SavvyAnalysis Puzzlers

The Wright Brothers were the first to fly on this date in 1903. Growing up in southwest Ohio, we didn’t give a lot of credence to those floating other versions of history. It was Orville and Wilbur at Kitty Hawk on December 17. Full stop. While we’re at it, let’s not forget their mechanic Charles Taylor who patiently got them back into the air time after time. Coming from the world of bicycles, the Wrights used a chain drive to turn their propellers. A few weeks back, I got engine data from an airplane powered by two Lycoming IO-390 engines and data from a Dynon Skyview. My rule about using data for a Puzzler is to identify the aircraft type and the engine type, but never use the N number. I’m sticking with that. The overspeed on this 390 isn’t a scenario that the typical 390 owner is likely to encounter, but in honor of first flight day I thought I would show you what happens when the chain breaks on a Wright B Flyer. The ticket title was “Severe Engine overspeed, dieseling, and smoke” so it sort of had Puzzler written on it from the get-go. Plus, I was […]

Is there a mirror image for Occam’s Razor? Seems like if we have a name for relying on the least complex explanation for a scenario, then we ought to have one for the most outlandish, unlikely, logic-defying explanation. For instance, recently we got engine data for a Cessna TR182 Turbo Skylane RG. The shop replaced a failed CHT probe and all of a sudden that cylinder’s CHTs were about 100º hotter. EGT’s hadn’t changed so the problem seemed to be cooling related. We recommended checking baffles and – although no exhaust maintenance had been performed – checking for an exhaust leak. After coming up empty checking all the logical explanations, somebody at the shop noticed that the new CHT probe in cylinder 3 didn’t look like its neighbors in cylinders 1 and 5. Installing the wrong probe gave us bad data and a pretty good goose chase. And something to consider next time we see that scenario. Let’s take a closer look at the data from the Turbo Skylane. The engine is a Lycoming O-540 and data is from a JPI 830 with a 1 sec sample rate. EGTs on top, with CHTs and FF below. Max CHT is 491 […]

When we started our engine data analysis service in 2014, I wondered if we would see seasonal patterns in the data. For instance, would we see more probe failures in the heat of the summer? Would we see more stuck valves in the colder months? Would we see more anomalies in general in the fall months after the busy flying season? Data supports that last one, which shouldn’t surprise anyone. But in ten years of working tickets I haven’t yet been able to connect the dots on other seasonal patterns. I’ll keep looking because my spidey sense tells me there’s something to find. Meanwhile, this month’s collection covers a lot of ground in terms of issues and powerplants. We’ll start with data from a Cirrus SR20 powered by a Continental IO-360 with data from an Avidyne Entegra with a 6 sec sample rate. EGTs on top, with CHTs and FF below. Nothing special about curser placement – just moving it out-of-the-way. After about 15 minutes of warming up, CHT 5 starts jumping between 500º and zero. After about 15 minutes of that, it settles down and tracks CHT normally until shutdown. Meanwhile, CHT3, which was tracking normally until takeoff, is […]

We rolled out Savvy’s Borescope Initiative in January. In the ensuing 7 1/2 months users have uploaded over 65,000 images to the repository. If you’re still waiting to participate, or if this is news to you, check out Savvy’s Borescope Initiative. Before the rollout, we suspected that most of the valves we looked at would show normal – symmetrical – heating patterns. We also suspected that we would catch some valves in various stages of burning that may not have been detected through compression checks or oscillations in EGT. We were right on both counts. A borescope image report prepared by an analyst will classify exhaust valve head images Normal, Early/mid-stage uneven heat pattern, or Advanced uneven heat pattern. We hope the following examples assist you in deciding what intervention, if any, is appropriate. If you are a SavvyMx client, your dedicated account manager can help you make this determination, and if you are a SavvyQA client, an account manager on our team will provide guidance. Normal heat pattern A normal heat pattern in the exhaust valve head deposits will show symmetrical rings of varying color, or possibly uniform color over the entire valve head. The colors are not so […]

I’m brushing up my French for an upcoming trip and was reminded that Mayday comes from M’Aidez – Help Me. When I started flying in the ’70s the conventional wisdom from the old timers was “whatever you do don’t declare an emergency”. Now that I’m an old timer, the prevailing attitude in my pilot circle is to declare the emergency and get whatever help you can – and you might be surprised at how little follow up you’ll encounter. Engine failure, structural failure, smoke in the cockpit; those are easy. Declare and get the help. What’s more challenging is when you see something troubling on the engine monitor – but you’re not sure if it’s a real problem or just a sensor acting up. Declaring entitles you to priority handling, and if you declared but it turned out to be nothing more than sensor failure, that would likely be the end of it -unless you had a history of crying wolf when there was no wolf. In your preparation for 91.103 compliance – a.k.a. “all the information about the upcoming flight” – you could review what scenarios would trigger an emergency declaration. It’s a tougher decision in the moment. We’ll […]

If an engine monitor probe or sensor is going to fail, would you rather have it just fail and be done with it, or have it be reliable some of the time and distracting some of the time? Neither has much appeal, but at least with the failure, you can skip the detective work of does it only happen in flight, or when the engine is warm, or when the tanks are full, or…? With intermittent data, it almost always makes sense to check the connection first. Sometimes the data can give you clues as to the odds of it simply being a case of a loose or dirty connection. Sometimes the data is accurately logging the intermittent failure of an engine component. For each case, a decision tree helps guide the process. In this puzzler we’ll analyze some examples and see what they can tell us. Here’s data from a Cirrus SR22 TN powered by a Continental IO – 550 turbo normalized engine and data from a Garmin G1,000 with a one second sample rate. Here are EGT, CHT and FF for the 35 minute flight. Something happens at the 10 minute mark, but I put the cursor on […]

Hockey playoffs are in full swing. It’s usually about this time of the year that I rewatch my favorite hockey movie “Miracle”, about the 1980 Olympics. So many great performances, including fellow pilot Kurt Russell as Herb Brooks, but none better than Eddie Cahill playing US goalie Jim Craig. And now you’re wondering, how is he going to tie hockey to engine data analysis? Here’s how. Sometimes analysis tells us exactly where to look to fix a problem. a clogged injector, a bad spark plug, a pinched cooling baffle – all pretty easy to spot in the data and gratifying to know that we helped save the owner time and money. That’s offense. Equally gratifying from an analysis point of view is when we keep an owner from scheduling and paying for unnecessary maintenance on something that isn’t broken. That’s defense. First up is a Piper Saratoga powered by a Lycoming IO – 540 and data from an Insight G4 with a 1 second sample rate. The flight is only about seven minutes starting at the 15 minute mark, but this includes ground ops before and after the flight. EGTs on top, CHTs and FF below. Cursor is on the […]

The analysis team sees lots of malfunctions in the data we review. No surprise – most clients request analysis when something’s not right, so we’re not seeing the thousands of flights that go off without a hitch. So you could say we’re spring-loaded for malfunctions. This month we had a surge in things that were broken – not just malfunctioning. Surely the engine data offered signals that something was about to fail. Either in real time, or in post flight review. Yeah but it’s not that simple – and as they say in the movies, don’t call me Shirley. This month, Savvy’s Director of Analysis Joe Godfrey looks at data from a Cessna 206, a Bearhawk, a Cirrus SR22T and a Beech Bonanza 36. First up is data from a Cessna 206 powered by a Lycoming IO-540 with data from a JPI 900 with a 1 sec sample rate. EGTs, CHTs and FF. There’s an early indication of trouble in EGT and CHT 3. EGT 3 is low until takeoff, then it’s up with the others. CHT 3 is also low until takeoff, then it jumps past the others and stays high – until it fails at about 2 hrs […]

Lately the news is filled with examples of deep fakes. This is nothing new for pilots. We practice partial panel approaches for scenarios where technology is trying to convince us that something phony is real. It’s nothing new for engine data analysts. We start the process with evaluating the reliability of the data. The difference is the perpetrator. A good deep fake usually has a bad actor or a comedian attached to it. We almost never see either in an airplane – it’s just bad data from a loose connection or a failing part. Or is it? This month Savvy’s Director of Analysis Joe Godfrey looks at data from a Cessna 182, a Beech Bonanza 36, a Cirrus SR22T and a Cessna 210. Let’s start with a Cessna 182 powered by a Continental O-470 engine and data from a JPI 900 with a 1 sec sample rate. Rather than take up space depicting CHTs and FF for each flight, let’s stipulate that both looked normal for all cylinders. The is EGTs only for four consecutive flights. There are a few things in this data that we could examine, but let’s stick with the green trace of EGT five. On the […]

When the analysis team started looking at engine data about 10 years ago, we could confidently recognize the patterns for clogged injectors, spark plug misfires and detonation events. Other patterns, like broken valve springs and induction leaks, came later. Last month the analysis team began reviewing borescope pictures. Just as with engine data, some conditions are easy to spot. And I’m confident that, as we gain experience will be able to identify more anomalies, and take advantage of the synergy of having pictures combined engine data. We’ll start this month with data from Cessna 210 powered by a Continental IO-520 and data from a Garmin 275 with a one second sample rate. This is a 30-minute segment of an hour-long flight – displaying only EGTs. All EGTs are a little noisy, but there’s a distinct and recurring pattern in the red trace, which is EGT 1. The original FEVA 1 algorithm scanned for this pattern. FEVA 2.1 uses a machine-learning algorithm designed to spot exhaust valve failures earlier in their demise. Both algorithms spotted this one. That green spot means this valve is too far gone for lapping compound to bring it back. Next is data from a Cessna 320 […]

Scottish poet Robert Burns wasn’t a pilot – but his words are good advice for flight planning. This Puzzler is about anomalies that happened in flight. 14 CFR 91.103 directs pilots to become familiar with all available information concerning a planned flight prior to departure. We plan and prepare for what we think and hope will unfold, then we use our training and experience to deal with what really happens. Captains Al Haynes and Sully come to mind. And Apollo 13. It’s a January tradition to restate the mission of these Puzzlers. Our goal is to help pilots recognize data anomalies in real time to make good decisions about the safety of the flight, and help owners to make informed maintenance decisions. This month we’ll look at data from a Cessna 182, A Commander 114 and a Cirrus SR22TN. How often do you review your engine data? If you had a runaway CHT event in flight are you sure you would have noticed it in real time? A CHT excursion would not necessarily lead to poor performance after the event. It might, but it might not. It might just be a warning sign for more serious event to come. That’s […]

I was channel surfing recently and landed on a Perry Mason rerun. First of all, it’s one of the great TV themes of all time. Hats off to Fred Steiner. Private Detective Paul Drake just oozes cool with his wild sport coats and convertible sports cars. Della Street was smart with just the right amount of allure. Poor DA Hamilton Burger – like Wile E. Coyote and the Washington Generals – he was destined to be outplayed. Over the course of the hour, Perry would often get foiled – sometimes by Lt. Tragg and sometimes just by events – but would eventually solve the case and get the murderer to admit guilt on the witness stand. When we’re analyzing engine data, most of the time we can spot the guilty party, and confidently make a recommendation for the appropriate maintenance. We know your time is valuable, and we know operating and maintaining an airplane isn’t cheap, so the goal is to only make recommendations that we are confident will solve the issue. Sometimes the analysis team will consult each other, or one or more Savvy account managers to try and get the Perry Mason moment. When we’re all stumped, our […]