Technically the title is The Christmas Song but when a musician calls it on a gig it’s Chestnuts — as in roasting on an open fire. It’s a good lyric full of holiday imagery and it’s a great tune by a guy better known as a singer, arranger and entertainer than a songwriter — Mel Torme. The first two notes are an octave apart so right away the song tells you it’s different. You hear it a lot this time of year which also makes it the other kind of chestnut…a familiar story, tale or joke. I took over this monthly column in late 2014 so by my math this is Puzzler #112. Figure each one averages three cases and that’s a lot of chestnuts.
Early on it was mostly clogged injectors, weak spark, cooling issues and failed probes. Then we started seeing patterns for broken valve springs, induction leaks, detonation and burned valves. We’ve had data from some partial and a few catastrophic engine failures. We’re now two years into our borescope initiative and that opened up a new area of intrigue. At Oshkosh somebody asked me how I choose the cases for each Puzzler. Obviously it starts with having compelling engine data or borescope pictures. Then it’s a concentrated effort to include singles and twins, turbos and air-breathers, certified and experimentals, high performance and trainers. Some months that’s easier than others – some months it feels like all the interesting stuff came from SR22s or 182s. Then it’s a mix of anomalies – it can’t be all clogged injectors or failed probes. So yeah building a Puzzler is sort of a puzzle.
First up is a Glassair III powered by a Ly-Con 540 EXP and data from a Garmin G3X touch with a 1 second sample rate. Here’s a baseline flight with EGTs, CHTs and FF. I used the V tool to mark the highest value for each depicted trace. The cursor is on the high mark for most of the CHTs – although the high mark for CHT 2 was a little earlier during the climb and CHT 4 peaks just before 01:50.
Let’s start with CHT’s because that’s where the biggest problem is. Look at the first value for CHT 3. It’s lower than all the others. It looks normal during taxi, then goes low and stays there during the flight, then joins the others again after landing. In cruise it’s running about 130º cooler than its neighbor CHT 1 and 220º cooler than its neighbor CHT 5. Something’s fishy. CHT 4 has some noise from takeoff until about 01:05, then a little more around 01:35. The EGT traces and the FF trace are showing a steady small oscillation from the FF drop at 00:55 until about 01:30. I didn’t depict Bus Volts but it’s steady as a rock at 28.3 V. The takeaway from this baseline flight is that there’s a big cooling problem and some credibility issues with some of the data. Here’s the next flight a few days later.
I put the cursor at takeoff so the data block didn’t cover up the event at the 00:32 minute mark. CHT 5 peaks at 440º. It has the shape of a detonation event, but we’re not seeing the sudden drop in EGT 5 that often accompanies detonation – we think that’s because this event wasn’t the result of a cracked spark plug insulator. It’s the result of continued poor cooling reducing the detonation margin, and the power change at the 27 minute mark seems to have been the trigger. It didn’t like that FF setting and the bump at the 30 minute mark came too late. Here’s what they found when they pulled the cylinder.
Next is data from a Lancair EX-P powered by a Continental TSIO-550 with data from a Garmin G1000 with a one second sample rate. EGTs including TIT 1 and 2, CHTs and FF.
I put the cursor on the spot where FF shows 0. There’s one other spot earlier around the 11 minute mark where it drops to 0, then it happens again just before 02:20. It’s pretty clear that the engine’s still running so we know it’s not 0. It wasn’t showing up and down spikes indicating a loose connection before it dropped off. It always makes sense to check the connections first and especially with suspicious behavior like this.
There’s an ongoing jumpiness in CHT 1 that happens in flight – not seeing it as much much before takeoff and during approach. Now that we know that’s there let’s zoom into the earlier part of the flight for a closer look at other parameters.
CHT 2 has two small excursions that look like connection issues. The first lines up with the end of the runup and the second happens just before takeoff power is applied – maybe as the plane taxied into position. After that CHT 2 behaves.
The wavy pattern in CHT 1 looks like a loose cooling baffle. We see waviness for a while in CHT 3, too. There’s also a steady rise in CHT 1 peaking at 30 minutes and again we’re not seeing corroboration in EGT 1. This doesn’t have that detonation event look about it – looks like a cylinder that got hot for a while.
The elephant here is the pattern in TIT 1 and 2 that begins as EGTs rise and FF drops – starting at about 22 minutes. Altitude is not depicted but it helps to know that the pattern begins as the plane climbs through FL110 on its way to FL190. That “a-b-c” pattern – about 40 seconds at “a”, then about 25 seconds at “b” then about 15 seconds at”c” – is too regular to attribute to a loose connection. And both TITs are showing the pattern. But it’s not showing in any of the EGTs – which is surprising given the connection of EGTs to TITs.
You might wonder if the pattern shows up in other parameters like oil pressure or MAP. Let’s take a look.
So the answer is yes for oil pressure and no for MAP. What about fuel pressure? Let’s take a look.
So fuel pressure is a hot mess for a while then settles down then hot mess again then settles down. Hot mess occurred during climbs and settle downs occurred during level-offs and eventually cruise altitude. The owner mentioned “I noted in the aircraft pre-buy inspection that both fuel vent tubes had kinks and recommended replacement. I will do that tomorrow” A few days later he posted “I replaced both wingtip fuel vent hoses and the problem resolved. Test flew up to 14,500’ and all parameters held steady.”
Truth be told, wingtip fuel vent tubes were not even on our bingo card when we first saw these TIT fluctuations. Sometimes the dots are hard to connect. And the TIT fluctuations are still there even when fuel pressure settles down. But we’re glad that the owner can log a win. And now vent tubes are on our bingo cards.
We’ll close out 2025 with an exhaust valve on a Piper Comanche powered by a Lycoming O-540 with data from an Insight G2 with a 2 second sample rate. Here are EGTs, CHTs and FF for a recent flight.
The engine data didn’t flag an oscillating pattern in EGT 3. The shop found the problem during borescope inspection as part of the annual.
It’s a small spot so still a great candidate for lapping. Presumably as the spot grew it would have showed up in engine data but better to catch it now.