cooling fans

Homebuilt Aircraft & Kit Plane Forum

Help Support Homebuilt Aircraft & Kit Plane Forum:

pfarber

Well-Known Member
Joined
Feb 21, 2019
Messages
1,031
Location
Dollywood
Doesn't seem that op didn't do any research and will fail without massive alterations.

The general rule is 1 to 2.5 cu/in of rad area per HP. That should be a starting point for a radiator with good clean airflow. This idea of sucking air through the cockpit is a non-starter.

My 200hp motor needs is planned to start with 450-ish cu/in but will have a very clean belly scoop with the proper profile in and out.
 

raymondbird

Well-Known Member
Supporting Member
Joined
Nov 29, 2008
Messages
178
Location
Belleville, Ontario Canada
A careful look at the racer cowls and you will see the outlets are much bigger than stock. Inlets are usually bigger too. Get some dimensions and run the math. I think you will find that the usual cowlings keep the engines adequately cooled for the life expectations at each power and speed.

Look up posts by RV6EJGUY on radiator size and ducting. Ross is the proprietor at SDS, has a long history of running EJ22 Subie on his RV6, and has really useful advice for anyone doing liquid cooling. Russel is one of his customers, runs a 6 cylinder Subie Glasair and wins all sorts of races with it. He has an almost ridiculously small lookin
And from Ross's cooling video #2: "2835 CFM is way more than needed".
He was referring to his engine at 120hp in cruise. Hate to waste you time with my ignorance but that would indicate 300hp doesn't need anywhere close to your estimate of 12000 CFM . . . ?
 

Vigilant1

Well-Known Member
Supporting Member
Joined
Jan 24, 2011
Messages
8,140
Location
US
And from Ross's cooling video #2: "2835 CFM is way more than needed".
He was referring to his engine at 120hp in cruise. Hate to waste you time with my ignorance but that would indicate 300hp doesn't need anywhere close to your estimate of 12000 CFM . . . ?
As I read Billski's estimate of the required airflow, he made it clear that there were plenty of places in his methodology that would depend on the particulars of each situation (radiator efficiency, temp of ambient air, etc). So, I read it as a rough estimate good enough for the present discussion. He wrote:
  • This 300 hp engine will need more like 7500 to 11200 cfm to even have a chance at cooling a 300 hp engine at full trot, like we do in airplanes for the first five minutes of every flight. Better go 12,000 cfm with room to go higher still.
Ross's 2835 cfm for 120 HP = 23.625 cfm per HP.
For your case of 300 HP that would be 7,087 cfm.
Important: if we look at Ross's plane it still has air duct openings at the front of the cowl, and he mentions the hot air leaving the cowl. That's because he has another heat exchanger in there for his oil. That air also circulates around the engine and PSRU and cools it. I'm sure Ross took this other cooling into account when determining that the 2835 cfm through the belly scoop would be more than sufficient.
All that has to be added back into the heat load unless your plane will also have supplementary direct air cooling.
Billski's estimate of 7500-12,000 cfm is 25-40 cfm per HP. Ross's number (23.6) would fit into that window if we added in the additional direct air cooling of his engine surfaces and the oil heat exchanger. FWIW, It all greatly exceeds the 4000cfm / 300hp = 13.3 cfm/HP proposed in your OP.
 
Last edited:

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,293
Location
Saline Michigan
And from Ross's cooling video #2: "2835 CFM is way more than needed".
He was referring to his engine at 120hp in cruise. Hate to waste you time with my ignorance but that would indicate 300hp doesn't need anywhere close to your estimate of 12000 CFM . . . ?
You are being critical of a first trip through the particulars with rough guesses for heat rejection and heat exchanger performance to see if 4000 cfm was even in the ballpark. Even this simple analysis shows a need for a lot more air. Remember if you put up too little air, you can barely fly the bird. You have engine break-in and airframe debugging to do too, so you have to be able to fly the bird.

Do remember that a primary approach to a design search is to cast the net broad enough to include the solution. I did that and my simple pass suggests 4000 cfm is substantially undersize. If an analysis no better than this one is used, 12000 cfm would be conservative. Once flying, you can reduce fan speed until get to thermostat regulation, and know how much air flow you actually need, then modify the design accordingly.

Do a much more "downtown" job of determining the actual amount of waste heat, establishing radiator performance, etc, and you can certainly refine all of these numbers, giving a much closer estimate of how much air is needed. Have at it! I still recommend going big initially so you can actually fly the beast while you break in the engine and work out the kinks.

Billski
 

raymondbird

Well-Known Member
Supporting Member
Joined
Nov 29, 2008
Messages
178
Location
Belleville, Ontario Canada
As I read Billski's estimate of the required airflow, he made it clear that there were plenty of places in his methodology that would depend on the particulars of each situation (radiator efficiency, temp of ambient air, etc). So, I read it as a rough estimate good enough for the present discussion. He wrote:

Ross's 2835 cfm for 120 HP = 23.625 cfm per HP.
For your case of 300 HP that would be 7,087 cfm.
Important: if we look at Ross's plane it still has air duct openings at the front of the cowl, and he mentions the hot air leaving the cowl. That's because he has another heat exchanger in there for his oil. That air also circulates around the engine and PSRU and cools it. I'm sure Ross took this other cooling into account when determining that the 2835 cfm through the belly scoop would be more than sufficient.
All that has to be added back into the heat load unless your plane will also have supplementary direct air cooling.
Billski's estimate of 7500-12,000 cfm is 25-40 cfm per HP. Ross's number (23.6) would fit into that window if we added in the additional direct air cooling of his engine surfaces and the oil heat exchanger. FWIW, It all greatly exceeds the 4000cfm / 300hp = 13.3 cfm/HP proposed in your OP.
Sorry but don't understand where you get "greatly" from? Again, Ross's comment was "way more than needed". The 2835 CFM was his "available cfm" at that particular speed and 120hp setting. Bill's math and assumptions look very accurate so just trying to understand what I'm missing.
 

Vigilant1

Well-Known Member
Supporting Member
Joined
Jan 24, 2011
Messages
8,140
Location
US
Sorry but don't understand where you get "greatly" from? Again, Ross's comment was "way more than needed". The 2835 CFM was his "available cfm" at that particular speed and 120hp setting. Bill's math and assumptions look very accurate so just trying to understand what I'm missing.
Here was your initial setup, from the OP:
Anyone think a ~4000 CFM puller fan and matching large radiator could cool a 6L LS engine ok? With no intake scoop I mean, just cockpit air with open canopy and exiting through louvers in the belly behind the wing.
You want to cool a 300 HP engine with no intake scoop and you are proposing to use a fan pulling 4000 cfm to do all that cooling, no other cooling air from other scoops, etc. That's 13.3 CFM per HP.

Billski estimated (with plenty of clarity on his assumptions and the uncertainties) that it would take 25-40 CFM per minute per HP. You say his method looks "very accurate," and it is double to triple the amount of your estimate in the OP, so I'm saying that it "greatly exceeds" your estimate in the OP. Fair?

Ross's cooling comes from two systems.
1) He has holes in the front of his cowling that push air through his oil heat exchanger and remove heat from his PSRU, engine surfaces. etc. Let's say the two openings are 5" diameter each (I don't know). That's 40 SQ inches (0.28 SQ ft). At 80 mph climb airspeed at SL, they would admit 2000 CFM of air. Let's assume that backpressure/resistance reduces that by 50%, so he's getting 1000 cfm to use for cooling through those ducts.
2) According to Ross's chart, he expects to have 2835 CFM flowing through his coolant radiator, and he believes it is more than enough. We don't know how much more, but it is adequate.
So, Ross's total available cooling airflow in climb is about 1000 + 2835 = 3835 cfm. That is for 120 HP, so at that rate (32 cfm per HP) your proposed 300 HP installation would require 250% more, or about 9600 CFM. Now, interestingly, this 32 cfm per HP is smack dab in the middle of Billski's estimate range of 25 to 40 cfm/HP. It is also well more than double your proposed flow rate of 13.3 cfm/HP.

And that's what I meant when I wrote that both estimates "greatly exceed the 4000cfm / 300hp = 13.3 cfm/HP proposed in your OP."

I'm sure you are not maintaining that Ross's "way more than needed" comment means that in climb his system is more than twice the size it needs to be.
 
Last edited:

raymondbird

Well-Known Member
Supporting Member
Joined
Nov 29, 2008
Messages
178
Location
Belleville, Ontario Canada
Here was your initial setup, from the OP:

You want to cool a 300 HP engine with no intake scoop and you are proposing to use a fan pulling 4000 cfm to do all that cooling, no other cooling air from other scoops, etc. That's 13.3 CFM per HP.

Billski estimated (with plenty of clarity on his assumptions and the uncertainties) that it would take 25-40 CFM per minute per HP. You say his method looks "very accurate," and it is double to triple the amount of your estimate in the OP, so I'm saying that it "greatly exceeds" your estimate in the OP. Fair?

Ross's cooling comes from two systems.
1) He has holes in the front of his cowling that push air through his oil heat exchanger and remove heat from his PSRU, engine surfaces. etc. Let's say the two openings are 5" diameter each (I don't know). That's 40 SQ inches (0.28 SQ ft). At 80 mph climb airspeed at SL, they would admit 2000 CFM of air. Let's assume that backpressure/resistance reduces that by 50%, so he's getting 1000 cfm to use for cooling through those ducts.
2) According to Ross's chart, he expects to have 2835 CFM flowing through his coolant radiator, and he believes it is more than enough. We don't know how much more, but it is adequate.
So, Ross's total available cooling airflow in climb is about 1000 + 2835 = 3835 cfm. That is for 120 HP, so at that rate (32 cfm per HP) your proposed 300 HP installation would require 250% more, or about 9600 CFM. Now, interestingly, this 32 cfm per HP is smack dab in the middle of Billski's estimate range of 25 to 40 cfm/HP. It is also well more than double your proposed flow rate of 13.3 cfm/HP.

And that's what I meant when I wrote that both estimates "greatly exceed the 4000cfm / 300hp = 13.3 cfm/HP proposed in your OP."

I'm sure you are not maintaining that Ross's "way more than needed" comment means that in climb his system is more than twice the size it needs to be.
Fair enough indeed and really appreciate all this time and trouble! Even a 3/4 scale 109 can have numerous cowling scoops and even a very large oil cooler though. Sorry I didn't mention that when quoting Ross's data. Enough said anyway though thanks. I will do more research and testing of course. 4000 cfm is just what a a few other experts have told me and seem to work for them in auto race applications and at more than 300hp average. BTW, as per cockpit air, I guess that hurricane going past my head when testing the prop and pressurizing the turtle deck was just my imagination. Dumb idea no doubt.
 

Attachments

  • bf109bulge_title.jpg
    bf109bulge_title.jpg
    56.1 KB · Views: 0

rv7charlie

Well-Known Member
Supporting Member
Joined
Nov 17, 2014
Messages
3,284
Location
Pocahontas MS
And, using cockpit air for cooling will cause the same issues that you run into with air cooled engines: virtually impossible to efficiently recover the dynamic pressure of the air, so cooling drag goes way up. Remember, the air has to get into the cockpit, and out to the fan/radiator. (Google 'cooling duct pressure recovery'. That should keep you busy for a year or two (or a decade or two).
 
Last edited:

Riggerrob

Well-Known Member
Joined
Sep 9, 2014
Messages
2,833
Location
Canada
Looking at your Me109 for reprise,

It looks like the lower surface closes out the top one with a 45ish diagonal. The radiator is back 15-16? inches, it's not on the lip. My thought is you will get fair torsion out of the original structure. Solving a problem that doesn't exist?
Yes, ME-109 radiator occupies the full depth of the wing.
Is there any way to transmit torsional loads through the external radiator fairings?
 

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,293
Location
Saline Michigan
4000 cfm is just what a a few other experts have told me and seem to work for them in auto race applications and at more than 300hp average.
I actually got into the issue of auto racing, when I was talking about duty cycle. A lot depends upon the type of racing. Circuit racers tend to appear undercooled for their hp until you look at the dutycycle they run. Power and thus waste engine heat while braking is close to zero, cornering is low power in most classes, and even straightaway power is usually traction limited in lower gears. Circuit racing rarely runs more than a few seconds at a time at 100% power. Yeah, Le Mans, and some of the super speedways can get higher, but most of us will never run there. Designing for 40-50% of max power is commonly all that is needed in most circuit racing. But, as I did mention earlier, airplane dutycycles run 100% for five minutes and then 75% for hours. We can easily overheat the engine and destroy it in those five minutes getting off the ground and out of initial climb if you are only capable of rejecting half the waste heat before boiling away your coolant.

BTW, as per cockpit air, I guess that hurricane going past my head when testing the prop and pressurizing the turtle deck was just my imagination. Dumb idea no doubt.

Most of that hurricane has to go into the free stream or it is not propelling the airplane. The portion going into a poorly ducted radiator system and needing a fan assist to go through the HX's is deducted from the propulsion. That is why the folks at Messerschmitt and Supermarine and DeHaviland and Hawker and North American worked so hard at making pressure recovery inlets for their radiator systems and velocity recovery outlets. Depending upon who you talk too, cooling in some of these systems either had very low net drag or actually had net thrust. All the other schemes make significant net drag, including fan boosted systems.

Now, this is not to say that you could not come up with a fan boosted system that is a net winner on energy, but the odds are against you.

Really, consider replicating the factory scheme for cooling your replica. We know it will work, and with some gentle adjustments (smoother expanding inlets, longer contracting outlets) using what we know now, you might even get closer to ideal than what was run in wartime.

Billski
 

Monty

Well-Known Member
Joined
Jul 15, 2010
Messages
1,365
Location
Fayetteville, AR / USA
I actually got into the issue of auto racing, when I was talking about duty cycle. A lot depends upon the type of racing. Circuit racers tend to appear undercooled for their hp until you look at the dutycycle they run. Power and thus waste engine heat while braking is close to zero, cornering is low power in most classes, and even straightaway power is usually traction limited in lower gears. Circuit racing rarely runs more than a few seconds at a time at 100% power. Yeah, Le Mans, and some of the super speedways can get higher, but most of us will never run there. Designing for 40-50% of max power is commonly all that is needed in most circuit racing. But, as I did mention earlier, airplane dutycycles run 100% for five minutes and then 75% for hours. We can easily overheat the engine and destroy it in those five minutes getting off the ground and out of initial climb if you are only capable of rejecting half the waste heat before boiling away your coolant.



Most of that hurricane has to go into the free stream or it is not propelling the airplane. The portion going into a poorly ducted radiator system and needing a fan assist to go through the HX's is deducted from the propulsion. That is why the folks at Messerschmitt and Supermarine and DeHaviland and Hawker and North American worked so hard at making pressure recovery inlets for their radiator systems and velocity recovery outlets. Depending upon who you talk too, cooling in some of these systems either had very low net drag or actually had net thrust. All the other schemes make significant net drag, including fan boosted systems.

Now, this is not to say that you could not come up with a fan boosted system that is a net winner on energy, but the odds are against you.

Really, consider replicating the factory scheme for cooling your replica. We know it will work, and with some gentle adjustments (smoother expanding inlets, longer contracting outlets) using what we know now, you might even get closer to ideal than what was run in wartime.

Billski
Billski,

I normally wouldn't argue with you, but on the airplane engines doing 75-100% duty cycle, I'm afraid I'm going to. Are we talking about 100% SL power? Is the engine boosted? Does it have a CS prop? Most aren't and don't. 75% rpm at 10Kft is nowhere near 75% SL power. This gets stated (incorrectly) over, and over, and over. It's simply false. Fixed pitch prop, normally aspirated NEVER sees 100% DC, except climb SL standard day at max power rpm. How often does that happen? It may see 75-100% of available power, (rpm) but not rated (manifold pressure). A boat on the ocean with a normally aspirated engine... might see 75-100%, all day, every day. The rest, we agree on, mostly...The OP should listen to you, especially that last paragraph!
 
Last edited:

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,293
Location
Saline Michigan
Billski,

I normally wouldn't argue with you, but on the airplane engines doing 75-100% duty cycle, I'm afraid I'm going to. Are we talking about 100% SL power? Is the engine boosted? Does it have a CS prop? Most aren't and don't. 75% rpm at 10Kft is nowhere near 75% SL power. This gets stated (incorrectly) over, and over, and over. It's simply false. Fixed pitch prop, normally aspirated NEVER sees 100% DC, except climb SL standard day at max power rpm. How often does that happen? It may see 75-100% of available power, (rpm) but not rated (manifold pressure). A boat on the ocean with a normally aspirated engine... might see 75-100%, all day, every day. The rest, we agree on, mostly...The OP should listen to you, especially that last paragraph!

The OP is talking about running a cooling system that appears to have cooling capable of somewhere around 33% to 50% of what his proposed engine is capable of. The limiting case is the one we like. The S-51's and T-51's and other WWII fighter look-a-likes running LS and bigger V-8's ARE running constant speed props and can make 100% power at sea level. Take-off and climb goal is to get safe altitude quickly, so we are all taught to fly max power for a couple minutes. After taxi and runup, even a fan cooled engine will be on the thermostats for coolant and oil before take-off. Sea level density on the ground will make 100% power when they shove the knobs forward for two minutes and the 33-50% cooling solution will result in boiling coolant. 75% power with the knobs forward won't happen until they reach 8000 feet, and a lot of coolant will be gone by then. The forward parts of the heads become uncovered first in climb, and engine failure is imminent when that happens, even at 33% power. This is an awful scenario and completely available to these guys.

Alternate scenario. Still on the thermostats for oil and coolant before takeoff. Climb prop allows 95%torque/90% rpm/86%power on the runway and 90-95% power in climb. Boiling is still happening until you get to 15000 feet - if the engine will run that long.

Another scenario. Cruise prop allows 75% power until 8000 feet and you still boiled off a bunch of coolant before power can be brought back to 40% in cruise. If the engine will run that long.

The only way to operate it and not boil coolant, uncover heads, and fail the engine is to hold it to 40% power all the time. Set power based on coolant temperature for all of the flight. A 300 hp engine at 40% is 120 hp. Do these warbird replica folk really want a 120 hp engine? Might as well run an EJ22 conversion instead of a Corvette conversion. Save a bunch of money and effort, lighter too.

I feel a need to remind us of the Raptor disaster. Cooling was unknown because the pilot would pull back fuel whenever coolant temps started up. Very restricted power output, leisurely climb, and low cruise airspeeds was all he could allow. And he still blew one engine and we suspect a second in what, 50 flight hours? That is terrible. Lucky the program did not KILL anyone. I expect similar issues with a cooling system capable of anything less than 100% duty cycle.

Even if a builder does a downtown job of assessing their duty cycle, sizing the cooling system, accounting for the heat sinking of the engine and coolant, and getting within a gnats eyebrow of coolant boiling, a small underperformance or a little excess power will still get a failed engine quickly. My guess is a 90% duty cycle is still flirting with disaster and is way more than what is being proposed. Once we have sized 90%, a 110% system is easy and only a little bigger.

Now for the rest of the picture. The OP is designing a new airframe with this novel cooling system. New airframes have issues to be worked out. Every automotive conversion is a new engine with issues to be worked out too. The more issues we have in a bird, the more likely we are to find they interact, making solutions difficult to identify and implement. The test pilot is attempting to learn a new airplane while dealing with this complex problem. And the OP is proposing to make this potentially deadly set of issues worse by putting in a known problem on purpose. Recipe for a terrible result...

Please oversize the HX system initially, and then tune for less drag once other problems have been reduced.

I propose that the known and period correct type radiator system be installed, using modern heat exchangers, well done diffusers and nozzles, and sized for excess cooling of the intended engine. Pick a version that put the air inlets in the prop blast (Ross' approach) and it will even cool during extended taxi. No big fans nor huge alternators needed to power them. High speed taxi and see that it behaves. Work out the bugs in the powerplant, then fly it, and work out the rest of the system issues. Be prepared to stop flying and fix cooling, running, and airframe issues. Accept that you may do a lot more rebuilding than flying for a long time. This simplifies your development and greatly increases your chances of finding robust solutions so you can fly the bird, take it places, and play with it in flight.

Billski
 
Last edited:

dog

Well-Known Member
Joined
Dec 29, 2019
Messages
740
the way that legacy aircraft power plants are installed and operated today is a result of countless investigations into "what went wrong"
(hereafter refered to as WWW)
and NO modern,electronic installation addresses
all of those things.
Its a brutal logical process,which takes no prisoners,and is enjoyable only if you can keep
a focus on your set goal
which in this case is lookin cool,goin fast,havin a heck of a good time
which luckily for the 9 year old I remember bieng
(ok still am),only involved swoopy hand movements and self generated aircraft sound track
the quality of the specific advice bieng provided
about how to cool an aviation plant here,is very
very good,and has little to do with opinion,see
WWW
 

addaon

Well-Known Member
Supporting Member
Joined
Feb 24, 2008
Messages
1,983
Location
Kanab, UT
A lot of us with CS props cruise (by the POH) at 75% engine rated power up to 8k or so, as confirmed by fuel flows, although I admit that I tend to cruise at 65% (“efficiency cruise” by my POH) more often. With a turbo, I’ll cruise at 65% rated power up to 14k. Yes, fixed pitch props mean that cruising at 75% at 8k requires advancing the throttle relative to sea level cruise (or, more often, just leaving the throttle full forward as you climb out) — but that’s pretty normal.
 

Riggerrob

Well-Known Member
Joined
Sep 9, 2014
Messages
2,833
Location
Canada
I have often wondered what would happen if you installed fans down-wind (er ... behind cylinders) in an air-cooled radial engine. I started thinking aobut this concept after seeing the inlet cooling fan on a Focke-Wulf 190, then comparing it to a Hawker Sea Fury.
What happens if you bury cooling fans int eh cowling behind the engine and use those fans to suck cooling air through the cowling during ground running and low-speed climb???????
 

Toobuilder

Well-Known Member
Supporting Member
Joined
Jan 19, 2010
Messages
5,763
Location
Mojave, Ca
I have often wondered what would happen if you installed fans down-wind (er ... behind cylinders) in an air-cooled radial engine. I started thinking aobut this concept after seeing the inlet cooling fan on a Focke-Wulf 190, then comparing it to a Hawker Sea Fury.
What happens if you bury cooling fans int eh cowling behind the engine and use those fans to suck cooling air through the cowling during ground running and low-speed climb???????

And that concept is what drove me to try the T-34 (and many other) augmentor tubes for cooling. It works - but you dont need a fan when you have a strong exhaust flow.
 

AdrianS

Well-Known Member
Joined
Jul 5, 2014
Messages
852
Location
Australia
I have often wondered what would happen if you installed fans down-wind (er ... behind cylinders) in an air-cooled radial engine. I started thinking aobut this concept after seeing the inlet cooling fan on a Focke-Wulf 190, then comparing it to a Hawker Sea Fury.
What happens if you bury cooling fans int eh cowling behind the engine and use those fans to suck cooling air through the cowling during ground running and low-speed climb???????
According to something I read, the FW190 used over 120 HP to drive the fan at zero airspeed, but power required dropped as airspeed increased.
 

AdrianS

Well-Known Member
Joined
Jul 5, 2014
Messages
852
Location
Australia
100% duty cycle is hard to cool.

I was involved with OEM engine testing, and part of that was 120 hours at WOT.

The engines were perfectly fine producing ~100 kW forever, but we had a big cooling tower to dump heat into.
We ended up having to eg water cool the sprag clutch so it didn't expand and lock up, fit radiant heat guards so the (orange-hot) exhaust didn't fry things, and watch for heat soak into parts that barely get warm under normal use.

The devil is, as usual, in the details.
 
Top