Clouds over mountains
11 min read

Structural ice is a known flight hazard and there are plenty of forecasting products to help a pilot avoid it. Failing that, there are volumes of text and hours of videos to advise the unlucky or the unwise on how to deliver themselves from its grip.

Curiously, there is another type of icing that has sent its share of airplanes to the salvage yard, and pilots to the graveyard. Because it is mainly an affliction of low-performance aircraft, it doesn’t receive as much attention.

Carb ice

Remember carb ice?

Many of us (I’d guess most) had our initial training in carburetor-equipped airplanes where the drill of applying carburetor heat before landing was strictly enforced. I did it mechanically with no conception of what would, or wouldn’t, happen if I ever stopped. We’d purchased a Cessna 150 before I began instruction and would be wedded to it for some time. The 150, being what it is, does not confer any claim to expertise on most aviation subjects. One exception is carburetor ice.

Conventional wisdom holds that certain engine designs are predisposed to carburetor ice. Notably the Continental O-200/300s and the older O-470 series, while Lycoming engines are resistant to icing because the induction piping efficiently transfers engine heat to the carburetor. I can’t dispute that but icing susceptibility also depends on the engine installation in a particular airframe and it can vary considerably between airplanes, even of the same model.

Sooner or Later

Our 150 had a Continental O-200-A and its Marvel carburetor would pick up ice under a cold stare. My first encounter with it was while practicing a hold on a clear, cool morning over a large body of water. I’d pulled the throttle back to 2200 RPM while circling but every couple of minutes the RPMs would slowly drop off and I’d need to add a touch of throttle. I tightened the throttle tensioner until my hand hurt but it still seemed to be creeping back. Then, and with no warning, the engine went silent. “This must be carb ice,” I thought and reached for the carburetor heat knob.

Sure enough, the engine came back for one second and then quit outright. In a panic, I tried to undo everything but before I could remove the heat, the slug of melted ice passed through the engine and full power was restored. I learned two things.

First, if there is a substantial amount of ice present, the engine will go into distress when carb heat is applied. Just wait it out, power is coming. Second, there is a fine emotional line between the smug pride someone feels handling a situation which is well understood, and the panic that sets in when it isn’t. The first lesson never needed reinforcing. The second seems to require periodic re-exposures – like a tetanus booster.

Fool Me Twice

Working toward a private certificate, I cannot remember anyone telling me that carburetor heat might not always resolve an icing problem. That it might have limitations. During my instrument training, one instructor mentioned in passing that he had an issue flying a Piper Apache (not a stellar performer in ideal conditions) where he’d been unable to remove, or even arrest, the ice building in both carburetors. He’d applied full heat, adjusted mixtures, manifold pressures, RPMs. No combination of anything helped. I asked what he did then. He answered, “I got scared.”

I took what he said to heart, but in the succeeding years I logged a fair amount of time flying the 150 in visible moisture without any ice control issues. We had a routine in IMC. Every one or two minutes I would pull the heat control, whereupon the RPMs would momentarily drop and then recover. Engine roughness and RPM decrease would determine how often the heat needed to be applied. Sometimes I just had to leave it on, lean the mixture, and live with the power deficit – undesirable in such a low-powered airplane.

Flat land clouds

Carburetor ice in the Low stratus over California’s central valley is easily managed – with some exceptions.

Much of that flying was done at low altitude in the stratus layer which often forms in California’s central valley. Those conditions were fairly benign but, eventually, my flights extended into more varied geography and weather. Two of those stand out. The first caught me by surprise while crossing the Sierras. The flight originated in Truckee, California, with temperatures in the mid-30s (F) and melting snow on the taxiway. I kept the carburetor heat on while taxiing until advancing the throttle for takeoff.

Climbing east-bound at full throttle over Interstate 80, the engine wouldn’t maintain power. Full heat, and jockeying the mixture, preserved what power remained but it would not clear out the accumulated ice. Over the next 15 minutes, the highway relentlessly out-climbed the airplane, raising the specter of a forced landing on that steep, twisting concrete ribbon. A few miles before reaching Donner Summit, the ice began to clear.

Entering drier air from the valley rescued my situation with a scant 500 feet between the landing gear and the pavement. Using mental sleight of hand, I convinced myself the experience was an aberration unique to mountain operations. I clung to that conceit for a couple of years until it was stripped away during an instrument flight across the Sacramento Valley.

This hop was at lower altitude in somewhat warmer temperatures but it was heading toward a deep low pressure system moving in from the Pacific. The FSS briefer assured me structural icing would not be a factor but gusting winds would.

With a clearance in hand, I launched from Auburn and contacted the approach facility at Stockton. As soon as we leveled off at 4000 feet, there was a perceptible decay in power. So the heat came on and the throttle went forward to the stop. It may have been the higher moisture content in that warmer air, or convergence around the low, or something else, but power continued to drop (although more slowly).

I asked Stockton for lower, but sector MVAs and conflicting traffic wouldn’t allow it. Intense wind held down the ground speed, further delaying any relief. Finally, negotiating 3000 got the 150 below the clouds and gained a little manifold pressure. But the engine was just holding its own. Meanwhile, twilight fell – along with the dew point spread and visibility – as a one-hour flight morphed into two. The next controller expedited a VOR approach into Tracy.

After leaving the approach fix, I suffered for a couple of minutes in genuine anxiety knowing that the missed approach procedure called for an initial climb which was, by now, beyond the capability of the airplane, and conditions for maintaining power were rapidly deteriorating.

That night the low deepened and passed directly over the airport, ripping an airplane from its moorings behind our 150 and flipping it on its back.

Cessna 150

The 150 is not known for its thrilling performance.

In order to gain the full utility from a low-performance airplane, you will need to explore some outlying regions of the operating envelope. These exercises can be a source of gratification but they also demand judgment. That was my last instrument flight in the C-150.

You’re kidding

Most carb ice encounters don’t induce sweaty palms. In fact, some deliver a dollop of humor.

Heading down the northern California coast, just inland of the coastal mountains on a pleasant spring afternoon, the RPMs began decreasing. I was a little surprised to see it at this temperature, but I realized an onshore flow was bringing moist Pacific air up over the hills. Giving the carburetor heat knob a good tug, I watched in dismay as the knob, the actuating wire and part of the sheath pulled out of the panel and fell into my lap. Santa Rosa airport was 10 miles ahead. I told the tower I’d like to make a precautionary landing straight in. The controller asked the nature of my problem and I did my best do describe what happened. Then he asked if I’d like them to dispatch the emergency equipment (I could hear laughter in the background). I demurred.

With the cowling removed, you could see where the actuator wire had pulled out of the B-nut securing it to the air box flapper. That was easy to fix. I also saw that the air box flapper automatically fell to the “full heat” position when it was disconnected – a smart fail-safe design. If I’d understood the system better, I could have proceeded to my destination and spared us all the trouble.

All you need

Even on a beautiful warm day, in dry air, it is possible to build up ice in a carburetor. All you need is a willfully negligent pilot.

The airplane had been sitting in the Bay Area for several rainy days when I picked it up at San Carlos. A faulty gas cap gasket had admitted an impressive amount of water into the fuel tanks and I spent half an hour rocking the wings and draining the sumps until I was positively convinced no water remained.

Yet the carb heat check showed a steep drop on the tachometer before returning to its normal drop of 80 RPM. It couldn’t be carb ice. The dew point spread was too wide. Since my ride had departed, and planes were stacking up behind me, I decided to figure this out back home and accepted a takeoff clearance.

The east departure from San Carlos keeps you down around a thousand feet while transiting the Bay under San Francisco’s Class B airspace. It’s a poor place to deal with an engine problem – even an impossible one – so I kept heat on the carburetor. Clearing the Sunol Pass, with more altitude, and in very dry air, I removed the heat and promptly experienced the same ice-like symptoms.

On the ground, the sumps were drained again with no hint of contamination. The air box, carburetor and linkages all seemed to be in order. I quit in frustration and went home. A couple of days later, during a routine preflight, I discovered the gascolator was nearly full of water. In my preoccupation with the fuel tanks, I’d forgotten to drain it at San Carlos. In response, the gascolator had been sending a steady trickle of water into the carburetor. Like a companion quietly reminding me to pay attention.

Pretty much

Clouds over mountains

The valleys paralleling California’s northern coastline contain favorable elements for carb icing (moisture, lifting, convergence) year round, but particularly during the winter months.

During initial training, one of my excellent CFIs, Mike Walling, was giving me tailwheel instruction in a Super Cub. Mike suggested I forgo the constant pulling and pushing of the carb heat control since, “Lycoming engines pretty much don’t ice up under power.” He was right. But pretty much, when used as a tool for aeronautical decision making, has its limitations. Twenty years later, I found a notable exception to Mike’s rule. By then I was flying behind a Lycoming 0-360 with a constant speed propeller and a carburetor air temperature gauge in the panel, which pretty much eliminated the possibility of icing. It was a routine IMC flight into Santa Rosa (a hard luck airport for my carburetors).

The assigned altitude was right at the freezing level, but the airframe was clean and there were no PIREPs. During cruise, the carb heat control was set to keep the temperature needle just out of the yellow arc (5 degrees C). A substantial portion of the traffic at STS consists of corporate jets and airliners so ATC expected my approach to be flown at speed with power on.

Descending out of 800 feet, the runway came into sight. I closed the throttle and decelerated through the final mile to touchdown – at which point the propeller stopped cold. Coasting onto a turnout I notified the Tower I’d experienced an engine stoppage – there was no way to tell when it actually occurred. It took less than a minute for the ice to melt out of the carburetor. The Lycoming started without complaint.

So what happened on this flight? Five degrees worked fine on countless similar approaches. There is no way to know for certain. The temperature indication in the carburetor throat was probably accurate, but there was a severe ice accumulation in another, unmonitored part of the carburetor.

So what is a safe carburetor air temperature on approach? I don’t know. While a carb temperature gauge is an invaluable tool for setting partial heat during the cruise phase of flight, applying full carb heat may be a better bet during the approach phase. One thing I do know: If the throttle had been closed a little earlier, I might’ve been writing this to the NTSB.

Finally, yes, a fuel-injected engine would put an end to these difficulties, but it would not satisfy the cost/performance relationship that governs my flying.

Kim Hunter
Latest posts by Kim Hunter (see all)
26 replies
  1. Anthony A
    Anthony A says:

    Great write-up. There’s a fair bit of mystery to carb heat because most of the time we don’t need it. I am glad to have learned from your analysis, thanks for writing.

  2. Steve Green
    Steve Green says:

    An outstanding discussion of carb icing! You’ve really covered a lot of ground with your examples and descriptive details. Should be required reading for new students, if there are any still using carburetors, that is.

    But perhaps the greatest pearls of wisdom in this article are the paragraphs describing the fine emotional line between smug pride and panic, and the paragraph discussing exploring outlying regions of the operating envelope, and when that exploration has gone far enough. Those points are enormously important and they are, sadly, not always well taught in many flight training environments.

  3. Fred O.
    Fred O. says:

    Nice work. Thanks. I agree with your suggestion that ice is possible at unexpected times. The Lycoming O-360 in my Mooney M20C will make carburetor ice at full power on climb out if the conditions are right.

  4. Ken S.
    Ken S. says:

    I have seen Lyc. engines get carb ice on initial climb out (before the oil warms) at temps up to 45f. Once the oil warms, they are resistant to carb ice in most cases.
    I used to fly aTwin Beech that had the same basic carbs as the 150 you mention. It would get carb ice if we within 10 miles of a cloud.

  5. Alfred Ross Thompson IV
    Alfred Ross Thompson IV says:

    Great read! Currently a student pilot flying a Cessna 150 with the o-200-a. My first solo cross country was 168 NM which went without a hitch until the final 78 NM leg home. That’s the first time I experienced carburetor ice. It was a scary moment because the engine almost quit twice! I ended up leave the carb heat on and diverting to a closer airport.

  6. Matt Thurber
    Matt Thurber says:

    My worst carb icing incident: CAVU 70-degree day, Piper Warrior powered by a Lycoming O-320, engine quit completely without warning at 2,000 ft msl. Be prepared, and look over this handy probability chart.,-a-,-proficiency/accident-analysis/accident-statistics/epilot-reports/epilot-asf-accident-reports-carburetor-what/nyc02fa025_1.jpg?la=en&hash=67C5994A58F45C3B9DA137FD038BDAA8805B2DCD

  7. Arlen Burns
    Arlen Burns says:

    My Cherokee 160 made carb ice @ 2400feet just below a cloud layer. Apparently that’s a high humidity area, frequently used for vfr flights.

  8. Jim
    Jim says:

    I was taught to never leave the carb heat on. There are circumstances when it is real cold that adding carb heat will bring the temperature up to where it will allow for the build up of ice. Then you are in real trouble.
    That advice has served me well in my 15000 hrs flying single engines here in Alaska.

  9. Suresh Kumar Bista
    Suresh Kumar Bista says:

    I began flight training in Victoria B. C. Canada. British Columbia is a great place to learn flying. Great places to see from the air; beautiful mountains, forests, lakes, towns and coastal Pacific. But amidst the excitement, many airplanes are lost with carb icing.
    This is once procedure that all my instructors repeated every time. “Don’t forget the carb heat”. I did that and followed their advise very diligently and here I am still flying today.

  10. Dave
    Dave says:

    Interesting stuff. I fly a Robinson helicopter with automatic carb heat assist, but the assist can be deactivated at the pilots discretion. Robinson also uses an automatic throttle management system. The quirk is that if the carb builds ice, the auto-throttle keeps adding power, so it can easily go unnoticed. The good thing is that a carb heat gauge is standard, and we are taught to monitor and manage the carb heat continuously. 4000+ hours in Robinsons, so I guess it works pretty well

  11. Robert Thomas
    Robert Thomas says:

    I’ve never understood why heat is not applied to the carb body and intake piping continuously. Applying carb heat in the form of heated air works, obviously, but it does reduce power. Maybe a better solution would be to heat the carb and intake piping? Ice cannot form on a surface that is above freezing. The velocity on the intake air is fairly high, so little heat would transfer to the air/fuel mixture, so the loss of power should be miniscule. Heating could be by way of engine oil, or electrically.

  12. Mike Crognale
    Mike Crognale says:

    Crossing Lake Erie in a Cherokee 180, I think, IFR when I noticed the engine rpm decreasing with roughness. Grabbed the carb heat and it cleared right up. Then I noticed that the airspeed was decreasing and, yup, the pitot heat was off. Lessons learned that day.

  13. Scott
    Scott says:

    Last week while cruising at 7000’ over WV’s mountains we picked up a load of carb ice while in VMC with 30F OAT. With full carb heat applied, and varied mixture/throttle settings the 0-470 was unexpectedly smooth running while we were not able to maintain altitude. This makes for confusion and an unsettled flight deck. Checking mags & switching tanks didn’t help. We declared an emergency and chose the only flat area(smallish field) and prepared for a forced landing. After losing hundreds of feet ample power was restored and we diverted to the 2nd nearest airport in an agonizing 15 mile slog. We left the airplane and drove the final 5-1/2 hrs to our destination. My wife took an airliner home while I flew the 182 checking carb heat every 5 minutes with no trouble. FYI Carb ice remains an issue today – it’s not just in training stories years ago.

  14. Gary Miller
    Gary Miller says:

    The Lycoming O-360 in my Mooney M20C is also a carb ice maker. I have a carb temp sensor and have found a 50 deg F setting prevents carb ice and also provides more equal EGT distribution and a smoother running engine. One must not forget to adjust the mixture as carb heat will change the fuel/air mixture.

  15. Kim Hunter
    Kim Hunter says:


    Thank you for making an excellent point.

    I haven’t encountered the low temperatures you describe during my own operating experience, but they are common in Alaska, Canada and the upper tier US. Your comment emphasizes the importance of local knowledge and experience – no one has all the answers.

  16. Dan Fregin
    Dan Fregin says:

    I had a student take a C-150 for some cross-country time from Red Bluff to Klamath Falls and back. Upon return he told me all was good on the way up but coming back (1000 feet higher) he lost a bit of power which carb heat brought back when used as trained. By the time he got back to RBL a few clouds had formed just to the East of Mt. Shasta.

    While giving a BFR flight I would say, “This is a simulation; what do you think it is?” as I eased the power back a little. If I got a wrong answer, I eased it back a little bit more. If the answer was, “Carb ice” and the pilot applied carb heat, I would reduce power to simulate the engine sucking water. I was amazed at how many pilots would cancel the carb heat before the “ice melted” and I brought the power back up. I did this to all my students as well.

    If induction system icing falls into this area, my former business partner was flying a Twin Bonanza from Portland back to Red Bluff in cloud when both engines lost nearly all power and the plane lost nearly all altitude. He worked the engine controls around in a way that produced some backfiring and managed to limp over a ridge into Roseburg. An AD came out shortly requiring alternate air be opened before entry into visible moisture, or something like that.

  17. ScottK
    ScottK says:

    I fly a Scout with a Lycoming O-360. It never got carb-ice until we rebuilt the engine and put new baffles in the cowling. For the next couple of years it routinely would develop carb-ice, until the baffles wore and again it rarely develops carb-ice.

  18. Tim
    Tim says:

    Great discussion on carb ice. As a retire military pilot, just about all my time has been in turbine powered aircraft. So getting back into general aviation aircraft with recips after over 30 years has its learning challenges. Thanks for the reminder that an “old salt” may have much to learn or re-learn..

  19. David Ward Sandidge
    David Ward Sandidge says:

    Every Private Pilot ground school should have an evening with the mechanic – with the maintenance technician being a guest speaker. He/she would lecture on the mechanical aspects of operating and maintaining an airplane, which would, of course, cover how airplane carburetors work and how carburetor heat works its magic. Ideally, the PP course would, at some point, end with a lengthy field trip to a maintenance shop.

  20. Ian
    Ian says:

    I wish there was a safe and reliable way to intentionally encounter carb ice during primary flight training. The build up is sneaky, the first symptoms can go unnoticed, and next thing you know you’re wondering whether “this is it.” I’ve only encountered it once (thankfully during my PFT with an instructor) and it scared the hell out of me.

  21. Kim Hunter
    Kim Hunter says:


    Unlike military and ab initio airline training, a GA pilots’ learning is largely self-provided. While most of us benefited from excellent instructors, those who stick with flying gain much of their skill on the job.

    Carburetor icing is a classic example. After my first exposure I went through about 50 hours of wondering, worrying and testing for induction ice. I probably pulled the carb heat knob hundreds of times on suspicion. Most of the time it showed me no ice was present – and occasionally that it was. Over those hours accumulated knowledge displaced most of my anxiety. I have every confidence your experience will be similar.

    Good luck.

  22. Bobm
    Bobm says:

    If in cruise, monitor the egt. It will gradually decrease due to ice blocking the airflow & resulting in a richer mixture. It will be noticeable prior to any rpm drop.

  23. Barry Gloger
    Barry Gloger says:

    “Finally, yes, a fuel-injected engine would put an end to these difficulties, but it would not satisfy the cost/performance relationship that governs my flying.”

    Reasonable Safety always trumps my cost/performance relationship. In today’s environment carb ice should be a non event because all aircraft should be flying with fuel injection, which ultimately pays for itself with increased fuel efficiency and resale value.

  24. Kim Hunter
    Kim Hunter says:

    Well stated Barry.

    In my case it’s not just the cost of the engine I’d have to consider, but also the cost of the airplane it comes with.

  25. Richard Wyeroski
    Richard Wyeroski says:

    While waiting for take off in a 172 Cessna with a 0-300 continental, the engine began to run erratically. The temperature was 50 degrees. Visible moisture was present. I applied power and the engine quit and we could not restart it!


Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *