Modifications: Bigger oil cooler for IO-390-EXP119, try #3 (sastifactory performance).

TLDR:  Oil temp was too high with EXP119 engine.  Installed larger 2007X, which failed in flight.  Tried 8001652 which was inadequate.  Tried 8001643 with more deliberately designed duct and oil cooler tray with integrated exit diffuser.  Oil temp problem resolved.  Read on to learn all the details.  

If you're interested in using my designs for your plane, please contact me.  I will happily share them with you!  I can even add your embossed N-number to the flange of the duct.  I also now have step-by-step instructions on how to get the parts fabricated, prepped and installed. I share my work for free in support of the RV-14 community. 

My "larger oil cooler" series:

1. Initial larger oil cooler, 2007X.  No longer in use.
• Airflow Systems 2007X oil cooler in-flight failure after 75.9 hours.
2. Second larger oil cooler, Aero-Classics 8001652, which was inadequate.  Not in use.
3. Third larger oil cooler (the post you're reading now), Aero-Classics 8001643, currently in use.

Recall that I increased the size of my cooler from the 2006X to the 2007X since my oil temperature was prohibitively high in the summer.  But then my Airflow Systems oil cooler failed in flight.  So for try #2, I decided to go with an Aero-Classics oil cooler that uses a different manufacturing process (KitPlanes has a great article about different cooler designs).  That cooler was inadequate.  So in this post I outline try #3, which uses an even larger cooler by Aero-Classics.

First, I show the final result of this third try at a larger oil cooler installation below.  Afterwards, I'll outline my installation approach.  For all of the details on why a larger oil cooler was needed and more explanation about my duct design, see the post that outlined the original larger cooler's installation.  To learn about my second attempt at a larger cooler, go here to read that post.

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

Here is a comparison of the coolers I've had in my aircraft, with the dimensions of the respective cores delineated.  The images are scaled and aligned relative to the center of the upper mounting holes.  The stock 2006X provided inadequate heat rejection.  The 2007X was adequate however failed in flight.  Given my disappointing experience with Airflow Systems in trying to learn why it failed, I found myself no longer comfortable flying behind their products.  So I contacted Aero-Classics, which has coolers that use a different manufacturing design ("bar-and-plate" vs. "drawn cup"), such that they wouldn't fail in the same was as my failed 2007X did.  Aero-Classics suggested the 8001652, which turned out to be inadequate.  I then upgraded to their 8001643 ($1,407.68, shipped) which thankfully was adequate.  The fins are very fragile and can be bent with the slightest touch.  Keep a toothpick handy to straighten out any fins that get upset.  The rows are also wider than the Airflow Systems coolers.

Comparing the dimensions of the above four coolers' heat rejection cores yields the following chart.  You can see that the 2007X is significantly larger than the 2006X, the 8001652 has a larger cross-sectional area and less volume than the 2007X and finally the 8001642 is significantly larger than the 2007X.  However, there is more to the heat rejection capacity than just dimensions.  Despite that, the below provides some comparative value when assessing coolers.

 

When considering the explanations below, please understand that I have no skills related to aerodynamics (or really, skills of any value whatsoever!).  I'm just a person flailing about in a world he doesn't understand.  Having indicted and disclaimed myself, let's get on with the details.

Let's take a look at the dimensions of the 8001642.  It's big.  I needed to design a new duct and mounting tray for this cooler.  Those endeavors are outlined next.

Below is a CAD render of my duct design when viewed from its top.  For this third oil cooler attempt, I decided to try three things in addition to the much larger oil cooler.  First, 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 ducts on the intake side (the round SCAT flange) and the exit side (the rectangular oil cooler flange).  

I wanted my duct design to have the cross-sectional areas of the inner and outer ducts on respective sides the same.  This is illustrated below (the shapes are properly scaled relative to each other).  In other words, 'area of dark blue=area of light blue' and 'area of dark purple=area of light purple' (where AOT=area outer top, AIT=area inner top, AOB=area outer bottom, AIB=area inner bottom).  The intent here was to help ensure that air coming over the face of the cooler is evenly distributed to help with heat rejection efficiency.  Update:  I later learned from my friend who studied CFD in grad school that my approach was wrong since the two different airflows are subjected to different parasitic forces due to the different exposed surface areas on the walls of the two ducts.  So it's not balanced like I thought it would be.  However, it seems to work well enough.

Second, I wanted the angle of the duct to be more shallow so the SCEET tube has a more gradual transition along its trajectory.  Thus, I made the intake face of the duct pointed closer to vertical than in my previous design.  Below shows a comparison of trajectories of the SCEET tube for the failed 2007X cooler and the larger 8001643 cooler.  The difference is minor, but visible.  Though I wanted the duct a few degrees more towards the vertical, there is a limit to that angle since the duct will impinge on the engine mount if it points too high.  In the images a below, you can also see that the 8001643 sits further forward than the failed 2007X (which itself was a little forward than the stock 2006X).

 

Third, I wanted a poor man's exit diffuser on the bottom of the oil tray to try to help suck air through the cooler by lowering the exit air pressure.  It was too complicated to design a separate diffuser in CAD or bother with mocking one up in fiberglass.  So I just added little "wings" to the bottom of the oil cooler tray where I could.  Let's be clear:  this represents the fits of an uneducated person grasping for anything, desperate to eek any additional performance.  This is not some wonderfully elegant design leveraging some profound insight.

Also, since the 8001643 was very long/wide, I had to place the cooler more forward and outboard and change the tray's geometry so it wouldn't impinge on the passenger side heater's SCAT tube and engine mount.  This required me to add some flanges on the tray and flip them to the top side of the tray rather than the bottom.  I also tilted the tray higher than stock to try to help air bubbles find their way out of the cooler.  Below are two opposing views of the new oil cooler tray.

Aero-Classics sent me a CAD model of their cooler so I could mock my designs up.  Here's what the stack would look like.  Curiously, Van's has a blanket policy of not sharing any CAD files (I had asked for the right half of the engine mount and the right aft engine baffling CAD files).  It would have made things easier for me with those data, but I could conclude my design without them.

I had the tray parts fabricated by SendCutSend in 0.063" 5052 aluminum for $75.57, shipped.  Following deburring, priming and riveting, here's what it looks like on the airframe.  The holes for the mounting bolts that attach to the engine mount were match drilled into the tray's side angles.  Because the tray is so much longer than the stock tray, a spacer (0.3125"x0.058" T6 aluminum tube) and angle were used to capture the inboard mounting tab on the engine mount.  I neglected to measure the spacer after I fabricated it, but it's about 1.5".  I got the tray (side-to-side) level to 0.1° degrees.  The tray is very stout and stable.

As a reminder, always use steel fittings on aluminum to prevent galling (AN822-8), despite what the Van's plans say.  Here is what the clocked fittings look like on the cooler (when installed, this would be the inboard side of the cooler).  The clocking was chosen to help the cooler stay in position on the tray when the hoses are connected, seeking to have the forces imposed by the hoses balanced out as much as possible.  I.e., the cooler sits neutrally in position when the hoses are connected even when the cooler isn't bolted down to the tray.

As I discussed in the post outlining the failure of the Airflow Systems 2007X oil cooler, one of the items I was concerned about was the angle that the intake hose took to come in to the cooler.  Vibration from that hose may have exposed a vulnerable aspect of the "drawn cup" oil cooler design.  So I had FF-00020-1 fabricated at an inch shorter ($204.13 shipped from Aircraft Specialty) to account for a larger cooler sitting further inboard than the stock cooler position.  This would presumably reduce the torquing of the fitting by the hose.

 

Here is the cooler placed with the p-seals glued on and curing (drill bits are used during curing to ensure the seals adhere without curvature).  It's a big cooler.

I had my duct design printed in 0.04" aluminum using Craftcloud for $188.24 shipped (with an additional $26.56 tariff that I paid, which I've been assured by our Dear Leader is somehow not a tax).  Usually I receive really high quality parts from the manufacturer, but this time, they (ProtoTi) didn't do as good a job.  Structurally it's perfectly usable, it just has some inconsistencies in its surface and shape.

 

As an aside, I marvel at the world we live in:  A dumb guy (me) off the street with no relevant skills spins his computer mouse around, clicking in a few spots and creates a model using software he has no chance in writing on his own.  Some factory in China uses aluminum powder and a laser to print (print!) the design.  A logistics armada brings it to my doorstep:  A plane carries it over the ocean.  A tariff is assessed for me to pay (which is somehow not a tax - magic!).  Trucks cart it over land.  Some dude literally throws it on my porch and I don my slippers to retrieve it.  Crazy. 

 

Here is everything assembled, including using SCEET tubing rather that SCAT, to help reduce turbulence in the duct.

 

This cooler provides satisfactory performance.  OilT is almost always 195 in level flight at a range of summer OATs.  I have tested it between 80 - 95 OAT in flight.  Example below from a hot day.  The x-axis shows minutes from engine start.

 

 

 

I did start work on a transition piece that connects to the back of the engine baffling.  My intent was to help further fair the air by smoothing the curvature down to the oil cooler duct.  Given how good my OilT is now, I don't think this is a worthwhile endeavor.  Here's what it looks like in CAD both separately and as part of the whole assembly (the latter shown without the SCEET tubing).

 

 

Below is a cost breakdown of the parts that are specific to my design and the stock design.  It's ~$450 more in incremental cost to use this larger cooler than the stock cooler.  The primary driver of this cost difference is the Aero-Classics cooler (which represents the difference in manufacturing technique for Aero-Classics' "bar-and-plate" design).  This is ~5x higher than the $89 incremental cost to use the failed 2007X cooler (costs outlined for the failed 2007X cooler towards the bottom of this post).

 

If you're interested in using my designs for your plane, please contact me.  I will happily share them with you!  I can even add your embossed N-number to the flange of the duct.  I also now have step-by-step instructions on how to get the parts fabricated, prepped and installed. I share my work for free in support of the RV-14 community.

 

Aero-Classics 8001643:

• Cooler:   $1,407.68
• Tray: $75.57 +  $3.42 (nutplates)
• Duct:  $188.24 + $26.56 (tariff that I paid which Dear Leader says is not a tax that I paid)
• Hose: $204.13
• Total:   $1,905.60 

 Van's 2006X (as of 27-Jul-25):

• Cooler:  $940.82 (EA00002/2006X)
• Tray parts:   $35.40 (FF-01404A, FF-01404)
• Duct:  $135.25 (FF-01406E, FF-01406C)
• Hose:   $258.24 (FF-00020-1)
• Total:   $1,369.71

Modifications: Bigger oil cooler for IO-390-EXP119, try #2 (inadequate performance).

TLDR:  Oil temp was too high with EXP119 engine.  Installed larger 2007X, which failed in flight.  So I tried the Aero-Classics 8001652.  Read on to learn all the details, but keep in mind that the cooler described in this post was inadequate.  See the three other posts below, including the last post where I successfully tried the larger 8001643 with a more deliberately designed duct and oil cooler tray with integrated exit diffuser.

My "larger oil cooler" series:

1. Initial larger oil cooler, 2007X.  No longer in use.
• Airflow Systems 2007X oil cooler in-flight failure after 75.9 hours.
2. Second larger oil cooler, Aero-Classics 8001652, which was inadequate (the post you're reading now).  Not in use.
3. Third larger oil cooler, Aero-Classics 8001643, which is satisfactory.  Currently in use.

Recall that I increased the size of my cooler from the 2006X to the 2007X since my oil temperature was prohibitively too high in the summer.  But why is this my second attempt at a larger cooler?  Because my "Airflow Systems oil cooler failed in flight".  So for try #2, I decided to use a different manufacturer's oil cooler.  

But first, the final result of this second larger oil cooler installation is shown below.  Then I'll get into some of the details.  For all of the details on why a larger oil cooler was needed, see the post that outlined the first larger cooler's installation.

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

Now let's get into some of the details on this second larger cooler installation.

I spoke with Aero Classics and opted to try the HE 8001652 (the "HE" is for "high efficiency"), a PMA'd cooler with a core dimension of 6"x6.29" and depth of 3.2" for $1,285.  Aero Classic's sentiment was that I should see a 10 - 15% increase in performance over the 2007X.  Even if it was 10% - 15% less, I was confident it would be sufficient for my needs.  The Aero Classics cooler has 3.5 square inches more cross-sectional area across the core (10% more) and about 3 cubic inches less volume in the core (2.5% less).  The Aero Classics engineer also believes their cooler has a better sealing interface to mitigate leakage around the core.

You can read more about the selection of that cooler and a poor-man's comparison of the 8001652 against the failed 2007X cooler towards the bottom of this post.

Here is the HE 8001652 (silver) next to the failed 2007X (gold).  Immediately, you can see difference between Airflow Systems' "drawn cup" design on their failed cooler and the "bar and plate" design on the Aero Classics cooler.  If the Aero Classics cooler fails, it won't be from a seam ripping open like what happened on the failed Airflow Systems cooler.  KitPlanes has a great article about different cooler designs.  The fins are very fragile and can be bent with the slightest touch.  Keep a toothpick handy to straighten out any fins that get upset.  The rows are also wider than the failed 2007X.

Since the HE 8001652 has slightly different dimensions than the failed 2007X, I needed to update my OnShape CAD models for the tray and duct.  Here is the model of all three (tray, cooler, duct) stacked together.  I ultimately removed the hole on the right support angle in the model in favor of match drilling.

Here's the new tray in place during final fitment.  Cost was $61.58 shipped from SendCutSend.  I then primed and riveted it together.  The image on the left shows initial placement.  The image on the right show how I setup to match drill the right support angle:  First, a 1.5" angle was fabricated from 0.063" 5052 aluminum and a #12 hole drilled on one side.  Next, a 0.058"x0.3125" T6 aluminum was cut to 0.6" to fabricate the spacer.  An NAS1149F0363P and NAS1149F0463P washer were place on either side of the spacer.  A #12 bit was passed through the hole and twisted enough to score the right support angle so that the hole could be then precisely located and drilled.

Here is the final try in place.  I got it (side-to-side) level to 0.4° degrees.

Here is the cooler in place with p-seals adhered.  Note that the outflow hose is shown fastened to the engine mount with two Adel clamps.  I ultimately removed those to ensure that the hose can offer up its slack when the engine moves during startup and shutdown.  You can also see the intake hose coming straight in to the cooler behind an engine mount support.  This was a custom length hose I had fabricated to try to mitigate vibration being transmitted into the cooler.  I discarded this approach as this hose had no slack to accommodate engine shake on startup and shutdown (see video in this post that shows just how significant this shaking is).

Below is the aluminum duct prior to priming for installation.   On my original design for the failed Airflow Systems 2007X cooler, I had the duct printed at 0.0625" thickness with the cooler flange at 0.125".  That was overkill and cost $314 shipped last year from CraftCloud.  The time around, I had the duct printed for this cooler at 0.04" thickness and 0.0625" at the flange.  This was $172.48 shipped, again using CraftCloud.  

As I discussed in the post outlining the failure of the Airflow Systems 2007X oil cooler, one of the items I was concerned about was the angle that the intake hose took to come in to the cooler.  Vibration from that hose may have exposed a vulnerable part of the "drawn cup" oil cooler design.  So I had FF-00020-1 fabricated at an inch shorter ($204.13 shipped from Aircraft Specialty) to account for the new cooler sitting further inboard than the stock cooler position.  This would reduce the torquing of the fitting by the hose.

The inflow hose fitting is at 90 degrees pointed forward.  As a reminder, always use steel fittings on aluminum to prevent galling (AN822-8), despite what the Van's plans say.  Here is what the fittings look like on the cooler (when installed, this would be the inboard side of the cooler). 

Here I show the intake hose going in to the cooler.

The outflow hose could be rerouted to come in from the top using a 90 degree elbow fitting clocked at 12 degrees clockwise.  With this configuration, the oil cooler won't twist and does not want to translate in any direction when it isn't bolted down.  Thus, the outflow hose is perfectly situated. 

Here is the aluminum duct in place where you can also see the outflow placement.  Again, note that the outflow hose is shown fastened to the engine mount with two Adel clamps, which were ultimately removed.

Final installation shown below.  Maybe this coming winter I'll design a carbon fiber transition duct from the baffle so that a straight section of SCEET tubing can be used.

Turns out that this cooler was horribly inadequate, yielding a cruise OilT of 223°F at an OAT of 85°F and a peak OilT of 230°F at an OAT of 86°F in a low power descent!  So I dispensed with this cooler and set out on my third attempt at finding a satisfactory cooler.  It's worth noting that Aero-Classics would offer me only about $150 in credit to exchange this cooler.  I found that offer really, really disappointing.

Cost comparisons: 

Aero-Classics 8001652:

• Cooler:   $1,285.00
• Tray: $61.58 +  $3.42 (nutplates)
• Duct:  $172.48
• Hose: $204.13
• Total:   $1,726.61

 Van's 2006X (as of 27-Jul-25):

• Cooler:  $940.82 (EA00002/2006X)
• Tray parts:   $35.40 (FF-01404A, FF-01404)
• Duct:  $135.25 (FF-01406E, FF-01406C)
• Hose:   $258.24 (FF-00020-1)
• Total:   $1,369.71

Powerplant: Switched to automotive plugs and harnesses.

My ignition harnesses were getting ratty with fractures in the shield (necessitating crimping a new connector on one line due to excessive radio interference).  Turns out that replacing connectors is not only expensive in its own right, the tool required to do so is crazy expensive.  So I opted instead to switch over to automotive harnesses and plugs.  

I used G3 Ignition's kit.  It includes:

• MSD Super Conductor 8.5 mm wire 
• Right angle MSD part 3311 boots and terminals
• 3 mm diameter, 10 mm long hollow copper rivets 
• 14 mm long, 7-turn conical springs (0.25" to 0.14")
  
• Rubber 16.5 mm OD by 8.5 mm ID, 6 mm long, 4 mm thick grommets

I purchased the MSD "Pro-Crimp" tool, part 35051, a device that takes some practice to use properly.

The MSD Super Conductor wire has a fascinating design.  Rather than being shielded, it uses an RF-choke approach.

First I pulled out the Slick harnesses, noted which hole went to which cylinder then measured the wire lengths.

The harness caps are drilled out to 5/16" then tapped to 3/8" NC.

I then cut the new wires to lengths slightly longer than the original harness wires.  They are screwed in to the magneto caps such that the attached springs (shown later below) sit at the same height as the original springs on the original harness wires.

The springs are screwed on to the rivet heads, then the rivets are pushed into the center conductor of the MSD wires.

The new caps and wires are then placed onto the respective magnetos and the wires are rerouted (I needed to drill out the plastic pass-through flanges that pass the ignition wires through the rear engine baffling - I think to 5/16" or 3/8").  The Adel clamps need to be upsized for the larger diameter of the MSD wires.  And I used additional clamps as well.

Connectors are crimped on (I didn't capture any images of that process), then plugs (in my case, Denso IKH27) are inserted into their adapters (I used E-Mag's "long reach" 18 mm to 14 mm adapters) finger tight.  Anti-seize compound is placed on the adapters only, then the plug nut is torqued to 18 ft-lbs.

Here's the final routing (the last image shows the MAP line to the right magneto loose, before it was ultimately fastened down).

Engine starts a lot easier, it seems to sound throatier (could just be wishful thinking), CHTs and EGTs and performance seem unchanged.  However, I'm looking forward to easier condition inspections going forward.