Modifications: Bigger oil cooler for IO-390-EXP119

TLDR:  Oil temp was too high with EXP119 engine.  Installed larger oil cooler.  Oil temp problem resolved.  Read on to learn all the details. 

This post will explain why and how I installed a larger oil cooler in my airplane.  The final product is shown below, in comparison to the original, in the next two images.

Note:  I completed duct designs to fit the stock AS-2006X cooler for the 5" and 4" SCATs.  Follow that link to download the models for printing if you'd like to try them.  I'm making them available to the community for free.

The image on the right below shows the installation of the larger oil cooler with a manually fabricated oil cooler tray and a 3D CAD-designed oil cooler duct.  The image on the left shows the original oil cooler installation.  Though that image is from prior to my EXP119 engine conversion (best picture I had), the comparison is still valid as it's still the same 5" SCAT design used in the EXP119.

 

The image on the left below shows Vans' stock oil cooler duct whilst the image on the right shows the duct I designed. 

 

 

 

Now for the details...

 

Following my conversion of the IO-390-A3B6 engine to the -C3B6 (-EXP119), I experienced excessive oil temperature in the summer, frequently hitting 225 degrees and staying close to that during cruise.  Below is an example at 115 IAS and 83 OAT during level flight following the climb (using my free flight data analysis program).  The x-axis is the time since engine start in minutes.  Takeoff was around the 10-minute mark.  The oil temp peaks at 225 and stays well above 210 for the flight (CHTs ~330-360).  I have several datasets with similar presentation during the summer.  

 

I appreciate that the below looks like the vernatherm is closing late, however I don't believe that was the case.  What might be happening here is the high power through the climb is of course increasing OilT but also the high pitch angle is restricting airflow through the cooler during that period.  Only until the nose is pushed down with a subsequent 10" reduction in MAP, does the OilT finally stop rising.  This effect wasn't as pronounced before the engine/cowl modification.

In my view, though Lycoming's maximum allowed temperature is 235 degrees, the higher OilT I was experiencing rendered the aircraft unflyable on warmer days, whereas prior to the conversion, oil temperature was not an issue in the summer as ~190 degrees during cruise was typical.  

In an attempt to diagnose the issue, the vernatherm was tested and found to be functioning properly (it was the same vernatherm from prior to the modification).  I even replaced the vernatherm with two others, one used and another new (from Superior, rather than Lycoming), to no avail.  The path of the oil through the engine and into the cooler was verified via guidance from an A&P.  Trying a new oil temperature sensor didn't make a difference either. 

I eventually removed the FF-01414 RV-14 X-Over Cowl Close-Out in an attempt to increase airflow through the cowl (no evidence of fluttering in the associated area was found when flying with this part removed).  This markedly improved the situation and allowed the aircraft to fly in warmer weather again.  However, I didn't consider this a long-term solution.  Turns out, the first time the plane sat on a ramp overnight during a cross-country without the close-out in place, I returned in the morning to find that a bird started building a nest in my engine compartment despite the use of cowl intake plugs.

I came to believe the new cowl design was responsible for a reduction in the effective cooling capacity of the stock oil cooler (an Airflow Systems' AS-2006X, a 13-row design).  This perhaps makes sense given that the cowl design associated with the EXP119 conversion seems to reduce the airflow through the engine by closing the exhaust tunnel in favor of the primary air exit being the conformal openings around the new exhaust pipes.  As a compensatory move, Van's includes a pilot-controllable "ramp flap" to allow air to pass through the now unnecessary exhaust ramp for additional cooling, however in my case, that was clearly insufficient.

Since I wanted the F-01414 close-out in place, there were at least two options I could think of to restore the cooling capability of my airplane.  The first was to open up the cowl more by louvers or cowl flaps.  The aesthetics and complexity of that option weren't appealing to me.  The second option was to increase the size of the oil cooler.  

The efficacy of a larger cooler assumes that the increased surface area of the larger cooler will increase cooling despite the presumed restriction of airflow through the cowl.  E.g., even with the same inadequate air volume moving through cowl, there might be better heat transfer from a larger oil cooler due to more contact between that airflow and the cooler.  Now, if I purport to have any skills at all (indeed, worthy of debate), it surely isn't anything related to aerodynamics or heat transfer.  However, I thought a larger cooler might be a reasonable path to a solution if indeed the cooler could fit.

I looked at the next larger cooler available from Airflow Systems, the AS2007X.  Here is a side-by-side pic of a AS2007X (left) and a AS-2006X (right) showing the two additional rows on the larger cooler.

I next needed to investigate if installing the AS-2007X oil cooler with its 2 additional rows was feasible.  That includes the following steps (the last two being additional available options if steps 1-4 prove inadequate):

1. Verify bigger cooler can fit without impinging on items already in place (engine mount, hoses, etc.).
2. Fabricate a larger oil cooler tray (to replace the stock FF-01404).
3. Design and fabricate a larger cooler inlet assembly (to replace FF-01406B, FF-01406C and "DUCT  FLANGE") to connect the SCAT tubing to the cooler.  This inlet should have a smoother transition than Vans' box-like assembly.  It should integrate vanes for structural rigidity and to help reduce air turbulence.
   
4. Use SCEET tubing rather than SCAT for "VENT SCAT 4X21" in an attempt to reduce turbulence inside the duct. 
5. Optional:  Include a bell-mouth or additional transition part from the baffle to the SCEET tubing.
6. Optional:  Include an exit diffuser to help air move through the cooler (by lowering the pressure at the cooler exit).
   

With everything above in mind, I will now describe my process to accommodate the larger cooler.  The order I took is not what is outlined below, but it's easier to follow when presented as shown.

Relative to item 1, fitment, the larger cooler has the same width and height and is a bit more than 0.8" longer than the smaller cooler.  I verified that the engine mount wouldn't impede fitment of the bigger cooler on the cooler's inboard side if it's not placed too far aft, though as I later determined and explain below, the cooler's aft hose fitting would impede placement without additional considerations.

For item 2, fabricating a larger oil tray, I used 6061T6 0.05" aluminum sheet and the stock FF-01404 as a guide to design a tray to accommodate the larger cooler.  This included using a metal brake, cutting the rectangular hole, adding lightening holes and match drilling holes for mounting to the firewall.  Care was taken to match the bend angle of the aft flange to that of the original FF-01404 (the angle is greater than 90 degrees).  I decided to use all six bolt holes on the cooler (the stock design for the smaller cooler uses only four of the six).  Note:  The position of the rectangular hole on the tray was located after the inlet duct prototype #3 was fabricated (duct prototypes are outlined later below) and positioned properly - again, I'm showing the sequence of events out of order for clarity.  Below illustrates the tray fabrication steps.

Because the larger oil cooler extends further inboard, its aft oil hose fitting will contact the engine mount.  Since I used AN822-8 steel fittings (Vans' plans incorrectly call for using aluminum AN822-8D which will cause galling), I simply ground down the aft fitting, then primed it, to provide additional clearance.  There is a lot of material remaining so I'm not concerned about the part experiencing an aneurysm from the oil pressure.  I ultimately clocked the aft fitting further clockwise (relative to its position on the smaller cooler) to accept the FF-00020 Hose since that hose has to accommodate the connection at a slightly different position with the larger oil cooler.  Likewise, the forward fitting was clocked more counterclockwise than for the smaller cooler.

Pictures below show the aft AN822-8 steel fitting before and after grinding it down, priming and clocking it.

The inboard mounting tab for the tray on the engine mount is of course positioned for the smaller cooler's tray.  So I used an aluminum standoff to allow the larger tray to capture that tab on the engine mount.  This standoff is 0.486" and has an AN-13A bolt with a NAS1149F0362P washer at the nut and a NAS1149F0332P washer between the spacer and right tray support.  Also note that I needed to cut the right oil cooler tray support (analogous to FF-01404A) to clear the SCAT tube for the right heater outlet, hence its odd shape in the image below which shows the design after initial cuts were made and before the final cuts were made.  Note that you can just see the white prototype #3 of the oil cooler duct sitting on the oil cooler (descriptions of the prototypes are found later in this post).

Following priming, below is the completed larger tray with lightening holes, side brackets and stiffeners including an additional arm to capture the mounting tab on the engine mount.  During fabrication and alignment, I was able to clamp and match drill this tray into place level to 0.1 degree.  It's worth noting that this was the second tray I constructed.  Though not shown, the first tray I made had the oil cooler positioned too far inboard such that there was no reasonable way to get the SCEET tubing to clear the engine mount.  Thus it's clear why Van's chose to use a rectangular flange for the SCAT tube connection on FF-01406E for the 5" SCAT tube (visible midway down in this link) whereas for the original design's 4" SCAT tube, FF-01406B is round (visible in the last photo in this link).

Here is a comparison between my tray and the stock tray.  This was before I added four additional rivets to better attach the aft side of the two tray supports (those rivets are shown in the above two images).

Below is the larger tray installed.  You can clearly see the approach I took to capture the inboard mounting tab on the engine mount.  My countersinking depths were unfortunately uncharacteristically not ideal but sufficient for the purpose.

Here is the larger AS-2007X oil cooler in place during trial placement.

With the tray now in place, item 3 from above is the next item of concern:  Designing a cooler inlet assembly.  This required me to learn how to do 3D CAD.  For my panel, I learned how to do 2D CAD with LibreCAD, which is relatively straight-forward.  However, I appreciated that learning how to do 3D work would take some effort and I was expecting to have to derive equations to describe the surfaces placed in space.  I didn't want to purchase any software so I looked for free options.  OnShape is a web-based SaaS CAD tool that is free for non-commercial use and is what I chose to use.  The designers of the Darkaero use this software, which is what brought it to my attention.  And the software didn't require me to use any math!

I liked the design of Heinrich's on VAF so sought to follow that.  He kindly sent me his CAD files, but the cooler he used was too different than mine to make the design adaptable.  So going at it solo, below is the prototype #1 design, printed and shipped for just $16 in PLA nylon with an excessive 1/8" wall thickness.  I wasn't yet ready for printing it in a more expensive heat tolerant material since I still hadn't learned how to add vanes to the design.  As an aside, Craftcloud is a great site to use to get quotes on fabrication in a myriad of materials; the Darkaero guys use Xometry.

For this prototype #1 shown below, the angle of the inlet was too aggressive.  And I couldn't place the cooler in such a way that the inlet could clear the engine mount without the oil cooler needing to be so far outboard that it also wouldn't impinge on the side wall of the lower cowl.  On the right image, you can see how one of the arms from the engine mount runs across the upper corner of the duct inlet, despite the inlet being far offset outboard.

For prototype #2, shown in the series of four images below, I changed the angle and length of the inlet and I added the internal vanes (the latter thanks to some very valuable input from a couple of proficient OnShape users).  I had this printed in MJF PA12 nylon (for $116 shipped) with a reduced 1/16" wall thickness (though the flange to the cooler was printed with 1/8" thickness).  This material has perhaps a reasonable heat tolerance for the milieu in which it is intended to reside (HDT at 0.45 MPa of 175°C which is 347 °F).  

This duct design demonstrated that I could 1) ensure the cooler won't impinge on the side of lower cowl and 2) have the duct clear the engine mount so that the SCEET tube could be connected.  This second point was true only if I offset the inlet fully outboard, rather than keeping it centered over the cooler, thus informing prototype #3's required design modifications.

The last picture in the series below shows that the duct inlet indeed has to be on the outboard side, rather than centered, for the SCEET tube to connect to the inlet, as the duct is offset outboard to accommodate the older SCAT tubing without the engine mount interfering.

So the next prototype needed to have the outboard side of the inlet duct be lined up with the outboard side of the oil cooler screen edge.  Using that approach, this final prototype #3 I had printed for $15 shipped in PLA nylon at 1/16" thickness (and 1/8" thickness along the flange to the cooler).  As visible below, this prototype has the inlet positioned outboard.

A 30-second video shows the design rotating through all dimensions.

With the oil cooler positioned so that it clears the side of bottom cowl and the engine mount and with the duct aligned with the cooler, this final design showed that it too can clear the engine mount, thus the SCEET tube can make the connection properly.  In its new home, the larger cooler now sits a bit further outboard and forward than the original cooler.

Using prototype #3 and the original SCAT tubing, here is the setup verifying fitment.

Since I'm not schooled in materials science, I decided to forgo printing the final design in MJF PA12 nylon and instead went with printing it in aluminum for $314 shipped. I maintained the 1/16" thickness and made the cooler mating flange 1/8".  However these thicknesses were excessive.  The duct is incredibly stout.  It weighs almost exactly 16 ounces.  1/32" all around would be far more appropriate (and cheaper:  $168 shipped in aluminum and $77 in MJF PA12 - nearly half the price).  

Below is the final part before priming.  It was tumbled with glass beads for deburring following printing, so no prep was needed before priming.  For fun, printing the 1/32" thickness version in other materials:  $540k for gold, $5.9k for copper, $4.9k for 316L steel, $4.7k for silver, $2.2k for inconel, $1.3k for glass fiber PEEK and $760 for titanium.

Installation of the aluminum duct following priming, complete with P-seals and RTV application is shown below. Not readily visible in these images is the use of double Adel clamps on the engine mount to isolate the oil input hose from contacting the mount.  That was necessary due to the mild rerouting of that hose.

The duct probably acts as a small heat sink given how much metal it has.  A friend of mine jokingly suggested to add cooling fins to the exterior of the duct.  That may not be a bad idea! 

Below is the complete design finally installed, using SCEET tubing from item 4. Following a ground run with the cowl off to ensure there were no leaks or other issues, the cowl was put back on the airframe for flight.

As shown in the image at the top of this post and shown again below, this design has the SCEET tubing taking a less aggressive path into the cooler, which presumably helps reduce air turbulence into and through the cooler.  And since there is a nice transition from the round SCEET tubing to the rectangular cooler, rather than the forceful "square peg into a round hole" approach of Vans' design,  presumably that too is beneficial.  

As originally shown at the top of this post, Vans' method is more utilitarian as shown more specifically below with their duct from my stock cooler on the left and my design for the larger cooler on the right.  Perhaps Vans' duct reduces the efficacy of the oil cooler due to increased turbulence whereas perhaps my duct increases cooler efficacy due to decreased turbulence.

Back of napkin cost comparison analysis, as of 29-May-24 (appreciating that deletions from kits are no longer available): 

• Vans' oil cooler inlet: $126.  Tray parts: $30.  Total:  $156.    
• My duct design at:  
• $113 for MJF PA12 nylon 1/16" thickness
• $161 for aluminum 1/32" thickness. 
• Larger oil cooler:  $84 more than smaller (per Aircraft Spruce).
• Additional incremental cost for my design over Vans' (includes duct and cooler, no tray):
  
• MJF PA12 print:  $41.
• Aluminum print  $89.
• Optional costs:
• Aluminum for fabricating larger tray (if you don't have scrap):  $25.
  
• Using SCEET:  $15 additional incremental over SCAT.
  
• Using all six bolt holes on the cooler rather than four:  $10.
• 6061T6 0.058"x5/16" aluminum tube.
• Two additional K1000-3 nutplates.

In the future, I may design and print a transition piece from the baffle so the SCEET tubing is completely straight.  This will avoid the scrunching accordion effect as the bottom side of the tubing bunches up on the inside radius where it turns to connect to the rear baffle.

So that was a lot of  thought and effort over 1.5 months from start to finish.  Most of that time was spent waiting for the various 3D designs to be printed, shipped and received.  However, having this pursuit stretched out over time gave me the headspace to consider things more completely.  

Did it solve the high oil temp issue?  Yes!  I'm seeing a 15-25 degree decrease thus far for similar flight conditions.  See below comparison taken from cruise portion of flights (close-out was in place in all flights, with ramp flap open).

• Original cooler:  11-Jul-23: 217 OilT, 79 OAT, 16 MAP, 101 IAS, CHTs ~330-350.
  
• Larger cooler:  5-Jun-24:  198 OilT, 79 OAT, 16.8 MAP, 105 IAS, CHTs ~316-348.

• Original cooler:  6-Sep-22:  225 OilT, 83 OAT, 19.3 MAP, 130 IAS, CHTs 335-365.
• Larger cooler:  11-Jul-24:  200 OilT, 87 OAT, 18.1 MAP, 111 IAS, CHTs 321-352.
  

Here is an example of departing with the ramp flap closed then opening the flap in cruise (x-axis is time since start in minutes), from 11-Jun-24.  The OilT hit 220 in cruise with 73 OAT.  At that point I opened the ramp flap, indicated by the red arrow.  OilT dropped 25 degrees in about 10 minutes, with no changes in power and atmospheric conditions!  Prior to the oil cooler mod, the ramp flap offered only 2 to 3 degrees in OilT reduction.  For the curious, the CHTs were 336-376 until the ramp flap was opened, then dropped about 25 to 30 degrees across the board too.

Design for Stock Airplanes with 5" and 4" SCAT

I completed duct designs to fit the stock AS-2006X cooler for the 5" and 4" SCAT diameters. The 5" design is shown below (the left pic is a view looking from front).  The duct is offset an additional 0.5" outboard to help clear the engine mount (it has been verified that the AN-41A bolts can still be placed through the cooler and into the tray).  It's worth noting that two pilots have tried this design in flight and report no significant change in OilT, though their planes may have other aggravating factors.  So it's still unclear if this duct can help reduce OilT on stock planes.

Here is the design for the 4" SCAT for those aircraft with the earlier design.  In this case, the SCAT flange can be located in the center of the duct since its diameter is smaller.  Though I have tested fitment of this design with a prototype print, no one has yet tested it in flight to determine its effect on OilT, if any.

CAD File Download for Printing Your Own

Are you interested in using my design for the larger cooler?  I'm happy to share my work for free to the community!  All models use 0.04" wall thickness and 0.08" flange thickness.  I highly suggest printing in aluminum.  If you use MJF, be sure not to tighten the SCAT/SCEET hose clamp too tight.  Click the clinks below to get the files (use the download link on the upper right once the new page opens, you don't need to request sharing access).

• 5" duct for larger AS-2007X cooler
• 5" duct for stock AS-2006X cooler
  
• 4" duct for stock AS-2006X cooler

The zip files include the models in both STEP and Parasolid formats.  You can adjust the wall thickness by looking at the variables in the model, but be aware that the fillets may break if the wall thickness is changed so you may need to tweak those (upload the .x_t files to OnShape to edit).  The models are ready for printing as-is without modification (Craftcloud is great for printing - they use .step file format).  Bolt holes should be final-drilled with a #12 bit.

You'll need to fabricate your own cooler tray if you want to use the larger AS-2007X cooler.  I thought about creating a CAD design for the larger tray, but what I fabricated is sufficiently adequate and functional.  A CAD design for the tray might be a challenge since you'd have to locate all the features precisely in addition to having to properly place the bends after fabrication.  It's probably easier to just manually fabricate the tray as I did.

So, now that I have 3D CAD design skills, I wonder what else I can make for the plane...

Duct "Analysis"

After the larger cooler and duct were installed and I had flown the plane with successful results, I wondered if I could have further optimized the design of my duct in retrospect. Note:  The following thoughts are not from someone with any background relevant to the topics presented.

 

If you think of the duct as two ducts (the inner duct suspended by the vanes and the outer duct which is the body/shell of the part), you can consider the cross-sectional areas of the duct on the SCAT flange and the oil cooler flange.  E.g., on the image below, the green area is the outlet of the inner duct and the red area is the outlet of the outer duct.  For the inlet side of this image below, the inner 3" circle would be the inner duct, and the 5" circle (minus the inner 3" circle) would be the outer duct.

 

If you calculate the four surface areas of the ducts (inlet inner, inlet outer, outlet inner, outlet outer), you get the following data for all three of my duct designs.  Then if you look at the ratio of the surface areas of the outer vs. inner duct for both sides of the duct (duct and cooler flanges), you get the numbers also shown in the table.  The ratio is much bigger on the outer duct.

 

I don't know what any of that means since whatever it was I went to school for, it wasn't aeronautical engineering.  But, my uninformed brain says that the pressure in the inner duct is significantly higher than in the outer duct since the outer duct has comparatively more air being pushed through a larger surface area on exit side.  What does that mean for cooling efficiency?  I don't know.  Perhaps I got lucky in choosing the dimensions I did for the larger cooler duct.  If that's the case, I hope I was also lucky for the ducts I designed for the stock coolers.  Regardless, the areas for both ducts are larger on the exit side so this should help move air through the cooler.

Taking an optimistic (an admittedly uninformed) perspective, perhaps the comparatively larger pressure in the inner duct helps increase cooler efficiency by moving more air through that part of the cooler.  And maybe the comparatively lower pressure on the outer duct also helps move air through and out the cowl (since there would be two separate airflows exiting the cooler, with the inner airflow having a higher velocity than the outer airflow).  I wonder what would be different if I made the entry and exit areas have the same ratios for each duct.  In any case, Vans' duct seems like a design that would create turbulence, so perhaps the dimensions of the inner duct in my design isn't critical when compared against Vans' duct.