Allison V-3420- Anyone have information on it? It seems like it would have been perfect for the XB-42.

[wuzak]

The V-1710-127 (E27) was based on the V-1710-109 (E22).

A mock-up was built in September 1944.
A contract for development was awarded in January 1945.
The test mock-up was completed in June 1945.
The complete engine was finished in September 1945. Which means the first run was in September 1945 or later.

Dates from Vees for Victory.

 

[Geoffrey Sinclair]

The allies in Europe standardised on MT80 for ground vehicles and 100 octane for combat aircraft. A common specification was prepared and used. The tanker shortage meant refineries tried to supply their closest theatre, so most ETO fuel came from the Americas, mainly the US while the Middle East refineries handled the Middle East and India and after the end of 1942 fuel from the Americas was shipped to the Mediterranean theatre.

In Britain all fuel arrivals went into a common pool, which caused a small paper war, Lend-Lease, therefore free fuel, was supplied to the pool and the users, like the RAF and USAAF were charged for their consumption. The US had the system changed. Is the idea the fuel supplied from the US to the ETO had a different formulation to that supplied elsewhere?

https://apps.dtic.mil/sti/pdfs/AD1030388.pdf Pages 103 to 106, talks about "special fuel" for the P-38.

The rest of the message is a quote from a few years ago now about the P-38J

"The -25 mainly fixed Lockheed's flight control issues. Allegedly the intake problem overstressed Allison's small engineering team. The behavior of liquid / gas mixtures under extreme conditions is the core of "rocket science" and it was very rare knowledge back then, before computers enabled Computational Fluid Dynamics.

The P-38 Lightning

"To be more specific, the foremost problem was the temperature of the compressed air from the turbosupercharger entering the carburetor. High carburetor air temperature (CAT for short) can cause all kinds of engine problems including detonation, which can lead to catastrophic engine failure. Allison recommended a CAT of no more than 45 degrees C."

"As it turned out high CAT was one of the major problems limiting P-38 performance through the P-38H. The root cause was, of course, the limited cooling ability of the wing leading edge intercoolers found in all early P-38s. They were a very clever design, inducing almost no aerodynamic drag, but they were designed for the 1000 hp Allisons of the late 1930s. By 1943 Military power was up to 1425 bhp and War Emergency Power was 1600 bhp. The increased power required higher induction pressure, which through compression by the superchargers heated the air by several hundred degrees. There is no way that the simple intercoolers could keep CAT below 45 degrees C. when operating at high power at altitude."

CAT wasn't the only problem though. The centrifuge effect of the compressor and the temperature variations in the plumbing could separate out the heavier octane enhancing additives and worsen the destructive detonation. The scientists who really understood this stuff were elsewhere separating Uranium.

The planes performed better in the warmer Med and Pacific so that's where they went. Doolittle and Spaatz made the decision to use the available and very suitable Mustang instead of waiting past D-Day for large numbers of a better P-38 in a paper that I haven't been able to find online. Dialup seriously limits my searching unless I already know the key words and names. This gives an excerpt:

http://www.amazon.com/forum/military%20history?cdForum=Fx5TS2W4P5EMS6&cdPage=2&cdSort=oldest&cdThread=TxZW9V919NXLHM

" A lot of this had to do with its poor intake manifold design, something that Gen. Jimmy Doolittle, who had a PhD in aviation science from MIT, pointed out in a report on the P-38's problems, written to Carl Spaatz in the spring of 1944."

"Doolittle specifically singled out the intake system of the V-1710 as a problem - the tetraethyl lead was somehow separating out with the uneven distribution which caused repeated problems with detonation in certain cylinders in the V-1710s (i.e., they would blow up and catch fire in flight - kinda nasty for the pilots, eh?)."

***"Doolittle's letter to Spaatz is about as clear cut of an explanation of the problems of the V-1710 - turbosupercharger combo at high altitude in Europe as any I've ever found. It's the smoking gun that answers the question of why the P-38 was pulled from almost all combat duty in Europe, remaining only as a photo-reconnaissance plane."***

My further efforts to find that letter suggest it's buried in a Washington archive that requires permission to search and has never been published on the Net.

The dive brake modification shows how long it could take to field a fix on that plane.

Fred Colvin's writing on war production illuminates the difficulty of finding the proper curvature for efficient turbocharger and jet engine compressor blades. Unable to calculate it, the scientists were demanding that the shop produce a variety of exotic and difficult-to-machine mathematical spirals to determine by experiment what was best.

http://people.math.umass.edu/~tevelev/475_2014/beittel.pdf

"However the Archimedean spiral is used in a variety of applications. One such interesting application is in scroll compressors."

End quote.

 

[Shortround6]

Unfortunately we are confusing several different time periods here.
The initial American 100 octane vs British 100 octane was taking place in 1939 or 1940. That was the US insisting that 100 octane fuel have less than 2% aromatics while the British wanting around 20% aromatics. There were several different standardized fuel specification between 1940 and 1943 when the P-38s ran into trouble. I believe there were 3 different specifications for 100/130 fuel for instance (in addition to a 100/125 specification) and it was the last that seemed to cause the most trouble. Allowable lead had gone from 3ccs a gallon (US) to 4ccs to 4.6ccs for instance, not all batches of fuel used the maximum allowable. There were a number of different aromatics allowed, the last specification allowed higher percentages of the heavier compounds. These changes allowed for greatly increase production of 100/130 fuel from the same amount (tonnage) of crude oil or from the initial distillation.
The P-38s troubles had nothing to do with British fuel available in 1940-41.
ALL US engine makers were aware of the differences. British fuel tended to dissolve gaskets and rubber parts in the fuel systems and the liners first used for self sealing.
Kind of hard to ignore.
Allison had a problem simply because they were the only US maker of high powered V-12 engines. The radial engines used a much shorter path from the supercharger to the cylinder intakes. Photo of model but it shows things well.

Each cylinder has a pipe directly from the supercharger diffuser to the cylinder.
Allison had to do this.

Supercharger is to the left. Mixture goes through the pipe just visible in the bottom of the V to the middle of the engine to where the pipe splits to left and right, comes up and slits again to forward and aft 3 cylinder units and then gets split into the 3 individual intakes. Problems were not as random as some accounts seem. the middle cylinders (3 & 4 on the on sides tended to run leaner than than the other cylinders and had a disproportional failure rate. I can't recall now but the left and right banks did not have the same failure rate.
It was this long, convoluted path that caused much of the problem but that was hard to avoid with a V-12 engine unless you used either fuel injection in the cylinders or 3 carbuertors per cylinder bank like the Hispano-Suiza

Note the supercharger blows through the carbs.

The potential problems with the change in fuel specification were anticipated (at least to some extent) and Allison was working on a new manifold design in 1943 but it wasn't finished until near the end of 1943? The new manifold was installed on all new Allison's regardless of having a turbo or not and hundreds were sent out to retrofit engines in the field.

 

[GrauGeist]

Not an aircraft engine, but comparable issues:
Chevrolet's L-6 engines (194, 230, 250, etc.) had the same intake issues with the carb sitting center on an intake that fed the cylinders in groups (1&2, 3&4, 5&6) with an open plenum, which resulted in 3&4 cylinders being too rich or 1, 2, 5 and 6 being too lean, depending on how the carb was adjusted.

 

[Shortround6]

Quite a few inline six and V-12s had similar problems.
An experimental Cadillac V-12 was found to have very different octane ratings in the mixtures in the different cylinders.

 

[wuzak]

It was this long, convoluted path that caused much of the problem but that was hard to avoid with a V-12 engine unless you used either fuel injection in the cylinders or 3 carbuertors per cylinder bank like the Hispano-Suiza

The Merlin's intake was somewhat less convoluted.

https://www.enginehistory.org/members/images/PerfAna/Fig-05-03.jpg

https://www.enginehistory.org/members/images/PerfAna/Fig-05-02.jpg

https://www.enginehistory.org/members/images/PerfAna/Fig-05-04.jpg

 

[Shortround6]

The Merlin's intake was somewhat less convoluted.

https://www.enginehistory.org/members/images/PerfAna/Fig-05-03.jpg

https://www.enginehistory.org/members/images/PerfAna/Fig-05-02.jpg

https://www.enginehistory.org/members/images/PerfAna/Fig-05-04.jpg

The taper in the main pipe is to try to adjust pressure in the pipe as it goes forward. You want equal pressure in each branch and a straight pipe won't give you that.

Unfortunately some of these manifolds work best at one airflow and not quite as well at others.

 

[GregP]

The Allison had 4 sets of intakes that fed three cylinders each. The shape of the intake manifolds caused the outer two cylinders to run a bit rich and the middle one to run a bit lean. Adding a turbulator in the center manifold caused the fuel not only to stay atomized (as was the intent), but also resulted in more even cylinder mixtures. I have disassembled both early and late Allisons, and I can tell you the late intake manifolds shows much more even mixture in the cylinders when you look at them after removing the cylinder bank.

Not exactly sure of the date when the intake turbulator was incorporated, but they came in on the V-1710-61 engine.

The V-1710 started with a 6-counterweight crankshaft. The 12-counterweight crankshaft came in with the V-1710-97 (G-series).

A 2-speed, single-stage supercharger was tried on the V-1710-57 (F-5A, XF-5D, F-38F) and the V-1710-131 that flew on the XC-114 (a DC-4) and YC-116 (never heard of it before finding the reference).

 

[GrauGeist]

The cylinders closest to the carbs should be rich while the cylinders furthest will be leaner. It's a difficult balancing act to tune the carbs to the engine, especially when anticipating the engine's demand under specific loads.

The only way to balance the air-fuel needs for each cylinder, would be direct fuel injection - otherwise there will always be a disparity in fuel delivery.

 

[Shortround6]

The cylinders closest to the carbs should be rich while the cylinders furthest will be leaner. It's a difficult balancing act to tune the carbs to the engine, especially when anticipating the engine's demand under specific loads.

The only way to balance the air-fuel needs for each cylinder, would be direct fuel injection - otherwise there will always be a disparity in fuel delivery.

from what I have read, repeat read, what you say is true most of the time. It might very well be true for the Merlin.
In a non-supercharged the cylinders closest to the carb/s have the shortest path and should get a bit better flow. But the flow is being directed by the suction from the cylinders (unless you have gotten tricky with a tuned intake) and the pressure differences in the air/gas stream are not going to be large.
In a supercharged engine, and that covers a huge area, you have a pressurized manifold system and air/gases are looing to escape wherever and whenever it it can and with more force than the piston sucking down can provide.

Manifold design and port design were not understood as well as they are today.
Going back to the Hispano (because it is a good example of bad practice)

the right bank is being fed by a manifold that is a different length than the left bank and the left bank has got an extra 90 degree (or 75?) bend in it.

The "log" manifold may work but again, having you airflow do 90 degree turns and abrupt changes in passage cross section does not make for good airflow at the entry to the carburetors. It is hard to tell from the photos of the "log" is made up from slightly different diameters of pipe or not. A decrease in pipe cross section would help keep pressure up in the forward end of the manifold. However using that almost cut off/capped end can't be good for the last carb. You are going to have a pocket of air churning in the bottom corner of the pipe as the bulk of the air turns and moves upward. You can probably (I would hope) tune the carbs for each pair of cylinders to cut down on the rich/lean problem even if the cylinders don't all get the same amount of air.

The Allison manifold was designed to be as equal as they could get it without too much spaghetti and at least they didn't end the manifolds in dead ends with holes in the sides.
Hey, for all I know Hispano welded plates or ramps inside the tubes to help straighten things out and you just can't see them.
But with a supercharger when the air is being forced/blown into and engine the air/gas is going to be moving faster than than a unblown engine (you are moving more air per minute at the same rpm) and it is going to have a tendency to go past the early openings.

Going back to the Allison

Air does it's turn up from the main pipe, does it's 90 degree turn to the upper forward manifold. Now it has 3 places to go. Shortest distance is to the No 2 cylinder, No 1 cylinder offers an easy path with a gentile bend before making the turn to the head. No 3 cylinder needs to do a 180 degree turn to get to the passage to the head. Reverse for the rear cylinders. As I look at this it is even a bit worse. The No 4 cylinder, from main pipe in the bottom of the V, needs to do a 360 turn to end up just above the fork/spit in the main pipe, followed by a 90 degree turn into the port in the side of the head. No 1 cylinder has about 135 degrees (?) less turning to do?

 

[GrauGeist]

I will say that Allison did a great job of trying to equalize the lengths in an effort to balance the Air Fuel mixture with their intakes with the day's technology

If I might stray back to the automotive realm for a moment:
The "big three" all had their intake designs for the L-6 engine (I touch on this because of visuals).
Ford's intake was much like Hispano's - a straight open plenum with right angles.
Chrysler's slant-6 had a wild looking intake that resembled a menorah, which provided individual cylinder feed, but still a longer run to cyls 1 & 6.
And GM's six intake that tried to take advantage of the firing order with their combo-port intake (mentioned in my post above).
Of the three, GM did try to address the issue with a dual carb intake, situation a carb between the 1&2 and 3&4 port and another carb between the 3&4 port and 5&6 port. It helped, but was a bit much for the smaller cid engines.

The only factory engine intake I've seen, that tried to provide an equidistant solution, is Chrysler's V-8 "cross ram" intake, which was two seperate manifolds that sat over the opposite head with it's own carb and equal length intake ports to each cylinder.
While the contraption looked ungainly, it was delivering a near perfect AFM to each cylinder.

In the aircraft world, however, there are weight and space limitations that dictate just how far you can go with perfection.

 

[SaparotRob]

As a former member of the Slant Six Club of America, I have to say the slant six was the finest engine ever made. Now that you mention it, the intake manifold does kind of look like a menorah.

 

[Shortround6]

I had Pontiac OHC six for short period of time with a 4-barrel quadrajet.

The Chrysler V-8s were attempting to do something else.

As the engine runs and valves open and close you get pressure pulses traveling up and down the intake tract. On a 2 barrel carb on a V-8 engine they tend to blend together.
One the Chryslers (and there were sometimes two different manifolds) distance from the carb openings to the valve meant that the pulses were coming down the runner hitting the open valve to shove a bit more air and fuel into the engine.
I could be wrong (memory not working) but it has to do with the distance and the speed of sound in feet per minute and the rpm of the engine.
The long RAMS like the picture were set up for torque lower in the rpm band, I believe there was a short RAM set up that maximized high rpm but lost some of the torque?
As can be seen you can need around 30 inches of intake to get the effect down into the torque range of the engine. Shorter distances at going to be used up around 6,000rpm.

Works in motorcycles at high rpm and is the principle behind velocity stacks.

 

[GregP]

The cylinders closest to the carbs should be rich while the cylinders furthest will be leaner. It's a difficult balancing act to tune the carbs to the engine, especially when anticipating the engine's demand under specific loads.

The only way to balance the air-fuel needs for each cylinder, would be direct fuel injection - otherwise there will always be a disparity in fuel delivery.

Makes sense to me, GrauGeist, but it's not the way it looks when you examine the cylinders upon removal. Rich and lean cylinders are fairly easy to pick out. The rich exhausts are dark black and wet and the lean ones are light gray and bone dry. The tops of the pistons also tell a story, as you well know.

 

[muskeg13]

I had Pontiac OHC six for short period of time with a 4-barrel quadrajet.
View attachment 694058

The Chrysler V-8s were attempting to do something else.
View attachment 694059
As the engine runs and valves open and close you get pressure pulses traveling up and down the intake tract. On a 2 barrel carb on a V-8 engine they tend to blend together.
One the Chryslers (and there were sometimes two different manifolds) distance from the carb openings to the valve meant that the pulses were coming down the runner hitting the open valve to shove a bit more air and fuel into the engine.
I could be wrong (memory not working) but it has to do with the distance and the speed of sound in feet per minute and the rpm of the engine.
The long RAMS like the picture were set up for torque lower in the rpm band, I believe there was a short RAM set up that maximized high rpm but lost some of the torque?
As can be seen you can need around 30 inches of intake to get the effect down into the torque range of the engine. Shorter distances at going to be used up around 6,000rpm.

Works in motorcycles at high rpm and is the principle behind velocity stacks.

Oh my! That upper photo brings back a blast from the past. My first car was a '67 Tempest with that engine. Did you experience any lubrication problems with the OHC?

 

[don4331]

Which one to start with...

Hispano - Is actually more scientific than you are giving credit. If you think of plumbing in your house - plumbers add a "tee" and short extension past the last tap to avoid "water hammer" when war is suddenly turned off. The short extension past the last carb is preventing "air hammer" at the last carb. p.s. Remember Hispano started as unsupercharged engine and carb being easily accessed was considered benefit on those engine.

Fuel injection does a much better job of accurately controlling liquid fuel to each cylinder, but if the intake manifold isn't providing equal amounts of air to each cylinder, you will still have. I would be willing to bet more Cross Rams were sold (Chevrolet's equal length, high rpm intake) than Dodge Hemi ones

The Ford Australia straight 6s have some interesting intakes in attempt to get equal fuel delivery.

Merlin won't work with turbine - for either turbo supercharger or power recovery.

If we remember back to our high school chemistry, heating water from room temperature (20 C) to 100C happens very linearly, but then temperature rise stops at 100C and no temperature increase occurs to the water until it is all vaporized (at which point which point, we can begin heating the steam). It actually takes more than 6X energy to completely boil 1 litre of water than it did to heat from 20 to 100C. Hence, why engineers were playing with evaporative cooling systems. On other hand, if we have heated the water at 100C for couple minutes, then turn off the heat, it will remain at 100C for couple moments before temperature begins to fall. Scientifically, its the enthalpy of vaporization.

RR engineers took advantage of this phenomenon to limit the amount water in the engine. It also works to allow reduction in size of radiator - as long as temperature of water is 100C, you have maximum temperature differential to cooling air. (I am aware RR pressurize cooling system which allowed higher temperatures, but concept is the same). Unfortunately, because of this, there is no extra cooling to accommodate the additional heat load caused by the back pressure of a turbine.

Allison did a wonderful job of developing a manifold with equal length runners, but they downplayed the tendency of spark ignition engines to occasionally back fire, especially if a cylinder goes lean. And it gets worse in a highly tuned engine, especially if throttle is rapidly changed - i.e. fighter in a dogfight. In a naturally aspirated engine, it just sounds bad. In a heavily supercharged engine, a back fire rips the intake off the engine e.g Top Fuel dragster blower explosions. The Fiat AS.6 is also very susceptible with a tube the length of 2 V-12s full of air/fuel mixture. The RR solution was to put "flame traps" on the intake so backfire was contained within just the port of the cylinder back firing. But the number of thin plates that the Merlin flame trap is made up of is significant restriction to airflow, so Allison engineers were loathe to implement it.

The speed of airflow in the intake doesn't significantly change as boost increases. 2 atm boost * 1 velocity get twice as much fuel in the cylinder as 1 atm boost * 1 velocity. So, supercharged engine really doesn't need bigger ports as boost increases - for both Allison and RR, the port size/valves of the initial 1,000hp engine is very similar in size to the final 2,000 engines.

Back to V-3420 - there were a couple used by tractor pullers in '80s; they weren't short on power...

 

[Shortround6]

plumbers add a "tee" and short extension past the last tap to avoid "water hammer" when war is suddenly turned off. The short extension past the last carb is preventing "air hammer" at the last carb

Water is incompressible (in most real world situations) and air is compressible.
If the French were worried about "air hammer" in a system that used about 5lbs of boost they had other issues.
When I was a plumbers assistant for a few years (around 30 years back) my boss never did that. Of course we strapped in the pipes ever so often so they would not rattle around and we took care to the bleed air out of the system. We were also using copper pipe, not plastic.
You can buy a an assortment of water hammer arrestors that use a small pre-charged air chamber that will absorb shocks in the system. A few extra inched off incompressible water is not going to help.

 

[OldGeezer]

They did a nice advertisement for the 3420...

 

[wuzak]

Merlin won't work with turbine - for either turbo supercharger or power recovery.

Why not?

 

[GregP]

Why not?

My question, too. All it takes is a new exhaust pipe routed to a turbocharger. Who knows, maybe they could turbocharge a 2-stage Merlin and make it up into the 50,000 feet area.