lml@lancaironline.net Mailing List Archive

<<<<<<<<<<<<<<<<--->>>>>>>>>>>>>>>> << Lancair Builders' Mail List >> <<<<<<<<<<<<<<<<--->>>>>>>>>>>>>>>> A couple of years ago I had a talk with the RAM folks (who do the work with TSIO-520's in twin Cessnas) at OSH, and the fellow made the passing comment that measurements of combustion pressures in these engines showed that even in "steady state" operation the pressures recorded in an individual cylinder varied substantially from cycle to cycle.<< This is called "cycle to cycle" variability. Cylinder fires off and makes 30 Hp one time and 25 the next and 32 the next cycle. Not good. If two cylinders in direct firing sequence happen to fire the first jug at 25 HP and the next jug at 32 Hp, your significant other is going to jump and say something like, "Honey! WHAT WAS THAT! Don't do THAT, ANYMORE!!!" His observation is correct. However, that variation is, for example, much larger in the typical Lycoming than the typical TCM engine, in my experience. Further, different issues affect that variability. One of the capabilities we have on our engine test stand is to measure, calculate and display in a trend graph the real time statistical analysis that gives us a "coefficient of variability" of the peaks of the combustion pressure events, and the coef of variability of Theta_P-P. It is VERY revealing of a lot of issues and how to make it better. He asserted that one could balance an engine and prop as much as you like, but these pressure variations would give rise to a vibration harshness that could never be controlled by balancing alone.<< Well... let's just say that that there is "less than good", "good" and "better" but it will never be perfect! I kept his comments in mind when I started reading some of the new engine text books and learning a bit about electronic fuel injection and the ability to more carefully control the mixture into each cylinder. I also note with interest the careful work done on intake manifolds in modern automotive engines. << This contrasts dramatically from the induction system of my previous plane, a TR-182 with a carbureted, turbocharged Lycoming O-540 engine. << THAT is an example of one of the "less than good" systems I mentioned above!!! This beast used the normal Lycoming intake manifold (buried in part in the oil sump) with six intake tubes radiating up to the individual cylinders. I had it fully instrumented, and fought boredom on long flights fiddling with throttle butterfly position, manifold pressure, RPM, mixture and anything else I could think of while recording pages and pages of data which I would later plot. << Want a job in our test stand? You sound like someone who is a glutton for that kind of punishment! <g> The EGT spread was simply awful under all conditions, and there was not much you could do about it. If you leaned it to "peak" (which peak?) as recommended in the book, it burned some exhaust valves after about 800 hours. Burning more fuel saved exhaust valves,<< I suspect, if you look at the data carefully, that when you had he engine at Peak TIT, that some jugs were operating at 50 to 100F ROP and others at 50 LOP. And yes, I would expect, you would burn some valves by following Lycoming's misguided recommendation in that way. MOST LIKELY THE VALVES YOU BURNED WERE THOSE OPERATING AT 50 TO 100F RICH OF PEAK, AND NOT THOSE JUGS OPERATING LEAN OF PEAK. The Continental 470 and 520 engines use a log manifold design under the cylinders as opposed to the Malibu 520-BE and the 550 engines which have a top intake manifold that has been more carefully designed to equalize air flow and mixture.<< Well... we see some problems with that design. In our experience, the long runner-updraft TCM induction system does a better job of getting the same amount of air to each cylinder than does the top intake spider system (which more closely resembles the Lycoming system.... BTW...) Long induction pipes promote uniform air distribution that is not affected by RPM "tuning". Good thing, that. The later engines are reputedly capable of MUCH smoother operation ...<< I don't think that is correct. It is certainly inconsistent with all of our experience. I fly TCM runner-log branch engines every day that are smoother than any top down induction system I have ever flown. ... suggesting the truth of assertion that mixture uniformity contributes to more equal combustion events cycle to cycle, and thus more smoothness.<< True assertion. You need several things to get these engines to operate smoothly, most importantly: 1) Uniform AIR distribution across a reasonable range of RPMs; 2) Uniform FUEL distribution; 3) Reliable-uniform ignition without cross firing, etc... 4) Balanced rotating components. 5) Sufficient total fuel/air charge of sufficiently uniform distribution so that the natural statistical variation in burn times is minimized. So here is my current hypothesis which I would like you to comment upon. In a primitive induction system (such as the Lycoming I used to fly or the 470/520 bottom induction Continentals) designed just to get air to the cylinders at minimum cost and complexity, the result is dynamic behavior with pressure waves ricocheting from end to end in a fairly chaotic fashion thus causing different amounts of air to arrive at a particular intake valve, varying from cycle to cycle. Thus, even with carefully balanced fuel injection that squirts precisely the same amount of fuel each cycle, there will be a difference in mixture strength cycle to cycle because the air delivered to the cylinder varies in a semi-random fashion. Consequently, combustion speed, peak pressures, total power delivered, etc. all vary from cycle to cycle, and from cylinder to cylinder in a semi-chaotic way varying over time as well as cylinder to cylinder.<< As to the Lycoming, YES, for the reasons you state. As to the 470/520 bottom induction TCM engines, an emphatic, NO, for the reasons I noted. Come fly with me in "bottom induction" TCM Engine. In fact, I have comparison data showing the statistical coefficients of variability of the combustion events from the Lycomings to the TCM engines. The LYCOMING TIO-540 data is pretty bad, by comparison. In my mind, this differs substantially from the nice, regular, standing wave behavior one would see from a modern auto engine with electronic fuel injectors and carefully designed intake manifolds with equal length runners, carefully designed inlets, and other careful attention to detail. I gaze upon these new plastic intake manifolds with their aerodynamic sophistication with some admiration.<< A lot of that automotive smoothness is due to the V-8 / V-6 configuration. Their combustion events are not significantly more uniform that a well set up TCM engine (bottom induction). The electronic fuel injection systems don't help much at all at any RPM above about 2000 RPM, and introduce enormous complexity to the aircraft system that requires no single point failure modes. One should be very very careful about adopting automotive technology and importing it whole sale into the aircraft engine world. That is, largely, what TCM has done with FADEC. In my view, there are some very significant issues there. So, my question (at last.....) - What kind of cycle to cycle variations do you see in your engine testing? I presume that your optical instrumentation permits you to get fairly high quality traces of cylinder pressure versus time (or rotation) and that with enough data storage in your data acquisition system, you can see how an individual cylinder behaves through a series of cycles. (Then again, this IS asking a lot of a measurement system.) << Yes. We do all of that (and more), cylinder by cylinder, in real time. It is the routine/norm on our test stand. I will email you directly some large graphical data files, when I get a chance, later. Can you view a power point file of several graphs if I email it? Wish I could post it up to the list. If so, what do you see in terms of variation cycle to cycle? How does this relate to intake manifold design? What other factors affect these variations?<< Rich mixtures promote more uniformity. Higher power promotes more uniformity. Dual sparks promote more uniformity. Uniform F/A ratios promote more uniformity. TCM runner log branch/updraft induction systems, with our GAMIjectors, do the best overall job of any induction system we have seen flying. Lycoming TIO-540 upside down engines in some configurations, are the worst we have seen. And do these variations show up in the form of engine smoothness (or lack of it) as perceived by the pilot?<< YES... VERY MUCH SO. However... BIG however: It is surprising, that once you get the system above a certain level of good uniformity, any further improvements are hard to perceive. There is sort of a "threshold". On one side of it, the engine is perceived to operate "smoothly" across a broad range of variations in cycle to cycle variability, and, if you end up on the wrong side of the threshold, the engine is "rough", regardless. There is some "middle" ground, but it is surprisingly small in range. Related: would you care to comment on the famous "Lycoming knock" reported by many? For those not "in the know," this refers to the tendency of Lycoming engines to cruise nicely, and then make about 10-20 knocking noises in a row which then disappears immediately after your ears prick up in concern. In my plane, the frequency seemed to correspond to one cylinder making the offending noise a few times in a row, and then returning to normalcy. It happens often enough that it is not just "automatic rough" (happens also day VFR), and it remains a mystery as far as I know.<< I have not investigated that issue and really could do nothing but add some informed speculation. Regards, George LML website: http://members.olsusa.com/mkaye/maillist.html LML Builders' Bookstore: http://www.buildersbooks.com/lancair Please remember that purchases from the Builders' Bookstore assist with the management of the LML. Please send your photos and drawings to marvkaye@olsusa.com.