In general, frosting has generally a negative effect on cooling. How much depends on the air flow in the freezer, as cold air has to reach every part of the contents easily.
Water vapour deposited from the air leads to frost
formation and has a great influence on heat transfer with
the following positive factors.
The latent heat (sublimation of water vapour)
increases the heat transfer coefficient at the initial stages
of frosting. The influence of the latent heat is included by
introduction of the 'sensible heat factor', i.e. the ratio
between the total transferred and the sensible heat
amount. With this factor it is possible to calculate by
approximation the nominal cooling capacity from the
'dry' cooling capacity. Manufacturers often correct with
Q0o = 1.25*Qdry
The heat transfer is further improved by the fact that
the ice crystals in the beginning lead to an enlargement
and more roughness of the surface.
These advantages however are soon followed by the
- The poor heat conduction coefficient of the growing
frost layer decreases the heat transfer. The evaporation
temperature goes down or must be kept constant by a
larger cooling surface.
The frost formation reduces the air flow rate and
enlarges the pressure drop. The air flow rate decreases fast
or slowly, depending on the fan characteristic. This can
result in too low an exit temperature and an undesired
increase of the non-homogeneity of the air condition in
the refrigerated area.
C. H. M. Machielsen and H. G. Kerschbaumer, Influence of frost formation and defrosting on the performance of
air coolers: standards and dimensionless coefficients for the
system designer. Rev. Int. Froid 1989 Vol 12 Septembre 283
The article further explains that there's an optimal frequency at which automated defrosting should happen: on one hand one would like to maximize the benefits of defrosting, but on the other automatic defrosting has a cost, because it slightly heats up the contents to remove the frost.
While the authors refer to automatic defrosting, manual defrosting has even higher costs in terms of increasing temperature, because of course it means bringing the unit to room temperature.
So, to answer your question:
- Does frosting decrease efficiency significantly? Yes, and the negative effect increases with time.
- Is regular defrosting a good practice? Yes.
- Is it possible to make a general statement that even 1mm of frost generates so-and-so costs? No, as the article shows, specific measurements are necessary to determine the exact effects of frost in a particular unit, and to determine when to apply defrosting in order to minimize cost.
My personal take on the article is that defrosting once a year in winter is a good thing, but that the claimed efficiency losses are dubious.