Carb Ice

Carb Ice
By Grace Ellen



Grace Ellen is a freelance writer and pilot. She was a newspaper journalist for 10 years but left the profession upon finally conceding that it was morally bankrupt. A three-day open cockpit trip home from AirVenture 2001 in N1777W crystallized her love of aviation and formal training followed in a 1943 Taylorcraft L2-M. Tutored by CFIIs Chuck Nelson and Ken Terry, she also received aerobatic training. Grace Ellen earned her IAC Smooth Patch flying Ken’s Christen Eagle. A member of local IAC and EAA chapters, she is close to completing her instrument and commercial ratings. MCM

William Wynne was the passenger in his Corvair powered Pietenpol when it crashed north of Tampa, Florida, on July 14, 2001. The engine cut off at 700′ AGL, which gave only 60 seconds to attempt a restart and execute a forced landing. To avoid people on the ground, the pilot tried a sharp bank and the plane spun in from 80′. The impact destroyed the airframe.

When freeing the trapped pilot, a fire started and ignited William’s fuel-soaked clothes. While extensively burned, both William and the pilot survived the accident. An investigation into the engine stoppage has indicated that carb ice  almost certainly was the cause. Because a few of the engine components were incinerated in the fire, no one will ever be able to say with 100% certainty that carb ice was the cause. The engine was recovered from a wrecking yard where it sat for months, placed on a test stand, and runs well. The remaining wreckage was examined very closely and no evidence of any kind of failure was found.

This mechanical integrity and clues such as visible condensation on the exposed intake manifold just prior to the engine quitting leaves carb ice as the only probable cause fitting not most, but all of the evidence, William said.

A single incident of carb ice hardly seems newsworthy. Most pilots and flight instructors feel they have an adequate knowledge of the subject. However, William’s extensive discussions with pilots and builders after the accident revealed that most traditionally trained pilots have little understanding of the subject. The rote memorization of a technique that works on a single aircraft and atmospheric setting does not constitute an adequate understanding of the topic. Further, experimental aircraft may have quite different engine management requirements.

The day of the accident was overcast with the cloud base at 800′ and dropping. The OAT was about 70F. The dew point was within 5 degrees. The trip had been 90 miles and the flight was within 2 miles of the destination. The plane had been throttled back to 60mph to allow traffic at the destination to clear. The carb heat control was in the rear cockpit and was not applied.

The engine, a direct drive Corvair fed by a Stromberg carb from a C-85, was  turning about 2,200rpm. It had at least 13 gallons of 100LL in the wing tank. Within one minute of the power reduction, the engine quit, William said.  It had dual ignition and a restart was attempted on each one. The engine  cranked normally, but did not light. Carb heat was not used. The impact broke  up the airframe and severed the 3/8″ fuel lines to the wing tank.

The  airframe ignited about a minute later when fumes from the spilling lines  reached a shorting wire. The fire burned for more than 20 minutes, consuming  most of the airframe but leaving the engine largely intact.

Many pilots interviewed later expressed the following thoughts:
1) Carb ice cannot form at 70F, and certainly not in Florida.
2) It would take longer than a minute for ice to block off the 34mm venturi.
3) The engine would “run rough” for a while before quitting.
4) Ice could not form at 2,200rpm.
5) Auto fuel would have about the same potential to ice as 100LL.

All five of the above thoughts are wrong. If you believe any of them,
you are a giant step closer to having your own version of William’s accident.
Here’s why:

1) Ice can form on warm days. Anytime a gas expands from high pressure to  low it will consume energy from its environment. In this case, the gas is the  air the engine is consuming and the pressure drop is from ambient to manifold  pressure, about 30″map to 12″map. The energy it consumes is ancarbicey form of     available heat. Most of the heat comes out of the air. This temperature drop  is instantaneous and can easily be more than 40F. Shoot a thermometer with a  CO2 extinguisher and learn.




2) Carb ice forms at the pressure drop point, which occurs at the
restriction to flow. At a reduced power setting, the throttle plate is the
restriction, not the venturi. At 2,200rpm and 12″map, the throttle plate is
barely cracked open. Right at this crack is the idle fuel port, a tiny hole.
A minute film of ice could cover it in an instant. The engine will stop
running because at power settings like these, most of its fuel comes from the idle circuit.

3) A certified four-cylinder engine of 7 to 1 compression and 190cid
swinging a 25-pound metal prop generally will sputter for a while when it is
experiencing carb ice. By contrast, a 9.5 to 1 six-cylinder 164cid engine
with a 6-pound prop will quit nearly outright. Many Lycoming pilots said
their engines gave warning. Lycoming carbs are bolted to the oil sump and
experience the onset of icing at a slower rate.

4) Icing has nothing to do with rpm; it results from the pressure drop.
Granted, an A-65 Cub with a certified prop is very unlikely to ice at
2,200rpm, but this is because 2,200rpm usually is associated with nearly open
throttle on this plane, and consequently very little pressure drop in the
carb.  However, any motor experiencing a large pressure drop in the carb is prone to ice, regardless of rpm. The motor in N1777W had a static rpm of 2,650. Any motor that has a prop that will allow a similarly high static rpm will be running low map at an rpm like 2,200. A manifold pressure gauge provides
useful information that a tachometer by itself does not.

5) Although auto fuel was not being used the day of the accident, pilots
need to understand how it can contribute to carb icing. Remember that at a
reduced power setting, the restriction is the throttle plate. And when
operating at reduced power, there is a large pressure drop at the plate, with
its accompanying temperature decrease. Fuel flows out of the idle port in a
mist. Misted fuel is still a liquid, not a vapor. 100LL under these
conditions remains a mist until reaching the combustion chamber. Contrast
this with auto fuel, which by design will vaporize readily under these
circumstances. It is a fact of physics that when the fuel changes from a
liquid mist to a gaseous vapor, it takes further heat from the surrounding
air. This is the cooling one feels when gas evaporates off the skin. This
additional temperature drop can produce icing when the same engine under
identical circumstances would not ice with 100LL.

In the past five years, N1777W flew with more than a dozen pilots and  logged hundreds of hours that included several very long trips. William  briefed pilots who flew the plane to use carb heat before any substantial  power reduction. The carb heat system on the plane was so effective that it  produced a 200rpm drop at idle, but still had to be employed as anti-ice, not
de-ice. The pilot involved in the accident only had about two hours in the  plane, including the final flight. Click here to view the NTSB accident report.

Although much of the physics of icing can be found in textbooks and  technical publications, William’s observations on the subject are based upon  years of work and actual flight testing. It is the nature of some to debate  anything and offer opinions extracted out of context from technical  publications. But William adamantly believes that this is a safety of flight  issue and people without flight testing experience debating esoteric details  dilutes the risk management message. William deems any commentary that  hinders the delivery of the message as amoral.

To reduce the possibility of a similar accident, William suggests that  potential pilots get a better understanding of icing, more thorough  briefings, a panel placard about when to use carb heat, and carb air  temperature gauges. William is working on a combined throttle/carb heat lever
which would move in concert and have calibrated linkages, but could be  manipulated separately for run ups. All of these are small efforts when the  price can be the destruction of aircraft and the loss of life.

4 Replies to “Carb Ice”

  1. Always an excellent, but painful article…glad to hear Grace’s name…haven’t seen or heard anything about her for some time…hope she is doing well and keeping the Christmas lights going 24/7. {;^)

  2. Thanks William,
    I’ll be flying my GN1 behind a C75 when I’m finished. I’ll be sure to do my homework on carb icing when the time comes. Please continue sharing your insights…

  3. Carb ice can be quite sneaky. My first experience with it was in a Cessna 150M running 2700rpm at 7500′ MSL over New York City. OAT about 50F (around 75 on the ground). Clear, with a broken base around 4000′. Didn’t run rough, just noticed a gradual decrease in engine rpm. Didn’t even register in my mind that it could be carb ice because the sky was so clear. I started checking my sectional and VORs to verify my position (didn’t have GPS at the time). Then the light bulb above my head lit up, and I pulled the carb heat. Nothing happened for several seconds, then a little stumble, then all smooth. Glad it was that easy to learn that lesson!

  4. AMEN!!

    The only thing that I’d add is to make sure that you don’t pull the mixture off instead of the carb heat. I was getting checked out in an unfamiliar plane in the pattern at night when I did that and the cooling fan up front went quiet. The instructor immediately checked the controls and pushed the mixture in.

    I will and never have piloted another plane without thoroughly familiarizing myself with the controls.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.