Improving Engine Cooling
I finally solved my cooling problem! The problem was that my cowl exhaust exits were undersized. This was not easy to determine, but now, it makes so much sense I wonder why it took so long to realize...
I've been battling higher than desired CHTs on my engine nearly from the first flight. Switching to the Fence Baffles helped out quite a bit, but as the outside temperatures started rising with the onset of Spring, the engine started running a bit hotter than I wanted. For instance, with OAT of about 70 degrees, the CHTs would range from the low 300s on the front cylinders, to about 380 on the rear. This did not provide the margin of cooling that I wanted, and I had been working extensively to improve the performance of the baffles to bring down the temps.
I devised and tested about 10 different iterations of baffles. Some were additions to the fence baffles, some were hybrid fence-box systems, but nothing I did significantly lowered cruise temps.
I turned to the NASA technical reports on engine cooling for possible answers. Reading NASA CR 3405 (7 MB .pdf) that the solution finally occurred to me. The NASA report is quite long, and far ranging, but the relevant info was contained in the sections regarding cowl flaps. Indeed, this is what attracted me to this report, and everyone said, "Maybe you need a cowl flap to get some better cooling..." I really didn't want to add the complexity of a cowl flap, nor did I want to casually cut holes in my cowl to add additional cooling area. Plus, no other Sonex needed cowl flaps, so why did I?
Something I found interesting was on page 59 of the report, which states in the section titled 'Cowl Flap Test Results' that cowl flap exit area (really, cross sectional area, measured perpendicular to the flow streamline, although not explicitly stated in this section of the report) rather than defection angle, is the controlling parameter to cowl flap effectiveness. What this meant to me was that I could scratch off the idea of increasing the size of the deflector lips in an attempt to create a "stronger low pressure area" to help extract cooling air from the cowl. Essentially, this would be the same as changing cowl flap deflection without changing area. In short, it wouldn't have helped!
Additionally, this section also turned me on to the real problem, namely that the exit area should be measured *as installed* on the airframe, and *perpendicular to the flow*. Since my cowling was slightly long and overlapped the bottom of the fuselage somewhat, thus effectively reducing the exit area. The sketch below (click to enlarge) illustrates what was happening.
Measuring my cowl exits in this fashion, I determined that I only had approximately 39 square inches of exit area, not the 47-50 sq in called for in the AeroVee installation manual! Essentially, I was 20% short, and this was causing the ineffective cooling performance. Temporarily wedging the cowling open to increase the exit area approximately 8-10 sq inched showed this to be exactly the problem. With the proper exit area, CHT came down an immediate 30-50 degrees! The engine was finally cooling properly, and I felt like I just won the lottery....
There are several take-away points here.
The next step is to permanently modify my cowling to increase the exit area. This will involve removing the existing deflector lips, cutting out some additional fiberglass, and then installing new deflectors.
NOTE: If you do not see a menu frame on the left, click here to reload the full page.
Updated: 28 Apr 08