Pelican's Perch #59:
A cylinder in your piston aircraft engine flunks its compression check, with lots of leakage past the exhaust valve. The mechanic says you probably fried the valve by leaning too aggressively. Wrong, says AVweb's John Deakin! Lean mixtures don't cause burned valves — lousy valve-to-seat geometry does. It's probably the fault of the factory or overhaul shop, not the pilot.
The term "Fried Valves" seems to be sneaking into the lexicon of engine terminology, and is most often used by those who speak in dire tones about LOP (Lean of Peak) operations. Next they'll blame LOP for tail flutter, vacuum pump failures, faded upholstery, and other such things. While were at it, let's blame LOP for the ozone holes, global warming, and the next ice age, too. It makes about as much sense.
I'd like to convince you that EGT, ROP, LOP, octane, and all the other "usual suspects" have little or nothing to do with valve temperature, valve recession, valve failure, or valve anything. To find out what DOES affect valves, read on.
A Little History, If I May?
There has been a rash of messages lately about LOP, and whether it will cure problems, prevent them, or cause them. For example:
"… IMHO, the jury is still out as far as LOP being the solution to the top end problems with these engines. LOP certainly saves fuel and may result in lower operating costs…"
Who the heck ever said LOP would cure problems built in by the factory?
Let's review some recent history.
During the past 15 years or so, quite a number of people (including me) have come to the following conclusions:
- Engines built by TCM prior to about 1991 usually ran to (and beyond)
full TBO without much work on the cylinders.
- Engines (and cylinders) built during or after the 1991 strike at TCM have consistently suffered from seriously excessive premature cylinder problems, with very few making more than about 500 hours.
In 1998 or so, TCM even quietly acknowledged that there was a problem, and they were investigating. They later stated there were no problems, but at the same time, announced some changes in the way they were making cylinders. There is some evidence those changes may have helped with cylinder barrel wear, but the problems with exhaust valves and guides continue to plague owners.
The important thing here is that these issues have NOTHING to do with lean of peak operations! The whole subject of LOP operations never came up until about 1998, when it was discovered that truly balancing the fuel/air ratios across all cylinders would, for the first time, permit most of the fuel-injected engines used across the general aviation fleet to even get there, and continue to run smoothly!
In other words, these engines were suffering from grossly premature top-end (cylinder) problems long before 1998 ... LONG before GAMIjectors were even a dream!
If you remember back before the Internet, when the only way you could read Trade-A-Plane was three times a month on yellow paper, you should remember the hundreds of advertisements for airplanes with "1200 hours since new, 600 hours STOH" and the like. That meant there were a LOT of engines that were getting premature top-end work long before TBO. That was unacceptable then, and it's worse now. Some sharp observers noted that cylinders built after about 1991 almost never went beyond about 600 or 800 hours without requiring that famous "top overhaul."
I am aware of one brand-new 1993 Bonanza that had to have new cylinders at 200 hours, and had the engine replaced at 500. That particular owner had flown his previous V-35 Bonanza to TBO without problems. Anecdotal? Sure. Typical, too!
I've been a regular on CompuServe's AVSIG (the oldest forum on CompuServe) for more than 20 years, and this subject has been debated and discussed there endlessly all that time. (AVSIG is still going strong, by the way, and remains the finest aviation resource I know.) When the Internet came along, this same discussion became universal on all forums and mail-lists where GA people are found.
I repeat, all this was happening long before GAMIjectors were a gleam in George Braly's eye, and long before LOP operations were seriously considered as a routine method of operation on any of these "flat" engines. (The Piper Malibu is a special case.)
It is beginning to occur to me that many have misinterpreted the purpose of LOP operations, and may have the perception that we believe LOP to be a cure-all for the factory errors. That's just not true!
I'm not sure, but perhaps those of us who were "early adopters" of LOP operations may have been a bit too strident in "selling" LOP. Some now seem to think we suggest nothing but LOP, and that ROP is always "bad." That's also not true!
At first, we probably figured that everyone knew all about ROP, and all we had to do was "fill out the other half of the story." But we're now finding out that a lot of pilots merely want to set one fixed power setting, and don't want to understand what it's doing to their engine, ROP or LOP. The truth is that BOTH LOP and ROP have their proper place. The two different methods are each useful tools in any pilot's bag of tricks to maximize the utility of his aircraft across a broad spectrum of operations. I have resolved to emphasize both types of operation in the future.
Even the factories have modified their stance! Lycoming tech reps used to scream (literally), "I wouldn't recommend lean of peak to my worst enemy!" Now, they are saying, "Well, yeah, it works, but pilots are too stupid to do it." Well, I guess that's an improvement.
There is also the crowd that isn't going to run LOP no matter what, and for them, we need to teach them how to run ROP a little better (usually a lot richer). Or maybe we can't teach them anything, and that's fine, too.
Okay, Now — Back to Those Valves ...
Here's what an aircraft engine valve looks like:
It is a finely machined part, and looks like a jewel when finished. Tolerances are very tight. The valve stem must be just the right size (when hot!) to just slide smoothly in the "valve guide," itself another finely machined part. Some are of solid high-temperature steel alloy (most TCM engines), others are hollow with liquid sodium inside (most Lycomings) to spread valve heat away from the head and to the valve stem.
(An old field test on sodium-filled valves was to drain a sample of the engine oil, and check it on the spot for the presence of sodium. If found, it would indicate valve damage. I haven't thought of that in decades!)
There are many variations, but the valves in most aircraft engines are pretty much the same, differing only in minor details. There is usually one intake valve that opens every other turn of the crankshaft to let the good stuff in, and one exhaust valve that opens every other turn to let the bad stuff out. This is the classic four-stroke "Otto Cycle."
Picture an upright cylinder, just after the exhaust stroke:
- Intake valve opens just before TDC (Top Dead Center), exhaust valve
closes shortly after TDC (yes, both are open for a brief time, and this is
called "valve overlap"), piston falls away, sucking in the fuel
and air (good stuff);
- Intake valve closes, piston comes up, compressing the mixture, spark
fires before TDC, combustion starts, reaches peak pressure after TDC;
- Combustion event (both valves closed) drives piston down, turning
crankshaft (can you spell "Rube Goldberg"?);
- Exhaust valve opens, piston comes up, pushing the "bad stuff" out.
The cycle repeats, 20 times per second or more at high RPM (more than 40 crankshaft turns per second), endlessly. Well, maybe not endlessly, but it must seem that way to the poor valve, with somewhere between one hundred and two hundred million cycles to TBO!
Even at idle RPM, and even though valves actuate only once in two engine revolutions, the mechanisms are a blur in action. All valves are held closed with very strong springs, and get pushed closed even harder by the pressure of combustion. They are opened by means of a "pushrod" driven by cam rings in radial engines or by camshafts in "flat" engines. The valve head must be at the top of the cylinder, so the valve stem projects out and away from the engine. The usual method of choice for actuation is to have a "rocker arm" with one end of the rocker on the end of the valve stem, and the other on a "pushrod" that rides on the cam ring or camshaft. The cam rings and camshafts are geared so that a "bump" or "lobe" comes up under that pushrod every other crankshaft rotation.
Even though the valves open and close in split seconds, the slope of the cam lobe "gradually" opens and closes the valve, reducing the impact forces transmitted through the linkage. The slope on the back side of the cam also lets the valve close "gently." "Gently" is a relative term here — at least it's more gentle than letting that valve snap closed under full spring pressure!
The interface between the valve face and the valve "seat" that is pressed into the cylinder head is absolutely critical, but not for the reasons you might expect. It's true that this metal-to-metal interface needs to make a good seal to contain the 800-1000 PSI combustion event, but that can be done with a very slim point of contact. The really critical purpose of the metal-to-metal interface is for cooling the valve face, which gets pretty hot.
For more on this, see the following two links to John Schwaner's Sacramento Sky Ranch web site:
How Valves Keep Their Cool
All valve heating comes from the combustion event, which takes place right at the valve face, inside the combustion chamber. The valve face is subjected to a momentary blast of 3,000°F to 4,000°F when combustion is taking place. As the piston drops away, the pressure (and with it, the temperature) falls dramatically. Once the combustion event is over, the valve opens, and burned gasses (at much lower temperatures) exit past that valve-to-seat interface, carrying some heat away.
Somehow, the heat in that valve face must be removed. There are only two paths for that heat to take: (1) via the rim of the valve face to the valve seat, and (2) via the valve stem to the valve guide. During the time the valve is closed (about 75% of the time), most of the heat is conducted from the hot valve face into the much cooler valve seat, then to the still cooler cylinder head and cooling fins. Once there, it is carried away by airflow (or liquid cooling on engines so equipped).
Some heat conducts along the valve stem, and if the fit in the valve guide is true and correct, that will also be conducted to the valve guide, to any oil bathing the area, to the cylinder head, the cooling fins, and away.
Estimates vary, but normally, about 75% of the remaining heat in the valve is conducted away by the valve seat, and about 25% by the valve guide. (Sodium-filled valves are a little different — maybe 65%/ 35% or 60%/40% on the split.) Now, the valve itself doesn't go from 4,000°F to 400°F degrees in an instant, and back again. The flash of the combustion event during the power stroke will heat the valve up a little, then the metal-to-metal contact will cool it a lot, and the process repeats, 20 or more times per second. The valve itself will stabilize at some intermediate temperature.
It's important to remember here that the crankshaft makes two turns for each combustion event (in one cylinder). That means the valve is closed (and cooling) for a bit less than 540° and open for only a bit more than 180°.
It should be intuitively obvious then, that the valve temperature will correspond most closely with the cylinder head temperature (not the EGT), and indeed, old data from Lycoming (1966) and the old manuals from the big radials confirm this.
Note that CHT, valve guide, and valve head temperature all increase together, all peak at roughly the same point on the mixture curve (just rich of peak EGT), and all fall together. It stands to reason then, that if your CHT is too hot, then less heat will be carried away from the valve. It's a double-whammy, because you're probably making more heat in the combustion chamber, which makes the valve hotter, and the hotter cylinder isn't able to accept more heat.
EGT plays a part in all this, of course, but contrary to nearly universal belief, a pretty minor part.
The key elements in good valve cooling are:
- Good valve face to valve seat contact (it needs to be nearly perfect); and
- Good valve stem to valve guide fit; and
- Cool cylinder head temperatures.
All these VASTLY outweigh the effect of EGT.
Like many parts of these engines, the miracle is not that it runs so well, but that it runs at all! But run they do, for millions of cycles.
If you operate any of the "big flat sixes," you probably see EGTs of around 1400ºF to 1550º F in normal cruise. Are you worried that 1550ºF may be "frying" your valves? Well, on the same engines, usually with the same valves, but with turbos and lower compression ratios, you'd see TITs (Turbine Intake Temperatures) in the 1600º to 1650º F range in cruise, and up to 1850º F at peak TIT at very high power settings!
With that in mind, if you run a normally aspirated engine (no turbo), your concern over EGT as an indicator of valve temperatures should be like Alfred E. Neuman's, "What, me worry?" It IS possible to "fry" a valve in a normally aspirated engine, but EGT is NOT the indication you're looking for. Pay attention to the CHT, and the valves take care of themselves.com
Fact is, those same part number valves in the high-powered turbocharged engines shed heat to the cylinder heads just like your normally aspirated engine does. They wouldn't survive if valve temperatures were determined by EGT.
What DOES determine exhaust valve temperatures are the three factors I've mentioned above. Good contact at the valve seat and the valve stem, and less heat in the first place.
What Makes That High Heat?
Glad you asked!
First and foremost, the heat comes from high combustion pressures. Think of an engine where the designer foolishly had the peak pressure occurring EXACTLY AT TDC (Top Dead Center). The pressure would build up before TDC, peak at TDC, and then the pressure would decrease after TDC. Of course, this engine wouldn't even run, but bear with me?
Can you see that the pressure (and thus the temperature) would be enormous at TDC, right where the combustion chamber is at its smallest? Can you see that NO usable power would be transmitted to the crankshaft (and prop)? Pretty silly design, right?
Now, can you imagine the heat that would produce in the combustion chamber, piston, cylinder walls, cylinder heads, spark plugs — and valves?
Now, let's use our imagination, and move the peak pressure out to around 15º after TDC. Can you see that the pressure would be a LOT less, and so would the temperature? It just so happens that the theoretically ideal point for the "peak pressure pulse" (PPP) to occur is right around 15º past TDC. At that point, the temperatures in all components have dropped off dramatically.
So, what are the factors that contribute to high peak combustion pressures in some of our typical engines? Take a look at this matrix:
Factors Contributing to High Peak Combustion Pressures|
|Factor||Normally Aspirated Engines||Turbocharged Engines||Comments|
|1.||High power||260 to 300 hp||285 to 350 hp|
|2.||Advanced spark timing||22º to 25º BTDC||20º BTDC||Some turbocharged TCM engines use 24º BTDC, which is probably a bad idea.|
|3.||High compression ratio||8.5:1 (TCM)
|4.||Fuel/air ratios that are in the "danger zone" from just LOP out to 125ºF or more ROP||Same effect||Same effect||NOTE: Setting the EGT at 25º to 50ºF ROP guarantees the very hottest exhaust valve temperatures possible!|
|5.||High induction air temperature||Not a factor||May be a factor, especially if no intercooler||This is not a major factor, but worth noting|
|6.||Lower-octane fuel||May be a factor if you don’t play by the rules!||Heaven forbid!|
Note the turbocharged engines have employed engineering design parameters (like lower compression ratios and retarded spark timing) that are designed to LOWER peak cylinder pressures compared to the normally aspirated engines. Among many reasons this must be done is to keep the exhaust valves cool!
Bottom line, from data from the test stand: Given two engines, one normally aspirated and one turbocharged, both running at the same cylinder head temperature and the same horsepower, the valve temperatures will also be about the same, while the EGTs are about 1600º F in the turbocharged engine, and 1450º F in the normally aspirated engine.
So, "Where's the beef?" What can cause problems with these valves? They run just fine throughout a wide range of temperatures and power settings. As long as that finely ground rim on the valve face plants itself squarely on the matching valve seat and the metal-to-metal interface is wide enough, and the valve stem rides smoothly against the valve guide, the valves won't give any trouble, even if you abuse the engine. (Yes, there ARE limits, but they're not critical for VALVES.)
How the engine is operated, whether ROP or LOP, high power or not, is far less a factor than simple CHT. Manage CHT properly (including well-installed and maintained baffling), and your valves will be just fine. If there is some pilot error that affects valves, there will be other damage as well, giving the clues.
Who's to Blame for Fried Valves?
Flat statement: I believe that virtually all valve problems originate with the factory or the overhaul shop.
The hole through the cylinder that takes the valve guide must be true, straight and centered. The valve guide must be true and straight. Finally, the valve rim must match precisely the face of the valve seat, and both mating surfaces must be wide enough to provide enough surface area to conduct the heat.
This calls for some very fine machine work, and sadly, the factories haven't done it very well. The main hole will always be microscopically off-center, and it will never be perfectly straight. Close doesn't count here. The guide must also be machined or honed to very tight tolerances, and it must be straight and true.
It appears that for many years, TCM has simply drilled the "big holes" in the cylinder head, stuffed pre-reamed valve guides in, and then installed the valves. When done this way, it's very unlikely you'll end up with a nice straight valve guide. The cylinder is heated, the valve guide is chilled with liquid nitrogen, and pressed into place. When the temperatures stabilize, the guides are very tightly gripped, and some distortion is inevitable. The results are highly unpredictable, unless you predict poor results, and early top-end work!
Doing it this way takes less care and is less expensive than doing it right, at least for the manufacturer or the overhauler. It may even work. It may work to 500 hours, or it may even go to TBO. But, it probably won't. When everything heats up to normal operating temperatures, there may be the slightest bit of abnormal slop in the fit of the valve to the seat or the valve stem to the valve guide. That may tear up the guide or valve stem. One part of the stem may be hotter than another, causing a subtle warping. ANY imperfection in the valve-to-valve-guide contact will reduce the amount of heat conducted away from the valve face. If you don't get the heat away from the valve face, it gets hot. Now the trouble begins.
A FAR better way to do this is to take all the above steps, but install a guide that is too small to accept the valve stem. Allow the guide to absorb any of the forces placed upon it, and accept the small, inevitable distortions. Once installed, THEN ream it out to the exact size needed for the valve stem. This "post-reaming" technique will produce that straight and true hole. Consistent reports from visitors to the TCM factory, and comments from TCM, reveal that they are NOT "post-reaming." Some have reported TCM people as saying, "Yes, we're about to start post-reaming," but to date, I've seen no evidence that they have.
A recent Aviation Consumer article reflected this in its findings of some really sloppy fits for the TCM cylinders that they compared to the "Millennium" cylinders made by Superior Air Parts. I'd guess that some bright bean counter at TCM figured out a way to save a step or two.
This is probably the biggest single reason these engines suffer so many premature valve problems, today.
But that's only part of the story. Once the valve is installed nice and true, there remains the task of making the correct metal-to-metal contact between the rim of the valve face and the valve seat (which is itself yet another insert that needs to be placed with great care).
The usual way of doing this is to put a fine grinding compound on the surfaces, stick the valve in, and spin it, or rotate it back and forth, so that the two surfaces grind away at each other, hitting the high spots on both surfaces, eventually leaving a perfect match. This is called "lapping," and it's an evil chore. As John Schwaner points out, if the person doing that is tired, or ready to go home, or just doesn't care, it's awfully tempting to grind away until it looks reasonable, then just stuff the valve in, install the springs and keepers, and go home. Having once been a line boy drafted to do this chore in my "spare time," I can relate to that comment! I always figured that quiet moments on the line were my reward for working hard when things were busy, but Clyde Jones, founder of Jones Aviation in Sarasota, Fla., my boss at the time, didn't like paying seventy-five cents an hour for a line boy doing nothing. Whenever he could catch me, I got drafted for other duties. Like lapping valves, and sandpapering parking meters by the thousands. (Now THERE's a story for another column!)
Not all the engine overhaulers "lap" valves anymore. Monty Barrett of Barrett Performance Aircraft (one of the few excellent engine builders) in Tulsa, Okla., feels that the abrasive compound gets embedded in the matching surfaces at the molecular level, and may eventually cause problems. He prefers a much more modern system for grinding valves and seats — I think it's called a "Serdi" system — that does not require hand work or abrasive compounds.
But, I digress, as usual.
Ideally, that area of contact has to be some minimum width, and it must be equal all the way around, or the valve head will be cooled unevenly. Uneven cooling will cause hot spots and cold spots, and the valve head will actually warp a bit. When that happens the contact is not even all the way around, and a microscopic gap opens. At first, the pressure of the combustion event is probably enough to smash the valve head closed and correct a small warp, but eventually, the gap will allow a tiny amount of the combustion gases to leak past. Once this begins, it's only a matter of time. This is where you begin to see a loss in compression, sometimes very rapidly!
A Case in Point ...
Here's a jug removed for that very reason.
Remember, that combustion event can be upwards of 4,000ºF and that heat blowing through a small crack will cause an intense hot spot on the rim of the valve. With the rest of the valve getting cooling, and this little arc not getting any, the result looks like this:
For those folks who didn't have a misspent youth working on aircraft engines, the small valve is the exhaust valve, and the larger one is the intake valve.
Oh, by the way, these pictures are of the SAME cylinder, and this cylinder was never operated LOP — only at 50º to 100º ROP, just like the factory recommends. Just a few engine-hours before this picture was taken, the cylinder passed a compression test during the annual inspection. Within that few hours, the static compression dropped to almost nothing.
When I first saw these pictures, I assumed the discoloration at the six o'clock position was the "fried spot." But engine expert Monty Barrett took one look at the picture and instantly identified that spot as "normal," and said that the true hot spot was clearly (to Monty) the portion centered around the 10:30 position. Without knowing anything but what was in the picture, he stated unequivocally that it MUST be repaired before further flight. George Braly still has the jug and valve, and confirms that, AND confirmed that's the location of the "sliver of light" in the picture from the exhaust port side. I learned something, so it's been a good day.
Once a valve has gone this far, lots of things can happen, all bad, and they're fairly unpredictable. Pieces of the valve can break off, or the whole head can break off.
That's called, in the vernacular, "Swallering a valve."
Now, the factories and the tech reps, wanting to deny any and all claims, may tell you that you ran your engine "too hot," and "fried that valve." Well, maybe running "too hot" will hasten the demise of an improperly installed valve, but I cannot bring myself to believe that running the engine continuously at even elevated cylinder head temperatures will cause more than modestly accelerated wear and failure problems with PROPERLY INSTALLED valves, guides, and seats.
I really get a little testy with the factory suggesting pilots are the cause of "fried valves" when I think of the factory redline limit for CHT, usually 460º F, or 475º F, sometimes even 500º F, and we who are trying to teach "a better way" are saying LOUD AND CLEAR:
"Treat 400F as a max continuous redline!"
Just who is kidding whom, here?
We even take that a step further, and suggest 380º F as a nice "target," so we'll never exceed that 400º F.
I'm also hearing rumblings that TCM may start suggesting that a little bit of leakage past the exhaust valve is acceptable. Now, I'm not a lawyer, and I don't play one on TV, but if I were, I think I'd be quietly suggesting to TCM, "I really don't think you ought to do that."
I hope I've convinced you that EGT, ROP, LOP, octane, and all the other "usual suspects" have little or nothing to do with valve temperatures, valve recession, valve failure, or valve anything. It's NOT how the pilot operates the engine that really affects VALVE health, it's how the engine builder put your engine together.
That said, you CAN "improve" your chances by keeping your CHTs well under 400º F during ALL phases of flight. You can do that LOP, or you can do it ROP, the choice is yours.
If I haven't convinced you, at least I hope you're thinking about it. Above all, DON'T let anyone blame valve problems on you!
Be careful up there!