Air-cooled inline engines with forced cooling

[gjs238]

Could the use of forced cooling, ala BMW 801, have permitted air-cooled inline engine types such as those manufactured by Ranger and Argus to potentially achieve power levels and reliability of their water cooled inline cousins, or air cooled radial cousins?

Just to be clear, I'm not suggesting sticking a fan onto a Ranger V-770.

 

[gumbyk]

I guess it could have been done, but how much power are you going to lose to the cooling fan? Cylinder spacing would have to increase to get enough airflow to the cylinders, so the engine would end up larger too. The radial layout lends itself to a larger cylinder spacing.

 

[Shortround6]

In line air cooled V-12s had several major problems compared to liquid cooled V-12s.

1. is that bore spacing had to greater on the air-cooled engines to leave room for the fins. This means either longer engine for the same cylinder size or smaller diameter cylinders and less displacement.
2. The air cooled engines are separate cylinder engines instead enbloc engines. The Liquid cooled engines are stronger (less liable to bend) than the air-cooled engines. Trying to cast a block of 6 air-cooled cylinders with close spaced fins is going to drive the foundry people nuts and give you a truly astronomical scrap rate.
3. You need the most cooling around the cylinder head and using 3-4 valves per cylinder and overhead cam/s can severely impact the area/volume available for cooling fins, even with forced airflow. A cam box can help strengthen the engine but it really isn't a substitute for 6 cylinder long cylinder block and head.

 

[Balljoint]

The Liberty engine was successfully converted to air-cooling – by Alison if I recall correctly. But, like multirow air-cooled radial engines, higher outputs tend to run into heat rejection problems

 

[Koopernic]

Could the use of forced cooling, ala BMW 801, have permitted air-cooled inline engine types such as those manufactured by Ranger and Argus to potentially achieve power levels and reliability of their water cooled inline cousins, or air cooled radial cousins?

Just to be clear, I'm not suggesting sticking a fan onto a Ranger V-770.

This s certainly possible. The Argus AS411 was an inverted V12. (Ie A12) produced 600hp with a weight of 385kg. This makes it have about 10% lower power to weight ratio than a DB605A at 1.3ata on the same fuel. The DB605A weighed 720kg. However if given fuel injection and an emergency rating it could've done much better. Furthermore it didn't require 100kg-200kg of radiator, plumbing and fluids.

Putting a geared up fan on the front of the AS411 or the slightly smaller Ranger V-770 would certainly work.

One would have to scale up these engines volumetrically about 2.5:1.

The resulting fighter engine would look a little like the Jumo 213 on a Fw 190D9 or the annular Tempest. The forced fan on the BMW801 also forced air through a double reverse flow annular oil cooler.

There is no such thing as an air cooled engine. They are oil and air cooled with the oil cooler handling up to 50% of the cooling.

I think they would've worked just fine. The tendency to perceive radials as good for slow transports and liquid cooled in lines for fast low drag installations in fighters meant this promising technology might have just fallen between the cracks.

I don't accept that these or any engines run into thermal or heat rejection problems if scaled. These air cooled V12s possibly had less problems with cooling air distribution than the PW R-4360 corn cob. Larger cylinders have lower surface area to volume ratios, that means they need to conduct less heat.

 

[pbehn]

There is no such thing as an air cooled engine. They are oil and air cooled with the oil cooler handling up to 50% of the cooling.


Larger cylinders have lower surface area to volume ratios, that means they need to conduct less heat.

I thought that meant more heat had less area to disperse it into?


All aircraft engines are air cooled, the cooling is a heat sink ships and submarines may use a different sink but with aircraft it is merely what is the best way to get the heat of the engine out to the air.

 

[kool kitty89]

The Rolls Royce Exe apparently used forced air cooling of some sort, whether a fan or some sort of ducted compressor I'm not sure. (or rather, a low pressure centrifugal fan with air ducted accordingly for optimized cooling -with likely compromises for complexity, weight, and bulk)

The Exe was also used 24 rather small cylinders in banks of 6, so cooling would be less of a problem than on the likes of similar displacement V-12s. (OTOH the Napier Dager of smaller displacement in H-24 arrangement had considerably greater cooling problems and trouble hitting power remotely close to the Exe)

An axial fan arrangement (or prop cuffs) like some radials used really wouldn't suit inline, flat, V, X, or H blocks very well, so a ducted radial/dentrifugal fan/compressor might be the most sensible option as far as forced cooling goes.

In line air cooled V-12s had several major problems compared to liquid cooled V-12s.

1. is that bore spacing had to greater on the air-cooled engines to leave room for the fins. This means either longer engine for the same cylinder size or smaller diameter cylinders and less displacement.
2. The air cooled engines are separate cylinder engines instead enbloc engines. The Liquid cooled engines are stronger (less liable to bend) than the air-cooled engines. Trying to cast a block of 6 air-cooled cylinders with close spaced fins is going to drive the foundry people nuts and give you a truly astronomical scrap rate.
3. You need the most cooling around the cylinder head and using 3-4 valves per cylinder and overhead cam/s can severely impact the area/volume available for cooling fins, even with forced airflow. A cam box can help strengthen the engine but it really isn't a substitute for 6 cylinder long cylinder block and head.

Aside from forced cooling air, and the above problems with conventional air cooled inline cowlings (or lack thereof), there's potential for cowling/ducting designs optimized for long banks of air-cooled cylinders that distributes air far more evenly than the front-to-back flow arrangement.

I believe De Havilland used ram air intakes outboard of the engine to duct cooling air to the Gypsy Twelve/King's cylinders with flow going from the outside (exhaust end) of cylinders inward before exhausting out a ventral semi-retractable nozzle. (the Albatross and Don seem to have used this method) I haven't seen detailed technical drawings of the internal workings, but from the photographs and diagrams of those aircraft I've seen, it's the best conclusion I could reach.

The other (more obvious) method would be cutting cooling duct intakes in the cowling in front of each of the cylinders, forcing air to flow inward before going rearward between the cylinder banks. You'd obviously need some experimentation to maximize cooling and minimize drag, but the same was true for adapting NACA cowlings to radial engines.

Having a cooling scoop at the front of the cylinder banks forcing air to flow in a typical continuous rearward 'straight through' path similar to most radial engines (or radiators) just doesn't make sense for an inline cylinder bank arrangement.

You could, of course, have a ram air intake at the front of the engine that fed to multiple inward-flowing ducts arranged to give relatively even mass flow over all the cylinders, but that's just a variation of the distributed inward airflow arrangement common to the direct external intakes and DH's arrangement. (more like DH but without the remote intakes feeding air from the rear sides rather than the front sides) I think the DH.88 Comet might have used an arrangement close to this, actually.


Any forced air arrangement (again a radial fan seems most likely to me) would logically also have individual ducts feeding to the cylinders and targeting to maximize flow over the exhaust ports.

 

[wuzak]

The Rolls Royce Exe apparently used forced air cooling of some sort, whether a fan or some sort of ducted compressor I'm not sure. (or rather, a low pressure centrifugal fan with air ducted accordingly for optimized cooling -with likely compromises for complexity, weight, and bulk)

I don't think the Exe had a fan

But the later Pennine certainly would have

The Exe was also used 24 rather small cylinders in banks of 6, so cooling would be less of a problem than on the likes of similar displacement V-12s. (OTOH the Napier Dager of smaller displacement in H-24 arrangement had considerably greater cooling problems and trouble hitting power remotely close to the Exe)

The Dagger was 17l compared to the Exe's 22l.

The proximity of the two lines of cylinders must have made getting the airflow where it was needed.

The Dagger was pressure cooled, by the ram air from scoops in the front of the aircraft. No fan. A fan may have been wider than the engine, so may have not been great for aero.

An axial fan arrangement (or prop cuffs) like some radials used really wouldn't suit inline, flat, V, X, or H blocks very well, so a ducted radial/centrifugal fan/compressor might be the most sensible option as far as forced cooling goes.

With the cooling fan you want air flow. An axial fan would be suitable for inlines, particularly the X engines. It should also be OK for V-12s, obviously lined up for the cooling air passage. For the H-block, like the Dagger and Rapier, you may need two fans, one for each pair of cylinder banks.

I believe De Havilland used ram air intakes outboard of the engine to duct cooling air to the Gypsy Twelve/King's cylinders with flow going from the outside (exhaust end) of cylinders inward before exhausting out a ventral semi-retractable nozzle. (the Albatross and Don seem to have used this method) I haven't seen detailed technical drawings of the internal workings, but from the photographs and diagrams of those aircraft I've seen, it's the best conclusion I could reach.

That's exactly how the cooling in the Albatross worked. Air is taken from inlets in the leading edge of the wing, ducted forward to the outside of the cylinder blocks, force to the inside of the vee, using the ram air pressure. It then exhausts under the engine through an adjustable flap.

The other (more obvious) method would be cutting cooling duct intakes in the cowling in front of each of the cylinders, forcing air to flow inward before going rearward between the cylinder banks. You'd obviously need some experimentation to maximize cooling and minimize drag, but the same was true for adapting NACA cowlings to radial engines.

Having a cooling scoop at the front of the cylinder banks forcing air to flow in a typical continuous rearward 'straight through' path similar to most radial engines (or radiators) just doesn't make sense for an inline cylinder bank arrangement.

You could, of course, have a ram air intake at the front of the engine that fed to multiple inward-flowing ducts arranged to give relatively even mass flow over all the cylinders, but that's just a variation of the distributed inward airflow arrangement common to the direct external intakes and DH's arrangement. (more like DH but without the remote intakes feeding air from the rear sides rather than the front sides) I think the DH.88 Comet might have used an arrangement close to this, actually.


Any forced air arrangement (again a radial fan seems most likely to me) would logically also have individual ducts feeding to the cylinders and targeting to maximize flow over the exhaust ports.

The Dagger used such a straight forward system.

You'll probably find that the air is actually forced outwards between the cylinders to exit the sides of the cowling.

The larger, more powerful radials used baffles to guide the air around the cylinders, maximising the cooling of the cylinder.

 

[Shortround6]

Even Ranger 6 cylinder engines didn't really use a front to back airflow.

Air went in the scoop/hole on the right hand side of the engine, cross flowed over to the left side (with the aid of baffles) and then went rearward out out the back of the cowling. dimensions inside the cowling were such that it was easier for teh air to flow sideways through the engine than to pile up at the rear. Much like modern air handling systems in building change the size of the duct right after an outlet/s to help force the air out through the sideways opening.

Most pictures of engines in books (and engines in museums) often have some or all of the baffling removed. Most visitors what to see the engine, not a bunch of sheetmetal pieces.

First picture is of a later model engine and the cylinders are obscured by the baffles. And those are the engine baffles. Fairchild/Ranger did quite a bit of work on studying air cooling. I have book that reports on a study of a Ranger inline six with thermocouples on each cylinder head and at the base of each cylinder showing the temperature variation from cylinder to cylinder and the difference on each cylinder from head to base. Ranger figured out that air that passed more than 3/16 of an inch from the engine did nothing for cooling. The baffles are to direct the air and to force the air through the fins rather than letting air go around the cylinders or just over the tips of the fins.

AS for " I don't accept that these or any engines run into thermal or heat rejection problems if scaled. These air cooled V12s possibly had less problems with cooling air distribution than the PW R-4360 corn cob. Larger cylinders have lower surface area to volume ratios, that means they need to conduct less heat."

I would really like to see the reasoning behind that because it makes no sense from the developing power stand point. It make make some sense in a collage class room while discussing ultimate fuel efficiency but fails in the real world if you are trying to compete with liquid cooled engines.

Every advance in engine power by air cooled aircraft engines was accompanied by (or enabled by) an increase in the cooling ability of the cylinder. The more square inches of fin area you could get on a cylinder of a certain size the more power you could pull out of it before melting things (pistons and valves) , or suffering a breakdown in lubrication on the cylinder walls (piston and/or rings galling on the cylinder walls). More power means more fuel/air burned in the cylinder per minute, higher rpm, higher boost or both. This is not a theoretical exercise in heat distribution (what percentage goes to the crankshaft, what percentage goes out the exhaust and what percentages go to the oil and the cylinder walls/head). It is about burning the most fuel possible in a given amount of time without melting engine parts or cooking the oil. As the oil heats up it looses it's lubricating qualities and in extreme cases it "burns" or cooks to a point where it leaves a solid, mostly carbon residue which can score the parts.

 

[Shortround6]

The larger, more powerful radials used baffles to guide the air around the cylinders, maximising the cooling of the cylinder.

to add to Wuzak's informative post, In some engine text books from the 40s they have temperature diagrams of cylinders. View from the top, temperature every 15 or 20 degrees of arc though all 360 degrees of the cylinder. yes the back of the cylinder was hotter than the front but not by as much as you might think.

 

[kool kitty89]

I don't think the Exe had a fan

But the later Pennine certainly would have

Wiki's mention of pressure cooling must refer to ram intakes used on the Exe then. I also may have jumped the gun on my axial vs radial fan comments. The Pennine's arrangement seems like it would be pretty efficient if they included appropriate ducting/splitters/baffles to channel the air efficiently towards the four banks. (and likely expand the channels outward to provide the most space/pressure at the rear cylinders or otherwise channel air for evenly distributed cooling. (plenty of detail changes possible there, just the general common factor being even distribution of cooling air flow and/or higher mass flow in areas air ends up pre-heated before reaching the cylinder)

Sleeve Valve engines of course also have the advantage of more evenly distributing heat throughout the cylinder rather than concentrating it so much at the heads (or on the exhaust side of the head in particular), so the Exe would have the advantage there too.

With the cooling fan you want air flow. An axial fan would be suitable for inlines, particularly the X engines. It should also be OK for V-12s, obviously lined up for the cooling air passage. For the H-block, like the Dagger and Rapier, you may need two fans, one for each pair of cylinder banks.

Yeah, I failed to consider a lot of arrangements there, including V-12s scavenging air up and over the crankcase and channeling it back towards the cylinders where it's most needed. (the key being in all cases wanting flow to go from the exhaust port side of the cylinders towards the intake and THEN exhaust out in whatever direction presents the least resistance)

The Dagger used such a straight forward system.

You'll probably find that the air is actually forced outwards between the cylinders to exit the sides of the cowling.

That arrangement doesn't make a whole lot of sense, the air is rammed into ducts between the cylinder banks on the intake manifold end and it looks like the cooling duct actually narrows as it reaches the rear cylinders rather than expanding to better distribute air where most needed (DH cowlings seem to have adopted tight front ducts and large rear ducts on both the front and rear intake arrangements). I would think cooling air intakes oriented on the sides of the engine above the upper and below the lower exhaust stacks would be most effective, either four individual intake ports or 2 large central intake ports (like already used) split off to the sides in the same manner as 4 individual ports would and have the air ducted rearward and then inward towards the intake manifold before exhausting out of outlets on the top and bottom of the engine.

You might end up heating the intakes manifold a bit more that way, but you keep the most vulnerable portion of the engine cooled. (plus RPM and boost pressure limits on air cooled engines tend to be less limited by carb/manifold temperatures and much more limited by cylinder head temperatures)

I wonder what sort of progress the Germans made with cooling air for the Hirth and Argus engines.

 

[Koopernic]

I thought that meant more heat had less area to disperse it into?


All aircraft engines are air cooled, the cooling is a heat sink ships and submarines may use a different sink but with aircraft it is merely what is the best way to get the heat of the engine out to the air.

In a piston engine the purpose of cooling is only to remove enough heat to keep the materials from being damaged. In closed cycle engines such as rankine, Stirling, closed brayton it's part of the cycle itself.

Less area to volume on a larger piston means less material to protect.

Either way if you make an V12 out of say R3350 pistons/stroke you get a V2230, about the size of a griffon or Jumo 213 of about 1600hp.

An air cooled in line should have a potential advantage over an air cooled radial and that is the possibility of a 4 or 3 valve head. The disadvantage over a liquid cooled in line might be the greater cylinder spacing required to allow the inter cylinder air gap.

 

[Shortround6]

(the key being in all cases wanting flow to go from the exhaust port side of the cylinders towards the intake and THEN exhaust out in whatever direction presents the least resistance)

Actually you want the reverse. The intake area being much cooler than the exhaust area the air flow, after picking up the heat from the exhaust area may do very little cooling of the intake area. Going from the intake area the cooling air flow has gotten warmer but still has a good temperature difference between it's own temp and the temp of the engine parts it is going over.
Add in the facts that by allowing the intake area to run a bit hotter you wind up with a hotter intake charge which means a less dense one for less total power and a hotter engine in general ( increase the intake temp 50 degrees and the peak temp in the cylinder goes up 50 degrees and the exhaust temp is also 50 degrees hotter) and you may have shot yourself in the foot by trying to cool the exhaust valve first.

That arrangement doesn't make a whole lot of sense, the air is rammed into ducts between the cylinder banks on the intake manifold end and it looks like the cooling duct actually narrows as it reaches the rear cylinders rather than expanding to better distribute air where most needed

It actually makes a lot of sense. The idea is to make the air go sideways though the cylinders to the other side so each cylinder gets the same amount of cool air. The rear cylinder/s are NOT supposed to be getting air that has flowed over the earlier cylinders. You need a duct that has about the same pressure from front to back to do that. Expanding the duct near the back (or even keeping a constant section) will lower the air pressure in the duct in the back and result in a lower mass flow through the cylinder fins of the rear cylinders as air mass and pressure bleeds off past the front cylinders.

Diagram of a modern flat six.

 

[Shortround6]

duplicate

 

[Koopernic]

Notable in Wuzak's diagram of the Penine is the stator behind the fan. A stator increases fan efficiency and pressure ratio. It must increase the ability to engineer a good balanced airflow. The BMW 802 was also to receive a stator though the BMW 801 did not have one.

 

[yulzari]

Quite some years ago I happened to meet an ex De Havilland engine chappie. Amongst other things he told me that, due to their success with the Gipsy Twelve De Havilland were asked to report upon the Napier Dagger's cooling in the Hawker Hector and Handley Page Hereford. Essentially De Havilland reported that the problem was not cold air entry but poor exit and suggested some simple ways of improving the exits that would solve the problem but all of this was late 1941 so too late as Napiers had moved on. Principally they suggested ducting the exits to existing low pressure areas of the airframe.

At the time I was asking about advice on low drag cooling of VW Supervee racing engines so this must have been in the late 1970's. Blackening the fins with selenium dioxide took the temperature down by 3 degrees and running the petrol through dry ice took another 2 degrees off the head temperatures. Looking back, had we ducted a common exit to over the gearbox we could have drawn the intake air from under the car and gained some downforce, especially with some door draught stop brushes sealing the underside from sideways airflow migration. A poor mans Brabham BT46 fan car but I fear I am wandering OT.

 

[Koopernic]

When I came across selenium compounds in the 1980s, usually in the form of a heat paste for electronics or perhaps as part of a rectifier stack, we were not far of from evacuating the room. The women were sent out.

 

[wuzak]

Notable in Wuzak's diagram of the Penine is the stator behind the fan. A stator increases fan efficiency and pressure ratio. It must increase the ability to engineer a good balanced airflow. The BMW 802 was also to receive a stator though the BMW 801 did not have one.

The stator would be there to straighten the flow, which reduces the rotation of the air flow. The air flow rotating would lead to two of the cylinder banks getting more favourable cooling than the other two.

 

[kool kitty89]

Air went in the scoop/hole on the right hand side of the engine, cross flowed over to the left side (with the aid of baffles) and then went rearward out out the back of the cowling. dimensions inside the cowling were such that it was easier for teh air to flow sideways through the engine than to pile up at the rear. Much like modern air handling systems in building change the size of the duct right after an outlet/s to help force the air out through the sideways opening.

I expect it was mostly (some) 2-cylinder inlines and flat fours that used simple front-to-back flow, obviously those without any cowls at all would apply there.

AS for " I don't accept that these or any engines run into thermal or heat rejection problems if scaled. These air cooled V12s possibly had less problems with cooling air distribution than the PW R-4360 corn cob. Larger cylinders have lower surface area to volume ratios, that means they need to conduct less heat."

I would really like to see the reasoning behind that because it makes no sense from the developing power stand point. It make make some sense in a collage class room while discussing ultimate fuel efficiency but fails in the real world if you are trying to compete with liquid cooled engines.

Higher surface area to volume ratio (ie smaller cylinders) would tend to be much easier to cool, the opposite is true for larger volume cylinders attempting similar power/volume. Now, using larger cylinders but the same or only modestly lower power (much lower volumetric efficiency) is another matter. (but keeping frontal area and weight down becomes more difficult)

And that's not even getting into piston speed or flame front propagation.

Less area to volume on a larger piston means less material to protect.

Internal volume or surface area doesn't matter in terms of combat vulnerability, at least not in this context. It's overall external diameter, length, and general engine geometry that limits vulnerable area. (that and oil cooler capacity)

Sure, on a single row radial engine you have SOME chance of bullets or shells passing between cylinders entirely, but on nearly any multi-row radial (or inline) that's not the case.

If you compare different engine arrangements like inline vs V vs X vs H, there's certainly differences in vulnerable area, but also other issues like an X or H engine more likely to take hits to cylinders vs V and inline having more change of taking a direct hit to the crankcase. (all if unprotected) And air cooled engined tend to be able to power through damaged cylinders better (and drain oil more slowly) than holes punched in the crankcase or oil cooler.

Actually you want the reverse. The intake area being much cooler than the exhaust area the air flow, after picking up the heat from the exhaust area may do very little cooling of the intake area. Going from the intake area the cooling air flow has gotten warmer but still has a good temperature difference between it's own temp and the temp of the engine parts it is going over.
Add in the facts that by allowing the intake area to run a bit hotter you wind up with a hotter intake charge which means a less dense one for less total power and a hotter engine in general ( increase the intake temp 50 degrees and the peak temp in the cylinder goes up 50 degrees and the exhaust temp is also 50 degrees hotter) and you may have shot yourself in the foot by trying to cool the exhaust valve first.

If this is the case then why do radial engines position their exhaust valves at the front of the cylinders (at least on engines that don't have both intake and exhaust ports on the rear side of cylinders) and why did De Havilland opt to use an exhaust to intake manifold flow as well? This is particularly dramatic with Bristol's exhaust arrangement and their exhaust collector ring design, sometimes even embedded into the leading edge edge of some of their cowling designs. (which also kills exhaust thrust and adds weight, but that's another story -the ring COULD potentially be used to exploit venturi style augmented thrust if it had multiple annular jet exhaust ports ducted INSIDE the cowling but behind the cylinders, provided the fuselage can withstand the added hot air+exhaust and exhaust doesn't get sucked back into the cockpit)

It actually makes a lot of sense. The idea is to make the air go sideways though the cylinders to the other side so each cylinder gets the same amount of cool air. The rear cylinder/s are NOT supposed to be getting air that has flowed over the earlier cylinders. You need a duct that has about the same pressure from front to back to do that. Expanding the duct near the back (or even keeping a constant section) will lower the air pressure in the duct in the back and result in a lower mass flow through the cylinder fins of the rear cylinders as air mass and pressure bleeds off past the front cylinders.

Using larger ducting at the back makes sense, yes, I must not have been clear about that previously (I suggested that widening the duct around the rear cylinders would be preferable)

Quite some years ago I happened to meet an ex De Havilland engine chappie. Amongst other things he told me that, due to their success with the Gipsy Twelve De Havilland were asked to report upon the Napier Dagger's cooling in the Hawker Hector and Handley Page Hereford. Essentially De Havilland reported that the problem was not cold air entry but poor exit and suggested some simple ways of improving the exits that would solve the problem but all of this was late 1941 so too late as Napiers had moved on. Principally they suggested ducting the exits to existing low pressure areas of the airframe.

Wouldn't individual exit ports cut around the exhaust stacks be a simpler solution? Sure it'd add suction drag, but that same drag/low pressure area would be pulling air out just where it's needed (especially with decent aerodynamic shaping of the ducts). You'd still either need cooling flaps to control airflow either in the intake or through larger cooling exhaust flaps, but it seems like the more foolproof option if less than ideal drag-wise.

Using flaps along the exhaust stacks seems like trouble, so those cooling ports would likely remain static exits, unless you used a single long shroud around all the exhaust ports ( a shroud cowl with individual cooling air + exhaust manifold exiting and expanding into ) with the outlets to the rear of the engine and regulated by flaps as well. This seems less efficient given individual small cooling ports could use stub exhaust where that might be problematic (and less effective) in a single long shroud.

 

[wuzak]

If this is the case then why do radial engines position their exhaust valves at the front of the cylinders (at least on engines that don't have both intake and exhaust ports on the rear side of cylinders) and why did De Havilland opt to use an exhaust to intake manifold flow as well?

If you look at the Albatross drawing and pictures you will see that the exhaust of teh Gipsy Twelve was indeed on the inside of the vee. It would appear that the intake was too.