In the video ice was forming faster when the engine was revving higher…
Seems to suggest that Carb heat ‘on’ when downwind/base and ‘off’ on final when the throttle is idle (pfl) should reduce the chance of carb ice.
The motorcycle carb example is particularly interesting because it is a CV (constant vacuum or constant velocity) carb, on which the slide rises automatically to maintain the same venturi effect regardless of rpm or throttle opening (i.e. regardless of power setting). Therefore I think what causes increased ice formation in that case is additional fuel flow through the jet. It may not be directly transferable to a fixed jet aircraft carb but none the less shows that fuel latent heat of vaporization is a significant factor. Rotax four strokes use CV motorcycle carbs so the situation is directly applicable.
On engines that are susceptible to carb ice (small Continentals, for instance) some people apply heat before closing the throttle completely, for instance when doing circuits they might do it when reducing power from 100%, directly after climb. The throttle body and fuel are thereby pre-warmed while hot exhaust gas is still available at the heat exchanger, before the throttle is closed for descent.
Regarding whether fuel evaporation were the whole story: if the engine were operating stochiometrically then the temperature drop due to icing would be identical whatever the throttle setting because you’d have the same ratio of fuel evaporated to air aspirated and therefore cooled.
Even if fuel evaporation causes the major component of the temperature drop, you might expect other factors e.g. the venturi effect to explain variation in susceptibility to icing at different throttle settings.
Or is it just that a partially closed butterfly valve is more susceptible to icing because of the smaller cross-sectional area of the airflow? Or perhaps because of turbulence causing supercooled water vapour to actually turn into ice?
edit…‘the temperature drop due to evaporation’ would be identical
‘the temperature drop due to evaporation’ would be identical
Yes, on the average. The latent heat per unit air mass flow is the same. But it seems to me the fuel would not be uniformly distributed into the incoming air initially, and the jet might see the coldest air prior to downstream mixing. In that case increased fuel evaporation would reduce the local surface temperature in the place where the ice forms and blocks fuel flow. Its clear in the video that the area around the jet is colder than other surfaces.
Perhaps the reason injected engines don’t have induction icing is simply because the fuel supply pressure is much, much higher. It would blow off any ice that might accumulate at the injector nozzle. Something to think about, I guess.
Just read CAA safety sense 14, which claims that icing is less common at high power settings primarily because the carburettor is likely to be hotter
b) Engines at reduced power
settings are more prone to icing
because engine induction
temperatures are lower. Also, the
partially closed butterfly can more
easily be restricted by the ice
build-up. This is a particular problem
if the engine is de-rated as in many
piston-engined helicopters and some
aeroplanes.
It also mentions carb surface roughness as a factor in icing.
In a venturi, the gas speeds up, so its temperature drops (straight thermodynamics). That is why carbs suffer icing. To a first order approximation, the icing isn’t anything to do with the presence of fuel. Without fuel you would still get ice buildup. The fuel evaporation just makes it happen a bit sooner.
The reason you get less carb icing at larger throttle openings is because the velocity increase is smaller so the temperature drop is smaller.
The reason why fuel servos (fuel injection) don’t normally suffer icing is because there is no venturi so the gas (the air) doesn’t cool down on the way through. You can still get icing on the fuel injection nozzles; right on their ends, but this is rare and is too small to cause a blockage of the entire port.
A fuel servo doesn’t use vacuum to meter the fuel/air ratio. A carb does that by sucking the fuel out of the chamber underneath, using the vacuum created by the venturi. A fuel servo does it by metering the air velocity and opening a valve for the fuel; this needs the fuel to be available under pressure otherwise it won’t come out.
At least that’s how I understand it…
I would never buy a plane with a carburettor. It is completely pointless and just a big extra risk. Loads and loads of people have been killed by carb icing, and the evidence is gone by the time they get to you. Even if you come out of it OK, you will never really know if the engine stoppage was due to carb icing, or due to some mysterious fuel system blockage which cleared itself, perhaps by the obstruction having moved elsewhere where it will block the flow another time.
Incidentally there is a famous carb icing diagram, all over the internet, which has the incorrect axes on it, but for years or decades nobody realised it and just kept reproducing it. It even featured in flying theory textbooks 
I would never buy a plane with a carburettor. It is completely pointless and just a big extra risk. Loads and loads of people have been killed by carb icing, and the evidence is gone by the time they get to you.
It highly depends on the installation. My carb never shows icing, not even in IMC. Other installations ice up on a CAVOK day during taxi. With the “I would never buy an airplane that has…” you end up building your own. Life is a series of compromises In my case, the only disadvantage I can see is not being able run LOP (only a bit) because of inaccurate fuel distribution. The 182 airframe did not have the ability to feed back fuel into the tanks (a requirement for fuel injection) and Cessna was in a big hurry when they brought the 182 RG to market. The restart 182 have that capability and come with injection engines but no more retractable gear and bad useful load.
did not have the ability to feed back fuel into the tanks (a requirement for fuel injection)
It isn’t. I don’t have any return to tanks.
All you need to avoid a tank return is a constant pressure fuel pump. It is systems that use a fuel pump without a pressure relief that need to return fuel.
In a venturi, the gas speeds up, so its temperature drops (straight thermodynamics). That is why carbs suffer icing. To a first order approximation, the icing isn’t anything to do with the presence of fuel. Without fuel you would still get ice buildup. The fuel evaporation just makes it happen a bit sooner.
You are right about venturi, but wrong about carb icing. The reason is the actual values you get. The pressure drop is way too small to have a large effect. The cooling due to fuel evaporation is an order of magnitude larger.
Throttle Body Injectors (TBI)s are much less susceptible to icing. The reason is supposedly that they have no butterfly valve for the ice to form.
