What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

A Dan Horton question ?

Larry DeCamp

Well Known Member
How much air to feed an oil cooler. There are many comments on OT control and 13 row oil coolers. I returned my standard issue 7 row to Van and will purchase AFS 10 Row for my parallel 0360. A 10 row face area is only a 2 3/4"D supply from the plenum. A 13Row face area only equals a 3"D supply.
OK, there may be some loss in turns and friction and there is a lot of reference to 4" tubing. If you tapped the plenum with nice curved entry and supplied the cooler with a "Dave Anders" 10* divergence, how do you decide how big the hole should be in the plenum ?? Dan has talked a lot about pressure differential etc. , any suggestions would be appreciated and hopefully put a pencil to the subjective approaches referenced in the archives.
 
I just Google'd "air flow through oil cooler" and was astounded by the number of technical papers out there on the subject.
 
Hi Larry, I'm sure Dan will be along shortly with enough science to satisfy your need but I will give you my recent experience with the oil cooler inlet. I have a 15 row cooler mounted on the firewall fed by a 4" opening in the baffles. I have found that I can "neck" the flow down to 3" without any performance degradation in flight.

Here is the new cooler fed by a 4" opening that has been reduced to 3.5" with a 3D printed insert.

20180621_000932091_iOS.jpg


Here is the same 4" opening that has been reduced down to 3" with another 3D printed insert.

20180621_003803374_iOS.jpg


So far I have seen no difference in temps for any in-flight phase. However, I did see my temps at idle go up. What I can't tell you with certainty is whether or not a 3.5 or 4" opening would have done any better at idle.
 
How much air to feed an oil cooler.

As much as required. Seriously...consult the data sheets. Heat transfer capacity rises with mass flow.

The SW charts tell you how much mass flow you get for a given pressure drop, but those numbers are measured at the faces of the cooler. Additional restriction due to a too-small duct would mean you could not rely on the deltaP between the plenum and the lower cowl volume.

There are many comments on OT control and 13 row oil coolers. I returned my standard issue 7 row to Van and will purchase AFS 10 Row for my parallel 0360. A 10 row face area is only a 2 3/4"D supply from the plenum. A 13Row face area only equals a 3"D supply.

First, if I got one of my many wishes, folks would stop describing heat exchangers by number of rows, and start using manufacturer and model. Rows tell us nothing, even from the same manufacturer. Offhand example; an SW-type 8432 and a 10599 are the same size, but are not the same capacity.

Area; might want to recehck your math. The 8432 and 10599 are roughly 4" x 6" face area, or 24 sq inches; I think they are 10 row, but I don't know for sure. A 4" diameter tube has an area of 12.56 sq inches.

(hint...the numbers you posted as diameter were radius. The tube diameters would be twice as large to equal the exchanger face areas you mentioned)
 
Last edited:
Thanks Dan, will check my math

Anyway, I just talked to air flow Performance and they said experience suggests use a 4" supply to assure there is more than enough. Logically. the cooler performance potential is determined by the static pressure on the face, so if you can get a 4" hole in the plenum, why not ? Thanks to all for your comments.
 
First, if I got one of my many wishes, folks would quite describing heat exchangers by number of rows, and start using manufacturer and model. Rows tell us nothing, even from the same manufacturer. Offhand example; an SW-type 8432 and a 10599 are the same size, but are not the same capacity.

Absolutely. And manufacturers should advertise their coolers by stating heat rejection at temperature differential and mass flow! A simple set of curves could be generated to assist in sizing. Why not give us the data like heat transfer coefficients to make selection simpler?
 
No matter the size, you can bet that those adapters Darin printed helped flow into the SCAT tubing a bit. They look a lot like a properly designed intake runner, inside a plenum. Almost makes one wonder whether the reduced turbulence into the SCAT made up for the smaller passage.
 
It is a common myth that the smallest restriction somehow sets the flow rate for the whole system. The transition shapes are probably much more important than some minimum section area, particularly in short ducts like these oil cooler ones. The restrictions shown in one of the posts might make the system equivalent to a much smaller, but well transitioned, hose. The transition back to the full rectangular face of the cooler is perhaps more important than the inlet transition.
 
Transitions are brutally important in flow systems. A basic rule of thumb for ducts carrying gases at near atmospheric pressure is never go more than a 30° included angle on an expansion in a duct (15° per side), and never more than 45° (22.5° per side) on a contraction.

This of course gets modified for curved ducts, elbows (which greatly benefit from turning vanes), and odd shape transitions.
 
The data I found for diffusers say max expansion is actually less than half that; 7 degrees from streamline. And for wedge diffusers (where the stream enters the diffuser at an acute angle to the face of the heat exchanger), one of my mentors told me (and I later verified with testing) that the far end of the diffuser needs to be pinched down radically, to preserve equal flow across the face of the heat exchanger. I haven't tested a typical a/c oil cooler, but I have checked a radiator. Unless the far end pinches down so it almost touches the core, virtually all the flow is through the last 1/3 of the core.
 
The data I found for diffusers say max expansion is actually less than half that; 7 degrees from streamline. And for wedge diffusers (where the stream enters the diffuser at an acute angle to the face of the heat exchanger), one of my mentors told me (and I later verified with testing) that the far end of the diffuser needs to be pinched down radically, to preserve equal flow across the face of the heat exchanger. I haven't tested a typical a/c oil cooler, but I have checked a radiator. Unless the far end pinches down so it almost touches the core, virtually all the flow is through the last 1/3 of the core.

Yeah, the numbers I quoted are rules of thumb for large industrial ducts (like 20 or 30 feet diameter) where smoother transistions get to be so large that they are not economically justifiable. I probably should have mentioned that, and that smaller ducts can benefit from even more gradual transitions at lower cost.
 
Look at the Anequim project for an optimally shape oil cooler duct. You'll notice it is long with gentle transitions on the inlet AND exit sides of the HX.

We have enough good flying examples of efficient duct designs to know how to do it, the problem is that space is restricted in a typical RV. You can't build something as good as what a clean sheet design like the Anequim has.

Guide vanes are your friend when you have large divergent angles and short distances to the HX face. People seem to have an aversion to them though, preferring to use the crossed fingers method in their designs, hoping the large, square HX face will be fully wetted over a short distance by a 3 or 4 inch round hose feeding it, sans separation. Likely it won't, from my extensive tuft testing.

Cooling is one aspect, but we'd also like to minimize momentum loss (drag) while doing it.
 
Last edited:
...but we'd also like to minimize momentum loss (drag) while doing it.
In that case, shouldn't we use something a heck of a lot less draggy than scat tubing between the plenum and the cooler?
 
In that case, shouldn't we use something a heck of a lot less draggy than scat tubing between the plenum and the cooler?

It depends on the flow regime (laminar or turbulent). Our short runs to the oil cooler aren't long enough in all likelihood to develop the flow regime, so it's probably mostly just chaotic...
 
Wiring

Think I would worry more about the wiring laying across the motor mount and it appears one wiring clamp has some not protected by the clamps plastic cover too. Just something I noticed from the picture in the background.
 
Back
Top