, and at low rates the virtual leak dominates. The result is then a property of the fixture, and comparing two materials in it compares their difference, not their rates.
Bound
H3761. A cryopump's ultimate pressure is bounded by the vapour pressure of what it has captured at the coldest surface it offers, not by its pumping speed. A pump with enormous nitrogen speed can therefore be useless for hydrogen, and the single number on the specification sheet hides exactly that.
Saturation
H3762. A cryopanel's capture coefficient for a gas falls once the adsorbed layer becomes mobile, at a coverage of roughly a monolayer (estimate). That threshold is why pumping speed measured on a fresh panel does not hold on a loaded one, and why regeneration is scheduled by coverage rather than by time.
Inversion
H3763. Warming a cold surface slightly can improve pumping of one species while ruining another, because the species that matters is the one whose vapour pressure is nearest the surface temperature. Regeneration is therefore species-selective, and a schedule optimised for water will fail for hydrogen.
Scale argument
H3764. At conductance-limited pressures a cryopump's effective speed is set by its mouth's molecular conductance rather than by the pump, so enlarging the pump buys nothing until the mouth is enlarged. The ratio of the two is Knudsen-number dependent, which is why the improvement saturates instead of scaling.
Double correction
H3765. Thermal transpiration and the gauge's accommodation-coefficient correction are numerically the same factor for a monatomic gas, so a chamber that needs both corrections is being corrected twice. Applying one and checking against a gauge held at the chamber's own temperature decides which is real.
Crossover
H3766. Below a pressure the dominant residual gas changes from water to hydrogen, and the two have opposite temperature dependences of desorption, so a bake-out that helps in one decade can hurt in the next. The crossover is why bake schedules optimised at one base pressure fail at a lower one.
Selection effect
H3767. Published outgassing rates are a survivor population of well-cleaned coupons, so a design that assumes them will under-predict a real chamber's first pump-down by a large factor (estimate). The bias is not random and does not average out across a bill of materials.
Degeneracy
H3768. Leak rate and outgassing rate are degenerate in a rate-of-rise test, since both produce a linear rise, and only a helium spray or a repeated pump-down separates them. A test reporting a leak rate without either is reporting a sum with an assumed split.
Hysteresis claim
H3769. A getter's capacity appears larger on a second exposure because the first exposure passivated its surface, so the difference is activation rather than capacity. Re-activating identically before each exposure removes the apparent hysteresis, and usually removes the apparent capacity gain with it.
Unit trap
H3770. Conductance tabulated for air is routinely applied to hydrogen without the square-root-of-mass correction, a factor of about four (estimate). The error is silent because both numbers look plausible, and it always errs toward optimism about pumping speed. Converting one tabulated conductance by the mass ratio, then repeating the pump-down, settles it in an hour.
Sign prediction
H3771. If a chamber's pressure rises when a cold trap is filled, the trap is desorbing rather than pumping, and the sign of the response to filling is diagnostic on its own. Reading the sign costs one valve movement and prevents a week spent chasing a leak that is not there.
Negative claim
H3772. No elastomer-sealed system can reach a base pressure below the vapour pressure of its seals, taken at the warmest surface in the chamber, so the floor belongs to the elastomer rather than to the pump. A metal-sealed construction changes the floor by orders of magnitude, which is the only test the claim needs.

Cryogenic insulation and thermal links

Hidden term
H3773. Blanket heat leak is quoted per unit area, yet on a real dewar the seam and penetration area dominates, a small fraction of the area carrying a large fraction of the leak. The two cannot be separated by one boil-off test, only by a test with the penetrations blanked.
Reversal
H3774. Differential contraction opens layer spacing at a blanket's seams while closing it at its centre, so a blanket qualified at room temperature is qualified at the wrong spacing. The effect scales with seam length, so small coupons look better than tanks built from the same blanket.
Inversion of standard practice
H3775. A copper strap's conductance at cryogenic temperature is limited by its bolted end contacts rather than by its length, so adding cross-section in the middle changes nothing while lapping the ends changes everything. The rule that straps should be short is a statement about the wrong term.
Measurement artefact
H3776. Temperature differences blamed on insulation in a dewar are often conducted in along the sensor leads from the radiation shield, so the sensor reads the shield rather than the bottle. A differential pair routed together on the shield shows it, since both members then read the same wrong temperature.
Ratio invariant
H3777. A mechanical support's heat leak is set by the ratio of its strength to its conductivity, not by its strength, so the best support material is selected by that ratio. This is why titanium and glass-fibre composites appear in designs that also claim to prioritise stiffness.
Crossover
H3778. Above a certain neck diameter a dewar's conduction down the neck exceeds its radiation through the bore, so widening the neck to cut radiation eventually loses more than it gains. The crossover is computable from the neck's length and the bore's view factor, and it sits at a narrower neck than most designs assume.

Materials at low temperature

Artefact
H3779. Cryogenic embrittlement is attributed to a ductile-to-brittle transition, but at liquid-helium temperature part of a specimen's lost toughness can be pre-stress from grips contracting more than the gauge length. Testing the same material with compliant and rigid grips separates the material from the fixture.
Hidden term
H3780. Thermal expansion measured by dilatometry at cryogenic temperature includes the push-rod's own contraction, and below about 20 K the rod's term can exceed the sample's (estimate). The correction is not a constant but a function of the rod's history, so one room-temperature calibration does not remove it.
Bound
H3781. A metal's low-temperature conductivity is bounded by its residual resistivity, so two samples of one alloy with different impurity content differ by more than the temperature dependence suggests. The bound is the Wiedemann-Franz ratio at the residual value, which makes a purity specification a thermal specification.
Threshold
H3782. There is a strain amplitude below which a cryogenic fatigue crack does not propagate, because the crack tip cannot form a plastic zone at that temperature. The threshold follows the yield strength's temperature dependence rather than the stress intensity, so it moves when the alloy changes and not when the geometry does.
Inversion
H3783. Annealing copper at low temperature can raise its residual resistivity when the anneal mobilises dissolved gas into voids, which scatter more than the gas did in solution. The expectation that annealing always lowers resistivity assumes the gas is already out, and at cryogenic purity levels it often is not.
Identity claim
H3784. Kapitza resistance at a solid-helium boundary and acoustic mismatch at a bolted joint are one phonon transmission problem at two scales, so a joint's conductance and a coating's interface conductance should collapse onto one curve against the ratio of acoustic impedances. One cryostat and two specimen types would test the collapse.
Exponent misassignment
H3785. A metal's heat capacity is electron-dominated below a few kelvin and phonon-dominated above, and the crossover moves with impurity content, so a single power-law fit across it misassigns the Debye temperature. Fitting two terms and reporting the crossover as a fitted parameter turns a nuisance into a measurement.
Hysteresis claim
H3786. A composite's thermal expansion on cool-down differs from its expansion on warm-up because the matrix yields in one direction, so the loop area measures interfacial shear strength at temperature. That makes a dilatometer a mechanical tester for interfaces no other instrument can reach.
Finite-size effect
H3787. At specimen sizes below the phonon mean free path the measured conductivity becomes size-dependent, so a foil's conductivity at liquid-helium temperature is not the bulk value. The discrepancy is a boundary effect rather than a materials property, and reporting it as a property is why foil data scatter.
Conservation audit
H3788. A cryostat's heat balance must include the enthalpy boiled off to cool the current leads, and audits that debit only the static leak can report a heat load below the leads' own conduction. Closing the balance from the boil-off rate and the lead currents is a check that needs no new sensor.
Sign prediction
H3789. If a superconducting joint's resistance rises with field but falls with temperature, the cause is flux creep rather than a bad solder; a resistive joint would do the opposite in field. The pair of signs separates the mechanisms without cutting the joint open.
Impossibility
H3790. No structural material can combine the lowest thermal contraction with the highest strength at four kelvin, because both are set by the same atomic bonding and improve together only by accident of chemistry. Claims of both are measurements made at different temperatures, and the best supports are acknowledged compromises.

Frost, defrost and moisture

Surrogate error
H3791. Frost growth in a display case is credited to the air's humidity, but a comparable share can be warm moist air infiltrating through the door gasket, which the case's own sensor cannot see. A gasket-sealed run at the same internal humidity separates infiltration from condensation.
Cost misallocation
H3792. The defrost heater's energy is booked as the cost of defrost, while the reheating of the cabinet afterwards and the load imposed by meltwater left on the coil are larger terms that no meter on the heater sees. Counting the compressor's recovery work as part of the defrost cost changes which control strategy looks best.
Threshold
H3793. There is a frost density below which defrost is unnecessary because the layer's insulation is small, and above which it is mandatory because the fin gap closes; that threshold is set by the fin gap rather than by frost mass. Controllers that terminate on time are assuming a density, and the assumption fails in dry climates.
Design conflict
H3794. A hydrophobic coating can raise the energy cost of defrosting while reducing frost adhesion, because meltwater re-freezes as isolated droplets that resist drainage more than a continuous sheet does. The two objectives conflict, so a coating chosen on adhesion alone can make the appliance worse.
Geometric limit
H3795. At fin spacings below the frost thickness that blocks them, blockage is set by the spacing rather than by the humidity, so a tighter coil frosts at a lower absolute humidity. Geometry, not psychrometrics, decides when a case fails. Two coils of different fin spacing in one cabinet, at equal humidity, would rank the blockage by geometry alone.
Duration artefact
H3796. Coatings that look frost-free in a laboratory test usually show a longer induction time rather than a different equilibrium frost layer, so a short test measures the wrong quantity. Running coated and uncoated coils until blockage, rather than for a fixed duration, reverses the ranking about as often as it confirms it.
Lag
H3797. The cabinet's temperature response to a defrost cycle lags the heater by the evaporator's mass, so a controller terminating on air temperature terminates late by an amount that grows as the coil ices. The lateness is self-reinforcing, which is why iced coils get iced worse.
Count
H3798. A defrost controller has at most two independent observables available without adding sensors, the air-side pressure drop and a coil temperature, so a controller claiming three is inferring one of them. Making that inference explicit is what distinguishes a better algorithm from a better sensor.
Reversal
H3799. Raising the defrost termination temperature shortens the next frost cycle while lengthening the current defrost, and the sum of the two has a minimum. Controllers set for the shortest single defrost are usually running the longest total, which a week of logged cycles will show.
Bound
H3800. The rate at which a coil can be cleared is bounded by meltwater drainage rather than by heater power, so beyond a power the defrost time is set by the drain path's slope. Adding watts buys nothing once the water cannot leave, and the drain is the part nobody instruments.

Status and plan

One hundred entries, numbered in sequence. They were composed individually, with the logical form chosen before the subject, so that no two share an argument shape; the labels are printed for that reason and can be audited against the text.

The physics stays inside the cold chain: the compression cycle and where its losses are booked, the working fluids and the oil that travels with them, the evaporator as a distribution problem, the regenerative machines that reach temperatures no cycle can, the vacuum that insulates them, the materials that survive it, and the frost that undoes it. No claim here concerns a product, a manufacturer, or a person.

If one of these is worth an afternoon, it is the strap entry: a copper strap's conductance is usually credited to its length and cross-section, and the claim is that the bolted ends decide it, so lapping two contacts would change what twenty centimetres of copper cannot. It needs a cryostat, two straps, and a heat load, and it would either explain a persistent disagreement between calculated and measured heat leaks or retire it.

Incoprea Research · hand-considered hypotheses 3701–3800 · 3 November 2026 · ~/Desktop/Nov3/