(5.67b/IDA-1.5); Fri, 11 Aug 1995 12:09:12 -0400 (5.67a/IDA-1.5 for marcz@hpwarhw.an.hp.com); Fri, 11 Aug 1995 12:08:00 -0400 From: "Dewey Davis" Date: Fri, 11 Aug 1995 12:08:57 -0400 "Re: COZY list" (Aug 10, 17:23) Subject: Cooling On Aug 10, 17:23, Marc J. Zeitlin wrote: > Subject: Re: COZY list > Dewey; > > In looking over old newsletters (to get them on-line), I noticed that > you had written to Nat regarding an engine monitor you were going to try > to implement that would be more accurate than the commercially available > ones. I was just wondering if you ever got that working, and if so, > maybe you could describe it for us. Thanks. > > -- > Marc J. Zeitlin E-Mail: marcz@an.hp.com > >-- End of excerpt from Marc J. Zeitlin Yes indeed I have spent a lot of time trying to get an accurate measure of engine temperature. When I first started flying the COZY, I talked to a lot of other builders/flyers about their engine temps and I noticed a wide variation (cylinder head temps from the incredibly low 300s to the high 400s). My own CHTs were a little above average and I set out to discover why. The first thing I learned is that MANY buillders have no idea what their real engine temperatures are. The cheap Westach gauges that are so often used in homebuilts are notoriously unreliable (actually the standard fare in Cessna and Piper machines are not any better). But the biggest problem with accuracy is caused by our installations. Since most standard engine CHT and EGT thermocouples are designed for tractors rather than pushers, they are usually only about 4 feet long. I can remember talking to several builders that proudly proclaimed that their engines were running so cool that they never got above 330 F, even on the hottest days on the longest climbs. After looking at the installation it is obvious why that was so. The standard thermocouples are connected to the cylinder bayonet mount or spark plug and then run to a junction on the firewall. Standard wire was used between the firewall and the instrument in the cockpit. All any thermocouple can measure is the DIFFERENCE in temperature between one thermocouple junction and another, in this case betweeen the cylinder and the firewall. Once the installation is corrected by adding real thermocouple wire all the way to the instrument, the indicated CHT reads 150 F higher! Dangerously close to engine redline. This may sound elementary to some, but you wouldn't believe how many times I have seen this on pushers that have been flying for years with the builder oblivious to real engine performance. Another problem is that some of the instruments are not temperature compensated, particularly the cheaper ones. That means they are only accurate at one instrument temperature, usually 70 F. So if your cockpit panel is actually 95 F, your engine temp is actually 25 F higher than what your gauge is telling you. Even the expensive gauges costing a kilobuck or more with built-in temperature compensation can be misleading if not installed correctly. Another problem with our airplanes is that Lycomings are designed for downdraft cooling and we are attempting to cool them by running the air through them backwards. If you look at the fins around the cylinders, you can see that Lycoming expected the cooling air to enter from the top. That is also why the bayonet plug is on the bottom instead of the top. With regular downdraft cooling, the hottest part of the cylinder will usually be on the bottom. The CHT probe will measure the hottest part of the cylinder. With our cooling arrangement, the top of the cylinder is definitely the hottest, by a lot. I have measured the top of the cylinder under the top spark plug, the bottom spark plug, and the bayonet probe; all three at the same time. There can be as much as 90 F difference between the top plug (hottest) and bottom bayonet (coolest). Under most flight conditions there is about 60 F difference. I have talked to Lycoming about this and they have very little advice. They leave all the cooling design to the "airframe manufacturer". They can't even say what their spec (500 F) means in terms of where and how it should be measured, except to say that we shouldn't go over 500 F CHT. We are on our own with our cooling design. Anyway, I decided I didn't want to spend hundreds of dollars on a fancy temperature compensated engine monitor. I found a $14 chip from Analog Devices that is specifically designed for Type J thermocouple temperature compensation. They also have one that is good for Type K thermocouples (the kind you need for the high temp EGT probes). For a few dollars more you can get one with accuracy to the nearest 1 or 2 degrees. I designed a little circuit card to handle a couple of these chips, plus the surrounding circuitry for conversion from centigrade to Fahrenheit, plus a couple of rotary switches and a digital display. I put it all in a little box for panel installation. I even made a silkscreen panel for it. Actually I spent WAY too much time on the instrument itself (like a lot of other diversions in the building process). But I did end up with an extremely accurate gauge. I have calibrated it against other gauges (expensive ones I borrowed), some temp probe multimeters, and even the welder's chalk that will melt at a precise temperature. With 400 F welders chalk on the probe, I can heat the probe with a torch and watch the temp climb through 400 F and the chalk will melt right as the gauge changes from 399 F to 400 F. I started all this to find out if I can cool the engine a little better for longer life and better reliability. I could go on for hours and hours about the things I did to actually improve cooling. I bet I have tested at least thirty different mods from various baffling tricks to vortex generators in the scoop. But that is another story that so far has no end. I had intended to write it all up once I discover the smoking gun. But I haven't found one yet. I can only say that there is one hell of a lot of myth and fantasy in the cooling world. For instance, Nat will often say that you can regulate cylinder-to-cylinder temps by putting small ramps on the inside bottom cowling to direct air to each cylinder. That is not true in most cases. I'm not saying it didn't work for him; I believe him. But I also know that it had absolutely no effect on at least six other airplanes. Generally, there is no such laminar "flow" once the air gets about 4 inches inside the cowling. After that it is entirely unpredictable. The cowling is more like a turbulent pressure chamber than a flow guide. I also built a manometer to measure pressure and flow throughout the cowling. But again, thats another story. Terry Shubert has written about some of those ideas in the Central States Newsletter. My next attempt to improve cooling is to route the exhaust through the aft end of the cowl nearer the spinner, rather than use the separate exhaust exits in the standard cowl. Several other builders have tried this now and they are getting some encouraging results (30 to 50 F cooler CHTs) due to exhaust augmentation of cooling air flow. After nearly two years at this problem, there seems to be no end. I guess that is why they call these airplanes EXPERIMENTAL. Dewey Davis