My first attempt at sending this was too large; so I eliminated attach. 3. Re-number accordingly.

 

Subject: FW: Oil Cooler re-visit

 

 

Some of you may recall that I brought up the issue of marginal oil cooling in my Velocity 20B last fall after finding that I was limited in climb rate in the warmer summer weather.  I continued to enjoy flying during the cooler winter months, and it hasn’t been much of an issue.  Now as we approach summer again, I intend to institute some changes.  Before doing so I wanted some further data on the air flow.

 

The cooler is in the wing root; behind the strake, behind the spar. Inlet is under the wing behind the main gear door cover. The cross section of the inlet is shown in the Attachmt 1.  The sketch shows the original shape when pressures A (ram, in front of scoop), B and C (static) were made.  I then modified the upper profile as shown, and sealed all the gaps around the cooler, and measured the pressure at D. Cooling improved very slightly.

 

The exit above the wing is shown in attachmt 2.  A pressure measurement at the mid point of the outlet was about -1” of water – somewhat disappointing since the theory was that there would be lower pressure there.

 

I had concluded that the airflow in front of the scoop was being somewhat disrupted by the gear door, which does have a bump on it for the wheel toward the out-board portion of the entry scoop.

 

Paul Lamar was kind enough to ship his multi-manometer setup (attach 3) to me so I could get some more data.  One challenge was to find ways to get many tubes from the measuring point to the manometers. Some 4-wide color coded clear 1/8” ribbon tube from McMaster-Carr helped facilitate that. I then fabricated a couple of velocity rakes (attach 4, also one from 1/8” copper tubes), and we measured 8 points at one time. The tubes are at 5/16” spacing.

 

The wide inlet scoop (23”) is divided into three sections. We placed one rake at the middle of the in-board 1/3, and one at the mid point of the outer 1/3 (this one behind the bump in the gear door). We then took measurements at 120 KIAS climb (138 mph), and at 150 (173 mph) level. The results for 150 KIAS are shown on the upper chart in attach 5. I penciled in the curve to the 5 points including 0 at the surface. (I should point out that the measurements were made on the opposite wing which has a matching radiator installation, and still has the original non scoop inlet, in order to facilitate running the tubes.  The location behind the gear door is the same.)

 

The result shows a fairly normal boundary layer profile; at both positions.  The low velocity portion is less than ½” thick which is consistent with that computed by a little BL program that I have. The scoop extends into the airstream about 1 ¼”, at which point ther is essentially free stream velocity. What does it tell me?

1. My assumption that the BL was somewhat mixed and turbulent was not correct.

2. Segmenting the profile and determining an average velocity (about 135 mph) would give me very close to the desired air flow rate IF the scoop were operating at 90% efficiency.

It was also interesting to find the profile shape changed little with the planes airspeed, however; the profile at the lower speed was more like 20 mph less, when the planes speed was 35 mph less – indicating a slightly higher pressure under the wing at the higher AOA.

 

OK, that was fun. Next we set up to measure the air velocity profile at the exit of the core.  This was done by positioning 1/8” copper tubes as ‘pitot’ tubes close to the surface at different positions across the core, near the center of it’s 22” length.  That result is shown in the lower chart in Attach 5.

 

Clearly the flow is not very uniform, being quite highly peaked toward the center. This of course indicates the ‘diffuser’ is not being effective.  Again, by segmenting the profile, and adjusting to the actual flow area (minus the tube area, but not accounting for the fins), I could compute the approximate total flow rate through the core.  It is about 68% of the potential inflow for an effective scoop- telling me that some amount of air is flowing around (under) the scoop.

 

Two other relavent pieces of info: we measured the static pressure on the upper surface outboard of the outlet fairing, about 1/2” off the surface. It was consistent with the earlier measurement behind the exit fairing (about -1”).  We also tufted the back edge of the exit fairing with some 4” strings, and noted that they went straight back – indicating the air cleanly detaching at the edge, and no swirling down into the exit stream. These things suggest that the exit fairing is not inhibiting the flow. It is also not enhancing it, but does provide protection for the core – and looks pretty coolJ. We did not tuft the surface of the wing behind the fairing because that is not visible from the cabin.  We also did not use longer tufts because these were the longest we could scavenge from the dust mop, since the tufting was an afterthought and we did not come equippedJ.

 

So-o-o-o; before going into these measurements, I had pretty much decided to install an oil/coolant heat exchanger since my coolant temps run 20 – 40F lower than the oil, and have some margin.  That is a fairly major mod to the plumbing.  Trying to modify the air inlet to get more effective diffusion is also major, as it requires removal of the wing and the oil cooler.  This would require some reshaping, and a baffle(s) the width of the scoop to confine the slow BL to the upper portion of the inlet.  And ‘how’ effective it could be made within the confines of the wing root, and ingesting the BL, and without adding drag - is uncertain.

 

Therein lies the dilemma.

 

BTW; the placement of these coolers in the wing root in this manner was based on the testimony of another Velocity builder you had placed his Lyc oil cooler there, and said it worked great – even without a scoop – because of the pressure differential above and below the wing.  I don’t know why his experience was so different, but it does not surprise me that there isn’t much differential since the stake area on a canard airplane is basically neutral in level flight.

 

Anybody read this without falling asleep?

 

Al