SHROUD
Since there is no cowl per-se in the Safari, it actually uses a shroud around the normal top of the engine (forwards after vertically mounting the engine) to supply the baffling to direct the cooling air which comes from the fan/clutch attached to the prop flange. This shroud also provides the ducting to direct airflow through the oil cooler. Originally the factory just supplied paper templates whereby the builder then cut the various panels out of sheet metal and added the various bends before attaching the various pieces. They have since gone to water-jet pre-cut panels which I believe are available to builders, as is the option of having the shroud pre-fabricated. Although there are a lot of pieces and this appears very confusing at first, as one starts to build the shroud it eventually becomes very logical.
I chose to get the pre-fabricated engine shroud with the idea that this would simplify my building requirements and be one less thing to construct. Instead, it's turned out to be just one more area of frustration with ill-fitting parts that I've chosen to replace or re-work anyway. I first tried to mock the panels up on the bench without being attached to the engine. That's when I realized that some of the holes/nutplates had been pre-drilled/installed without the panels being joined since there was absolutely no way that there was an overlap let alone alignment for the screws. I was also disappointed to note that most of the nutplate rivet holes had not been deburred or countersunk thus causing the heads of all the flush rivets to stick up. Since this is in a high vibration area and I believe these rivet heads will just dig into the joined panels, I drilled out ALL of the nutplate rivets (~68 nutplates with 2 rivets each) in order to re-install them properly.
In fairness to CHR, it is my understanding that
steps have been taken by the factory to change how the pre-fabricated
sheet metal parts are assembled and the currently shipped parts
may not have the problems described herein. It should be noted
that the following are just my experiences with the shroud I
received and I have not physically seen what effect the factory
changes have made. Also, when I received my shroud most kits were
still being equipped with the O-320 engine and the use of O-360's
was still relatively new.
Update: I have received feedback from one builder who received a
much newer pre-fabricated shroud for an O-360 and he indicated
that his shroud fit quite well as received without any
significant changes required.
Perhaps I'm being a bit anal about the shroud, but I'm concerned about both protecting my engine investment and also preventing the precautionary or forced consequences of an overheating engine. Since one of the high power requirements is hovering and there will be a lot of this during initial checkout, I intend to monitor the CHT's closely. In the event there's indications of cooling problems, I have a couple of ideas that may be able to increase the air flow.
On a cynical note: All of the nutplates supplied
with the kit and pre-fabricated parts are of the floating variety.
I believe this just promotes sloppy alignment since the nutplate
movement can take out some misalignment. It's interesting that
I've very seldom seen this before and I believe that if parts are
carefully fabricated, clamped, drilled, cleco'd, etc. that fixed
plates should be just fine.
Update: I've had feedback from an experienced Safari owner and he
indicated that he was very glad that he had the floating
nutplates as the alignment has tended to slightly change with
time. It would appear that my previous comments might have been
made too hastily without considering and investigating all the
factors.
N.B. If a builder chooses to eliminate the walking beams and use straight control rods, then two notches will be required in the top and front pieces of the shroud. The exact location and size of these notches can only be determined with the engine, shroud and transmission installed in the frame. The shroud can be built without the notches and they can later be added once their size and location has been determined.
Oh well, time to see how the shroud actually fit on the engine. Note that this is easiest to start while the engine is still horizontal such as in the shipping crate, since the bottom pieces of the shroud encircle two of the conical bushings. In hindsight, I believe the most critical alignment is with the side panels and I believe it is extremely important to get these properly fitted and be able to remove and re-install them such that they are in exactly the same position. If starting over, I would definitely perform this step first.
Bottom Pieces:
The first task I performed was to mount the two bottom halves to the engine along the case/accessory housing joint. The manual talks about drilling and tapping holes into the crankcase, but I found that three of the mounting holes pre-drilled in the shroud somewhat lined up with holes already in my case and the fourth one would need to have a hole drilled into the case just below the top right conical mount. The first question was what to use as a reference before opening up the various holes for alignment. I chose to go with the openings for the conical bushings since they were the largest and should always be the same for all engines. Mistake #1! Later when I was trying to figure out why the shroud was slightly twisted, I discovered that the distance from the edge of the shroud panels to the bushing holes was not equal on both sides. Hopefully the factory will verify this and correct the alignment for future builders.
Note the angle bar clamped squarely to the mounts; ~1/4" forward on left side
Since three of the attachment holes were already in the case and through drilled, I chose to use bolt/nut combinations rather than tapped holes per the manual. Much as I dreaded it, I triple checked alignment and drilled a hole through the case for the fourth mounting bolt. After marking the position, I used plastic sheet taped to the engine to prevent any metal chips from getting into the various nooks and crannies around the accessory housing. Because my mags were already mounted and timed, I had to use a long drill to clear them when drilling; I can't remember whether I used an 8" or 12" drill bit. Finally I could actually mount the bottom halves of the shroud and adjust the various holes in it. I found that I also had to open up the screw hole for where the two shroud bottom halves are joined at the center. This worried me since the back of the shroud already had holes/nutplates in it for where it joins the bottom halves and I knew this would throw the alignment off. As a result, I tried to keep the changes to this hole to a minimum.
As I was trying to fit these bottom pieces, I realized that one of my problems was with the alignment and "springiness" of the curved pieces that wrap around the #4 cylinder. Since they were making it extremely difficult to fit the bottom pieces, I chose to drill out the rivets, remove these pieces and continue the fitting without their interference. Once all the other pieces are fitted, I re-attached these pieces (note the comments later on about #4 baffle). Finally the bottom halves were mounted and I hoisted the engine onto the stand in a vertical position to continue building the shroud.
Update: If I were to do this again, I would spend more time on the #4 baffle area and attach the sides to verify the bottom alignment before hoisting the engine onto its stand. In order to work on the bottom pieces, one has to remove them but they are trapped by the conical mount bushings. The way my engine was crated, the normally upper bushings were not used, meaning one can easily work on these shroud pieces. Once the engine has been hoisted onto its stand then the entire stand must be removed in order to work on the bottom pieces. This means the engine has to be repeatedly hoisted by the prop flange using the engine installation tool or by allowing the engine to rock 90° if using the lift ring. I find that neither of these give me a comfortable feeling when they have to be used for trial and error fitting of the shroud pieces.
The other change I would do is to carefully measure the offset required at the #4 cylinder for the joggle between the two curved pieces that form the baffling. Mine was shipped with ~1/8"+ too much of an offset (i.e. vertical gap) which I believe will reduce the cooling efficiency of what is reported to be the hottest cylinder. At first I thought this was just from a misbend of the bottom piece, but this offset is also reflected in the pre-bent shroud front. Since I'd already removed the curved pieces that wrap around the cylinder, I chose to make spacers so that this gap is reduced to the same amount as on the other cylinders.
The holes for the forward (engine top) ignition wire seals can be seen in the above picture and I used 375 seals from Aircraft Spruce (part # 07-01014). Before cutting the slots for these, one should look at a sideview picture of an installed shroud and note the location of the engine mount structure. The location in the above picture is a compromise between the mount clearance and internal clearance for the cylinders. I'll admit that I lucked out here since I was just working on the shroud and hadn't considered the engine mount when I cut the slots.
Side Panels:
I next chose to attach the side panels which even though they were water-jet cut, already had the cylinder mounting holes elongated for fit. Obviously the CAD templates need a bit more work. In order to mount these, I first had to remove the rocker covers. Oops, these holes needed even more adjustment on my cylinders, especially the couple that had never had a bolt in them. Since I was using the cast rocker covers which are slightly larger, I then installed them onto the cylinders sandwiching the side panels and traced around them on the shroud sides with a felt pen so the panel holes could be opened up for clearance. Interestingly, these marks were not symmetrical no matter how much jiggling I did.
Trim lines are there but very hard to see in this picture
Both of these panels showed "waviness" along various edges. Now that I've got more sheet metal experience, it would appear that this is due to various flexing gyrations that these panels must have been put through during their initial construction. Tightening nutplate screws with protruding rivet heads would have also contributed to this. If I were creating a "Show Special" then I'd probably remake these panels, but since this is tedious and time consuming I'll stick with the panels I have. I did add reinforcement tabs for the elongated cylinder attach holes that will also allow the panels to be consistently re-aligned during installation. These are like a large washer cut from .063" sheet and are ~ 1-½" diameter with a flat on one side for clearance. A couple of strategically placed rivets then hold these "washers" to the side panels. Even without the elongated holes, a builder might want to consider adding these tabs for durability and repeatability in alignment since the side panels are the foundation that all other shroud parts are referenced to.
Tabs are highlighted in black
At Sun-n-Fun '08 I noticed that the factory had made a change to sthe side panels on the red ship. Instead of the two holes per panel for plug access, they had changed it to a hinged rectangular section that also includes the area between the two plug access holes. Although I didn't get to see it in action, I assume this makes it much easier to both install / remove the plug wires and also to guide the plugs when first trying to install them. Even with a crow foot socket, I imagine that access to the plug wires through the round holes will be somewhat difficult. Hmmm ... one more thing for me to check out.
#1 Upper Baffle:
With only the basic structure in place, the top is relatively unsupported so I figured it was time to finally understand all the baffling and install the two top assemblies. The one on the right side from the rear (#2) fit not too badly but has a few gaps that'll need a bigger hammer to close. The one on the left side of the photo (#1) appears to be unusable without some major work!
The fit on the #1 upper baffle was extremely tight but I eventually managed to get it into place against the engine case, shroud side and top. However, the alignment was dismal to say the least. The screw holes that should align with the side panel only did so when pressure was applied to the various bend lines which will be inaccessible once the large front panel is in place and thus it would be impossible to remove / reinstall the side panels without removing the front (i.e. engine top) panel. This piece also had nutplates installed on a bend line etc. ... I think that once I re-make this assembly I'll donate the shipped one to the local technical college so they can use it in their sheet metal course as an example of how not to do aircraft sheet metal.
Update: I've now had a chance to re-make this baffle from scratch and have come to the following conclusions about the factory delivered one:
Here I'm checking the fit of the new baffle with the correct angles before making a new inner bracket. The mis-fit and wrong angle on the bracket that attaches to the crankcase can clearly be seen.
#2 Upper Baffle:
I've now had a chance to take a critical look at this piece to determine if any changes are needed. The supplied part was a bit wide (perhaps 1/16") but this could be changed by carefully reworking the outer bend lines that accommodate the nut plates for the side panel. The bigger problem was that the baffle sat over 1/4" above the cylinders; this is inconsistent with the other side and I was worried that it would create a less restricted airflow that would rob cooling air from the #4 cylinder which tends to run the hottest from what I've been told. In the end, I couldn't figure out how to simply and effectively close this gap so I chose to remake this piece. I added a projection on the rear of this piece so the cylinder wasn't in the direct airflow and possibly pre-heating the air for the #4 cylinder.
I also chose to add a bracket towards the crankcase similar to that on the other side since there was a gap of several square inches that I believe would have allowed for the loss of a lot of pressurized cooling air.
Note the thoughts later on under "Oil Cooler
Duct" on a possible way of slightly modifying this baffle
that might provide additional cooling air for #2 and #4 cylinders.
Update: I've actually made this change during final installation.
Basically it involved cutting a bit of the material out, adding a
couple of nutplates and then touch-up paint. If this had been
done before painting, it would have been more rigid to have
created a secondary bend line.
#3 Lower Baffle:
No matter how I wiggled and twisted it, I couldn't get this baffle to fit with both the bottom and side panel in place. Finally I drilled out the rivets and just installed the main baffle. When I then positioned the small angle bracket that bolts to the crankcase it was obvious that the rivet holes were off by about 3/16". Since new holes would have been too close to the old ones, I fabricated a new bracket. This also allowed me to get a much better fit for the bracket to the crankcase contour with a much smaller gap.
Weird perspective error along top from camera angle
#4 Lower Baffle:
This is actually part of the bottom panel with a couple of curved strips attached that wrap around the cylinder to force the air through the fins. I found that both of these strips were too far away from the cylinder for my liking and were inconsistent with how I'd fit the other three cylinders (starting at just under 1/8").
The curved piece closest to the crankcase was easiest to correct as it only required a piece of .063" aluminum under it to act as a spacer. Because of its proximity to the crankcase and the method of attaching the bottom pieces, this would be hard to do any other way. The other thing I noticed here was how flimsy and weak the bottom panel was in this area and it tended to deform as soon as any pressure was put on the curved strip. I chose to add a piece of sheet under the bottom piece that acts both as a reinforcement for the mounting bolts and also has a fore/aft bend for strength situated under the center of the inner strip.
The outer curved strip was a bit more difficult to deal with since it had ~1/4" tapering gap between the cylinder and the bottom piece. Ideally I should have just remade the bottom piece and changed the "joggle" to reduce the offset by ~1/8", but this would also change the "joggle" requirements on the shroud front. Also, I was once again running out of .032" material and it would have been a long drive to get some more. Just changing one of the bottom piece bends would have altered the width and required both changing the outside bends and relocating the pre-installed nutplates. In the end, I chose to just make a spacer from a combination of .090" and .032" sheet that now makes the gap consistent with the other cylinders.
Note the ~1/2" vertical part of the reinforcement under the inner (large) curved piece
Shroud Top and Fan Guard:
I mounted the shroud top halves with the pre-riveted fan guard to check the clearance around the fan. While there was a little more adjustment required in various places, it seemed I could make a consistent clearance around the fan of about 1/8" - 3/16". Interestingly, the factory chose to mount one nutplate vertically and one horizontally on the lip that actually surrounds the fan. Since the vertical one has minimal clearance to the fan, I will be removing it and changing it to horizontal per the construction manual ... two more unused holes will be left.
After making the new top baffles (#1 & #2), I tried re-attaching the top pieces to check their alignment before drilling for the new nutplates. As I suspected, these pieces did not fit very well without bending various other pieces such as the side panels as it would appear that the top pieces were made perhaps a 1/4" too wide ... they're supposed to fit inside the side pieces. I don't know whether this was due to the CAD templates or the manual trimming and riveting at the factory, but it now appears that I will have to re-manufacture these parts if I want them to fit without a lot of extra bending and stress.
Update: After debating various alternatives I finally came to the conclusion that it was best to drill out the rivets holding the fan guard to the top pieces, straighten the guard and then build new top pieces. In order to straighten the alignment of the fan guard, I laminated two pieces of 3/4" MDF board and cut a circular slot in it using a jig saw and center pivot. I could then hammer and coerce the guard to have a consistent diameter and get the flange bend as close to 90° as possible.
The new top was cut from a single piece of sheet, but I chose to initially leave extra material in the area outside of the circular cutout for the guard. Thus I could drill and cleco the guard as well as mark the bend lines for the outside edges of the top pieces while still dealing with a relatively solid single piece that would become the top. Once the bend lines were marked the top piece was trimmed to size.
Like the factory, I chose to split the top into two pieces along the narrow back edge since it was too flimsy to work on as one big piece. Ironically something still managed to shift during re-assembly and the split line at the back was slightly out (perhaps 3/32"). It was trivial to correct this, but to me it also indicated just how hard it is to keep all these pieces in alignment. Because of the distance between the split and the nearest rivets, there is a tendancy for the ends to lift. I chose to add a small capstrip that covers the split and is captured by the first rivets on either side of the split line. This small strip really helps to strengthen the whole top piece for handling before final installation.
One change I made when reworking the guard and top pieces was to notch the lip of the guard where it rests on the #1 and #2 baffle tops. In one of the previous photos you can see where this lip was shipped even with the edge of the top panels which creates a distinct "step" and tries to make the guard tilt outwards. Because of the guard rivet locations, I actually moved the edge of the guard lip back about 1/2" and also had to notch the top of the #1 and #2 baffles. This creates a much more consistent fit.
The guard lip also interferes with the shroud front and creates a "stepped" effect on both sides of center which causes the guard ring to twist. Because of how narrow the top panels are in this area, I was reluctant to trim / notch in this area. I chose to fabricate a small fill strip that extends the guard lip even with the back of the top panels and ends about half way between nutplates. After the panels were painted I came back to this area and used a bit of RTV to fill this step and create a smooth transition. By using some plastic on the shroud top as a release agent and applying the RTV to the lip on the shroud front piece, this created a smooth junction of the panels such that it won't be trying to dig gouges while twisting the guard and releasing the pressurized cooling air.
Tip: To install this filler strip, I used a trick shown to me by an experienced sheet metal worker. A total of eight 3/32" flush rivets are installed to hold this strip onto the shroud top. The top pieces are countersunk for the rivets and the bottom of this fill strip is also countersunk. By using very short rivets, both of the countersunk holes are filled and both sides of the finished assembly are flush. Effectively, the rivet makes an hourglass configuration after it has been squeezed. Although I would probably not consider using this technique on thin structural sheeting, I believe it is fine for this kind of area which is non-structural and already captured by the screws holding the other pieces together.
What is not shown in the various photos above, are the notches that allow for the straight control rod modification (see Controls sections). There is a notch in both sides, approximately 1-1/2" x 1-1/2", on the forward edge that allow for the rod clearance. I was also worried that some of the drilled out rivets in the fan guard had been over-bucked and the holes were no longer in good shape. While I could have re-drilled the holes and stepped up to 5/32" rivets, I chose instead to go with dimpled countersunk rivets. The dimpling process was done on both pieces simultaneously which helped to reform some of the holes plus the two "cones" essentially lock the parts into position. After dimpling, the pieces would repeatedly and reliably go together in exactly the same position.
Note the notches for the straight control rod modification.
Center Divider:
Early on I thought I'd trial fit the center divider to check it's fit and it turned out there was a slight opening of the case bolt holes required, but not enough to worry about. The top appeared to mate fairly well with the other pieces, but the bottom was a real problem since it was ~1/4" wider than the bottom pieces and prevented the large front panel from being installed properly. Perhaps this is due to a difference in the Lycoming vs. XP-360 cases (which I highly doubt), but this left me with three options:
After debating it for a bit, I decided to make a new center divider and this allowed me to make several changes. I used six attach holes, rather than four with the implication being that I can now leave the lift ring in place where I won't lose it. The standard divider uses the lift ring bolt which won't allow for the thickness of the lift ring being between the divider and the crankcase, whereas I chose to use the bolts on either side of the lift ring bolt. When re-making this piece I chose to finish all the fitting and alignment along the crankcase before carefully marking where the bend is required based on the top and bottom panel positions. This corrects the ~ 1/4" too wide at the bottom and the too narrow at the top conditions in the CHR-delivered divider.
The plans don't show any attachment of the divider to the large front panel. Since everything is already very rigid in this area, I chose to add only four nutplates along the divider to stop any vibration movement and further clamp everything together. While one would need a near-zero tolerance right-angle screwdriver to remove these screws with the engine installed, the top lip is probably a bigger issue when removing the front panel.
Update: Now that I've had a chance to see the shroud on the engine in the frame, I realize this may not have been such a good idea. While there is perhaps just enough clearance to perhaps get a 90° screwdriver in here, I didn't take into account the fact that its a long way from the centerline to the edge of the shroud. I'm definately going to have to re-think this one as I can't think of an appropriate fastner that would work easily. The only easily workable idea I have is to actually mount the nutplates on the forward side of front panel and actually attach the screws from the inside. Since at least one side panel needs to be removed in order to remove the front panel, this should work.
Update 2: I've checked to see if its possible to remove the crankcase nuts but leave the divider attached to the front panel when removing it such that it can be brought much closer to the side where one can get better access to the screw heads or even removed while still attached. This seems to work fine and the nutplates can be on the divider, but one needs to consider and allow for this if they are running either an injected engine or mounting primer lines on the normal "top" side of the engine. Any fuel lines that need to pass through the divider will need to be free to "slide" through their opening(s) in the divider.
I have a primer system which I ran into the injector ports rather than the primer ports. Although this is slightly less efficient, the major negative is that these lines will not be visible during a pre-flight, but this is no different than on an injected engine. In order to be able to remove the front panel and divider as described above, I made a small piece of the divider, in the bypass hole area, removeable. It is attached to the main divider with a couple of screws / nutplates and is used to mount an adel clamp on the primer line that passes from one side of the engine to the other. The screws are accessible with the shroud side removed and thus allow the divider and front panel to be removed without affecting the primer lines. The same technique could be used for injector lines.
Shroud Front
I also had a lot of trouble getting the large front panel to fit properly. There were several nutplates that needed to be moved since they didn't line up with their mating holes and regardless of how I tried to fit it, the side panels which form part of the baffle sealing were forced outwards. I believe part of this was due to the edge bends being at the wrong angle during manufacture as there was ~1/8"-3/16" gap from the edge of the lip to the side panel. While this bend could have been changed (i.e. opened up) I'm started to think that the bigger problem was with the initial location of the bend lines.
After trying to fit this panel again, my first task was to remove all the rivets holding the nutplates since their heads were all sticking up and needed to be properly countersunk. It turned out there was only one side whose bend line really needed to be moved about 1/8"+ or so. However, there was still the issues of modifying the side bend angles and of multiple nutplate holes that didn't line up and would require both the screw hole to be opened up and the rivet holes repositioned. In the end I decided it was just as practical to make a whole new panel and position the bends and nutplates properly.
In hindsight, I hadn't fully realized what a pain this panel would be to fabricate. I chose to make the bottom / top bends first and the first problem was making the bottom bend around the joggle. None of the brakes I had access to could handle the long bend at the bottom since the joggle projects beyond the bend line. Eventually two clamped bars and a lot of blue air got an acceptable bend. The remaining bends were then relatively straightforward after carefully measuring their position. Because of the "tilt" to the panel, the edge bends are actually at a changing angle from top to bottom. Also, the top bend is slightly over 90° while the bottom bend is slightly less than 90°.
After marking the nutplate location, a bunch of progressive work with the fluting pliers resulted in a good fit. The nutplate holes were marked / drilled and now this panel is ready for final prep / paint and assembly.
Update: It's a good thing that I hadn't painted this panel yet as I've modified it for the potential use of straight control rods (see Controls sections). Before fluting and drilling nutplates on this piece, one might want to get the measurements for these slots so as to avoid putting nutplates too near to them. I chose to make filler plates from .025" aluminum to fill in the notch on the top plus the slot in the front panel and going to the thinner metal made it much easier to form. And yes I did try to first make them from .032" material which then ended up in the garbage can. These filler plates are riveted to the front panel and each have one screw / nutplate into the top panel.
Tip: If one is considering the straight control rods, then they may want to consider initially running this panel through a slip roller. This will make it more difficult to form the top and bottom bends, but it will make for a better panel. If one is only using fluting or a shrinker, the problem comes in when one cuts the notches for the filler plates since if the panel hasn't been rolled then it will want to straighten across the notches and create stress points. This also makes it more difficult to fabricate the filler plates. Once the panel is finally installed, the screws connecting it to the top piece will relieve most of this stress.
Oil Cooler Duct
Although this piece could possibly be used as delivered, it looks like I'll probably be making some changes to it both to reduce the fan gap and to correct a canted attitude. In order to increase efficiency, the rest of the fan guard has had a lot of work to make it a consistent 3/16" (or slightly less) clearance. The oil cooler duct side of the fan guard has a much larger clearance; perhaps 5/16" or more in most places. Because of where the bend line was placed on the lower front edge, there is also a very distinctive upwards tilt towards the rear edge of the duct and approximately 1/3 of the fan blades are below the guard ring.
I elected to re-fabricate the top piece of the duct. First the rivets attaching the top to the sides and guard ring were drilled out and the lip on the guard ring was squared up. A new top piece was then cut and bent as required. If I was really trying to make a "Show Special" then I probably would have remade the side panels, but I found that by slightly splaying the tops of them outwards I managed to reduce the fore/aft tilt to an acceptable amount. Also, I considerably opened the hole for the top of my MagnaFlite starter which allowed the front edge of the cooler duct to sit 1/8"+ lower. The plans and construction manual show two screws on each side inside the duct holding it to the top part of the baffles whereas the received parts only had one; I elected to stay with just one screw/nutplate but may have to re-consider this at final installation.
During final installation, I had about a 1/4"+
gap between the edge of the duct and the oil cooler. This needs
to be filled with weatherstripping insulation similar to what is
already on the top and bottom of the oil cooler. I'm still not
impressed with the way the duct tilts upward at the back but it
would have required a major change during construction. I'll try
this one out and see how it works but if the oil temperatures are
too high then I'll have to consider re-making this duct ... not
something that I look forward to.
Update: The weatherstripping options I found were either a very
soft open cell (up to 1/2" thick but almost totally
compressible) or a fairly firm closed cell foam. I obtained both
some 1/4" thick and some of the 3/8" thick closed cell
variety ... it turned out that the 3/8" thick foam was the
best one for my installation.
Airflow Thought: While working on this duct, I couldn't help but notice the lip that is formed by the duct to baffle joint. Above the #1 cylinder is not an issue since with the direction of fan rotation this lip will tend to retain pressurized air for the #1 and #3 cylinders. However, it would appear to me that the lip above the #2 cylinder would just create an area of turbulence due to the upslope of the oil cooler duct. If I'm having any cooling issues with #2 or #4 cylinder, I'm thinking that I may try cutting this lip off the oil cooler duct and also the corresponding area on the #2 baffle. A piece of 3/4" x 3/4" x 1/16" angle stock would then be placed in the junction with rivets and screws / nutplates. This would effectively eliminate the internal lip and allow any air that is trapped along the upslope to spill over into the main shroud.
Yellow outline indicates lip that could be removed
Update: I decided not to wait on this change as it's probably just one more thing that I'll never get around to if it's left till after the machine is flying. It would have been easier to have done this before painting since the tab where it joins the #2 baffle could have been bent down instead of up. Since this is a retrofit, I've chosen to just cut most of the tab off and go with the angle stock. Similarly, part of the #2 baffle needed to be cut off and new nutplates installed ... I used two since this area has been somewhat weakened due to removing the bend. Once the engine is running, I'll try remember to record the CHT's here to indicate which is the hottest cylinder and whether this modification might have had any effect.
Angle stock being fitted and ready for riveting
& painting
Edge tab on lip was left so RTV could be used to seal the edge
gap
Heat Collector
I've just done a preliminary fit of this while trying to mount the oil cooler. I definately have to do some work on this piece as it was delivered with one of the side bends at the wrong angle such that it sticks out ~ 1/2 inch, there had been no deburring and there was about a 1/8" overhang on the riveted edge. The fact that most of the rivets had been over-bucked did not help the looks. I can see from the manuals how the two screws will hold it at the top but it does not appear that there is any default means of retaining the bottom except the pressure from the foam on the oil cooler and also the SCAT tubing. I'm waiting for final installation, but I will be looking at adding a removeable strap(s) that go under the oil cooler and locks this piece to the oil cooler duct.
One idea that I'd had was to try see if there was a way of incorporating the cabin heat selector within the heat collector. The idea was to try reduce the back pressure on the airflow through the oil cooler on hot days and not force it to go around bends and through the exhaust heat exchanger before being vented. This idea has been abandoned as unsafe since if there was a leak in the exhaust system there would be no way to prevent the CO / CO2 in the exhaust from entering the cabin. Its much safer to just remove the heat collector on hot days if there are any concerns about oil cooling.
Update: I decided to drill out all the factory rivets and separate these pieces so that I could work on them. The various pieces were then trimmed and edged. The rivet holes were dimpled for countersunk rivets which helped to clean up the holes and assisted in locking these pieces together. Although I riveted these pieces back together to expedite the painting, I think it might have been wiser to wait until final assembly and the exhaust system is installed. Until that point, one can't determine if the ring for the SCAT tube is in exactly the correct position.
The engine has now had it's final installation
and I installed the oil cooler and exhaust to check the fit of
this piece. As suspected, the factory positioned ring for the
SCAT tube is in the wrong place. It's a bit off to one side and
over an inch too far back. Since the distance between the heat
collector and the exhaust is very short (perhaps 4") this
makes the offset very noticeable. Not very impressive since both
parts were pre-fabricated by the same factory ... I've seen other
ships with the angled SCAT tube and now I fully understand why. I
would recommend that the hole for the SCAT tube ring not be cut
until after the engine and exhaust are in place and the proper
location can be marked ... it's a lot easier to defer painting (or
apply touchup paint) than it is to install a patch plate.
Update: I'd deferred changes to this piece as I was waiting for
the final installation of the oil cooler duct and I'd removed the
exhaust system due to a need for some rewelding which I assumed
might change the fit. Now that these tasks are complete, I re-attached
the heat collector. Relative to the exhaust, it looks like the
ring is mounted 1-5/16" too far back and about 1/16" to
one side. The sideway offset would be fine and can be attributed
to mounting differences but the fore-aft offset is excessive. The
choice is a heavily twisted piece of SCAT tube or one more piece
requiring further changes.
I knew the fore-aft alignment of the SCAT tube was going to bug me if I didn't fix it and I finally got around to it. There is no debate now about what is the ugliest piece of sheet metal on my craft ... that will teach me to assume that a pre-fabricated part had correct alignment and then go ahead with painting it.
Before touchup paint
The two vertical tabs on the sides need some kind
of anti-chafe material where they rest against the oil cooler.
I'm using some left over 1/16" rubber strip from the fuel
tank saddles but if this causes problems I can always mold a bit
of high temperature RTV.
Update: As I was getting ready to attach the rubber strip I
realized that this wasn't the best material. The heat collector
is firmly attached to the oil cooler duct which forms part of the
shroud which in turn is firmly attached to the engine. However,
the oil cooler is firmly attached to the frame and any movement
of the engine will result in movement between these shroud parts
and the oil cooler. I elected to use some 1/4"
weatherstripping material between the vertical side tabs on the
heat collector and the oil cooler.
To lock the heat collector to the oil cooler duct, I used three nutplates inside the top of the duct. That solved the top connection but still left the bottom of the collector "floating". I decided to add two brackets (one each side) under these parts to lock them together and due to installation issues clearing the oil cooler, these brackets are attached with screws / nutplates at both ends ... seems to work fine.
I've only done engine runs in relatively cool
weather at this time and have not experienced any significant
cooling problems. However I've been told (and it's pretty obvious)
that this heat collector is a major restriction for the airflow
through the oil cooler. I've noticed that several owners operate
their machines in warmer weather with this collector removed in
order to get maximum airflow through the oil cooler. As an
alternative, it should be possible to design and build a simple
trap door in the top of this collector that would allow for
direct venting of the air without having to remove the collector.
Project #73 on the "to investigate" list ...
Update: After doing some extended hovering,
I can agree that the heat collector adds significantly to engine
oil temperatures. It was under 70°F outside but I still had
thinner 80 weight oil (30°F to 90°F) since we had done our
previous testing in much cooler temperatures. The oil temperature
would slowly continue to climb during extended hovering. We then
removed the heat collector and the oil temperature remained rock
steady at 180° regardless of how much hovering we did. It's very
obvious to me that I need to do some more work on this piece if I
intend to leave it in place year round.
General
The more I read and hear about baffling, the more it appears that it is critical to fully seal all gaps since cool airflow is critical for an air-cooled engine and it would seem that even small gaps allow for a pressure drop. Due to the design of the shroud there are several places where there are small gaps of perhaps 1/8" wide and a couple that are 1/2" wide. While these do not appear to be very big individually, they quickly add up to several square inches. Careful construction will allow the builder to minimize some of these gaps, but many of them will still remain, such as where they contour around the engine. There is an article in the April '07 issue of Kitplanes magazine that makes it appear that baffle sealing is not very important and the emphasis should be on air flow. While I won't dispute the article, I think it is important to note that it refers to the ram air cooling on a fixed wing aircraft. In our case, we have a shrouded fan providing the pressurized cooling air.
During final assembly, after the shroud was
painted, I used high-temperature RTV to fill as many of these
gaps as possible. I chose to first roughen with course sandpaper
any of the pieces where the RTV would contact and often used an
excess of RTV which was later trimmed with an X-acto blade. We
use a very thin flexible release tape (similar to mylar packing
tape, but with stretch properties) at the hangar and this was
placed anywhere that a release barrier was required. I chose to
build pads about 3/32" x 3/8" on the ends of the pieces
that curl around the cylinders and then a narrow strip (perhaps 1/8"
to 1/4" wide) along the edges of these pieces. Anywhere that
I could see a gap of more than about 1/32", it was also
given the RTV treatment. This is a slow process that takes a lot
of elapsed time to allow the RTV to cure and I chose to do it as
a side-task during other final assembly tasks. I was amazed at
the number of areas that I did this on, but I feel it was worth
the time and effort since the total area that was sealed quickly
added up.
Update: The release tape that I used did not work very well with
RTV even though it is extremely effective for composites. I would
suggest that anyone using this technique should also use PVA as a
release agent. I've used UHMW tape as an RTV release tape in
other areas and this worked extremely well, but it's too thick to
go around compound contours.
Note the Permatex Ultra-Copper RTV along the cylinder area
In order to hold the baffles in position around the cylinders, most builders appear to use safety wire between the upper and lower flanges. It's interesting to note that the latest baffle kits from Van's now include a rod that is threaded 6-32 at both ends which allows one to add some rigid "kinks" to go around any obstacles. I would assume that 0.140" stainless rod would be the ideal choice for this.
Operational Results
During the initial hovering it became apparant that more work is required on the shroud. The first problem I faced was with the CHT probes which was fixed. During further testing, it appears that cylinders #2 & #3 are running significantly hotter (40° plus) than #1 and #4 and required a cool-down period after about 5-10 minutes of hovering. Cylinder #4 reads considerably lower and perhaps that is due to a different kind of probe in there ... it still needs to be investigated. Cylinder #3 will be investigated by closing up the slot in the center divider ... the thought is that this slot allows air to move over to the other side of the shroud and creates a low pressure area at the outermost part of #3's head where it is needed most. If that doesn't work then I'll look at creating some kind of ramp to smoothen the airflow under this cylinder rather than the vertical baffle that is under it.
The action plan for cylinder #2 isn't quite as obvious. I think the first step will be to remove the extension I added that stops the incoming air from hitting #2 directly and possibly pre-heating the air before it gets to #4.
Final Thoughts
Perhaps the factory has now changed the quality of their pre-fabricated shrouds, but based on my experience I would have to see one of these before recommending this $1,300 US option. For that amount of money, one could get a quality bending brake, a shrinker, a stretcher, rivet squeezer etc. and still have money left over. I believe it actually took me longer to use the pre-fabricated shroud since I spent a lot of time first trying to make it fit, then trying to figure out what the real problem was and finally just re-manufacturing most of the parts from scratch. Not one piece was used in it's as-delivered state!
While the factory appears to pride itself on championship awards and the manuals constantly refer to show ships, it appears to me that some of the parts I received were never meant to be used in creating a "Show Special". While I've heard factory representatives talk about such details as the number of turns per inch on safety wire for show ships, it makes me wonder how the judges feel about 1" obvious misalignments.
One change that I would prefer to see on pre-assembled parts is the use of #40 holes for 3/32" clecos instead of 1/8", and the parts being shipped without riveting or nutplates installed. There are several reasons for this:
If I were to do it again, I would consider starting with the pre-cut parts or just the templates and then very carefully check each piece and also be very careful with the various required bends while basing them on the existing structure rather than the pre-cut marks. CHR's website now lists the flat laser-cut shroud parts at $450; seems like I'd have to seriously consider just cutting the parts myself. There are three pieces that may pose difficulty in a home shop without a shrinker/stretcher; the two pieces for the fan guard and the top/bottom edges on the large curved front panel. Perhaps arrangements could be made with the factory to pre-form these, but I have not investigated this. I would not get all the edges on the large front panel pre-bent since the side bend line positions will be influenced by the positioning of the other pieces.
<NEXT> Controls and Drivetrain
<BACK> Engine & Related Equipment
Last updated: September 18, 2009