What was the problem with the allison engine?

[Milosh]

I wouldn't spend too much time arguing with this person, as this exact same argument happened here: Reliability of aircraft engines nearly 10 years ago.

With virtually the same figures pulled out of context to support the same circular argument

I will say that perhaps the single most informative comment in all 9 pages of that particular discussion, was in another member's reply:

That is some thread with Huck and Kurfurst being major participants.

 

[LisaM]

Not sure what the argument is over, everyone knows that the Allison was a very fine engine, but RR were the masters of supercharging with long accumulated experience and a depth of technical and scientific skills that Allison could not match (nor anyone else for that matter).

But on a different topic I think NA made a mistake with the P-51B, a better approach would have been to develop an interim type (like the Spit V or IX) based on one of the better Mustang X prototypes. It would have got it into service far faster (albeit with lower. but still very good performance), the engineering risks would have been reduced (much based on the tried and tested P-51A). It would have been a useful 'interim type', more than competitive with the 109s or 190s of the era, later superseded with a more developed version (and almost certainly due to the lessons learned in operations, more reliable than the first P-51Bs).

The 'undergunning' alone was astonishing given that the Mustang I had 8 machines guns and the II, 4 x 20mm cannon. There was a real weird element in the USAAF about guns which continued right up to the F-86.

 

[tomo pauk]

The experience of the RR was indeed great, they started designing and producing aircraft engines from the days of ww1, and between the wars they were the prime source of V-12 engines anywhere in the world. Both for military needs and racing. They reckoned early enough that low compression ratio plus a sizable/capable supercharger are the keys for the success of the supercharged aircraft engine.
The USAF was wrong to turn down the, in 1938 suggested, 2-stage supercharged V-1710, thinking that turbed one will do. Compare that with USN and indeed the RAF tht seized the opportunity with 2-stage engines as early as possible.

We might call the P-51B-K an interim type, the P-51F/G/H/J being wholesale redesigns. Until/unless there is no 2-stage Merlins available, there is no point in an even 'more interim' type. Eg, in July 1943, there were only 173 2-stage V-1650s delivered to the NAA, while there are 534 P-51B airframes completed - lack of 360 engines.
Experiences with Mustang X pointed out that cooling capacity need to be increased, in order to cater with increased cooling loads vs. Allison Mustang. The whole project of Mustang with 2-stage Merlin was a very swift thing IMO. What Allies missed was the Mustang with 1-stage V-1650 that ended in the indifferent P-40F/L, the 1st engines were evailable by the end of 1941. Or even with the P-51A/Mustang II instead of A-36.

The Mustang Ia was with cannons, the II was with 4 HMGs. The 8-gun Mustang sounds nice, but 4 of the guns were .30s, and 2 of the HMGs were slower firing due to being synchronised. USAF/USN probably settled with HMGs beacuse their 20mm cannon was not that a reliable thing, especially in wing installations. Despite trying hard with the 20mm.

 

[pbehn]

Not sure what the argument is over, everyone knows that the Allison was a very fine engine, but RR were the masters of supercharging with long accumulated experience and a depth of technical and scientific skills that Allison could not match (nor anyone else for that matter).

But on a different topic I think NA made a mistake with the P-51B, a better approach would have been to develop an interim type (like the Spit V or IX) based on one of the better Mustang X prototypes. It would have got it into service far faster (albeit with lower. but still very good performance), the engineering risks would have been reduced (much based on the tried and tested P-51A). It would have been a useful 'interim type', more than competitive with the 109s or 190s of the era, later superseded with a more developed version (and almost certainly due to the lessons learned in operations, more reliable than the first P-51Bs).

The 'undergunning' alone was astonishing given that the Mustang I had 8 machines guns and the II, 4 x 20mm cannon. There was a real weird element in the USAAF about guns which continued right up to the F-86.

The development of he P 51/Mustang was a fortunate comedy of errors, the A36 Apache version was ordered as a dive bomber purely because there was no money for fighters but there was some for bombers. The British were moving from 303MGs to canon so the Mustang/P51 had the pleasure of trying out many combinations.

 

[FLYBOYJ]

But on a different topic I think NA made a mistake with the P-51B, a better approach would have been to develop an interim type (like the Spit V or IX) based on one of the better Mustang X prototypes. It would have got it into service far faster (albeit with lower. but still very good performance), the engineering risks would have been reduced (much based on the tried and tested P-51A). It would have been a useful 'interim type', more than competitive with the 109s or 190s of the era, later superseded with a more developed version (and almost certainly due to the lessons learned in operations, more reliable than the first P-51Bs).

The 'undergunning' alone was astonishing given that the Mustang I had 8 machines guns and the II, 4 x 20mm cannon. There was a real weird element in the USAAF about guns which continued right up to the F-86.

Understand that many times these decisions are not made by the manufacturer but by the customer, including what type of armament will be utilized.

 

[XBe02Drvr]

...or sound as sweet!

 

[XBe02Drvr]

...or sound as sweet!

Ooops that was in reference to a hypothetical R2800 powered Lancaster/Shackleton several pages back. ("It might've worked, but it wouldn't have been as pretty!")

 

[kool kitty89]

I was considering adding to this topic: Single stage V-1710: options for improvements?
but this thread's been active a lot more recently and covers most of the same thing (and a lot more).


Addressing that specific topic of single-stage first:

The 7 notable mechanical limitations/problems the early (C series -mostly commonly the V-1710-33 of the P-40B/C/Tomohawk -Curtiss Model 81) were:

1. relatively weak reduction gearing making any stress of overboosting beyond the relatively modest performance limits very destructive to this engine (I believe this is what gained reputations for stripping gears with the AVG's rather abusive use of the engine, possibly some cases in the MTO as well) This was addressed with the E/F and later series engines to the point of often tolerating very heavy out-of-spec abuse. (this might have still been a bottleneck for overrev, but the prop/hub/blade-mounts would likely be a bigger bottleneck)

2. crankshaft strength and to some extent overall finish (I've been digging around for the disucssion that had more specifics on this -particularly the actual changes between the C and E/F series and later models of the E and F prior to the added changes with the G series, but I haven't found it yet). From what I recall, the F series received a smoother finnish than the C series and had strengthening allowing higher power levels. (the well-balanced design of the C series already made for smooth running and ability to overrev -unlike the Merlin and some others- and made such more tempting for pilots to exploit, or attempt to)

3. inability to use the smaller 14-tooth supercharger gear needed for 9.6:1 supercharger operation (8.8 uses a 15-tooth gear) due to structural limits of the existing gear pitch/width requiring a redesign of the mount to allow space for a thicker gear. The required changes would have disrupted production and added costs the Army was unwilling to pay. (this was the main reason for delay in adopting this iirc, there's been more accurate and detailed discussion on it before, and I might have the exact terminology wrong, but I think this is about right)

4. Lack of 2-speed drive. Similar to the gear-tooth issue but requiring even more dramatic redesign of the compact accessory section of the engine, presumably with greater overhead in production changes than most other design problems/fixes. (and wouldn't benefit turbocharged engines) The Army would likely sooner fund an auxiliary supercharger than 2-speed drive given the concerns over efficiency of production.

6. Backfire screens adding drag in the engine manifold and reducing manifold pressure and/or mass flow for any given altitude and supercharger configuration. These were later removed as unnecessary some time after the 9.6:1 supercharger had been introduced. (I think the P-39M is the only fighter to receive a 9.6:1 supercharged engine retaining those screens)

7. Kink/bottleneck in the supercharger intake manifold around the carburetor/throttle body. This was one issue that I don't believe was ever solved and was also mainly limited by the Army's unwillingness to invest in the necessary changes to production. (this problem was noted pre-war, perhaps even prior to any large scale serial production had started but possibly after initial production tooling had been laid out -I'm not sure, but that seems likely given the cost/delay for modifying a non-production design would be so much less as to make that ridiculous, even with a turbocharger-oriented altitude performance program)


The latter issue is rather similar to what the pre-XX series Merlin suffered (in terms of intake chocking the supercharger -noted by Hooker when he began work on improving it) with the exception that Allison's engineers and the USAAF were aware of the problem when RR seemed less so until Hooker inspected the design. So it seems they had the engineering ability to solve (or at least start work on improving) the issue but the Army declined to fund such changes. Solving that issue (like with the Merlin) should have reduced charge heating and increased critical altitude for any given manifold pressure, resulting in higher power levels at any given pressure and higher maximum pressures without suffering detonation. (the latter more important with 9.6 superchargers and possibly Pre-J P-38 given the poor intercooling and any little bit of reduced charge heating in the supercharger still being relevant ... though clearing that kink should make ALL turbo installations less bottlenecked and more efficient)

Lack of aggressive/exhaustive testing combined with relatively conservative ratings in general seemed to plague the V-1710 as far as WEP and take-off ratings went, possibly even maximum continuous power. The E and F series should have been capable of pushing well beyond 3000 (perhaps even 3300-3400) RPM and unauthorized operation at such speeds were noted (let alone post-war tractor pulls and such pushing them far far higher -even without the G-series crankshaft) but no official testing leading to up-rating seems to have been done on this. (reduction gear and prop would be the breaking points for overrev, but without proper testing of those limits, there'd be no reference for even designing around WEP limits and no context for aircraft designers to appropriately test their aircraft and propeller combinations -or note whether different reduction ratios would be needed) The smooth running nature of the engine might have made a 3000 RPM max continuous rating also possible, but specific limits on boost pressure at that rating would also have to be noted separately from military/WEP (or 30-minute limits).

Likewise, tests specific to 100/130 and 100/150 octane fuel should have been completed sooner (or used more incomplete results for operational clearance/expanded operational testing).

It seems the British Air Ministry and Rolls Royce were both more aggressive with testing WEP and general engine limits and more willing to push less than exhaustively tested ratings before putting them into practice. (given the more conservative and stringent nature of Aircraft design specifications the USAAC/AAF and USN had already established pre-war, this seems less surprising, and the V-1650-1 somewhat reflects this as well given the apparent lack of WEP clearance compared to the contemporary Merlin XX and 45 series)

Overrev is particularly significant as it would have boosted critical altitude as well, and partially side-stepped the 9.6:1 gear delay. (exceeding 3200 RPM with 9.6:1 gearing would also seem unwise as the impeller tip speed would exceed the speed of sound by a significant margin even taking generous charge-heating into account, and would also exceed any real-world tip-speed examples during war-time use, while 3200~3250 RPM itself would rather closely match the tip speeds employed with the V-1650-1's supercharger in high gear, or the experimental 10.25" diameter 2-speed single stage Allison supercharger at 9.6:1) On that note, an 8.8 supercharger on an engine running at 3300 RPM will result in a speed equivalent to 9.68:1 at 3000 RPM. (so very similar tip-speeds to 9.6:1 engines at normal military power operation)

Overly conservative ratings can be more dangerous in practice due to leaving the real, hard, extreme operating limits of the machines totally vague and questionable, resulting in a rather high incentive for individual experimentation and unauthorized operation in the field with no reference to how dangerous such operation was. (compared to pushing the Merlin beyond WEP specifications, which was much more reliably going to end in failure) The rather dramatic warning of 57" boost limit on the 9.6:1 engines seems to be among the exceptions of extreme limits being posted, but even that is exaggerated in its imperative as experienced showed pushing closer to 60" was still possible without detonation on those engines, and their notice doesn't note this behavior at all. (P-39N flight testing even lists use of 59.8" manifold pressure being used)

OTOH, there may have been cases of Allison recommending/clearing use of certain engine settings, but the USAAF never implementing them. (2000 HP with 75" MAP on 100/150 fuel was cleared for the P-38L's engines, but may not have been officially cleared for P-38 operations ... though obviously would at least have given a nicer guide for unauthorized use operationally)
150 Grade Fuel

Making improvements on the base, single-stage engine is pretty universally important as it extends usefulness to all existing designs using the engine (no serious modifications to address changes in weight or changes in overall dimensions) while equally extending those improvements to engines using turbochargers or external auxiliary superchargers.




Now, to the points beyond the basic single-stage engine:

The big overarching issue here is the Army simply should have supported Allison's R&D far, far better, and put those hyper-engine projects at much lower priority. This includes putting turbocharger+intercooler development on higher priority too, specifically for technology making it more suitable in small, sleek single-engine aircraft.

Turbos have 2 big issues for small fighters (P-39 is a great example in the extreme here), their own size/bulk, and the size/bulk/drag caused by the intercooler. The larger the turbo, the larger the intercooler and resulting ducting, and restricting turbocharger design to relatively large units precludes the possibility of medium-altitude optimized variants with reduced size/weight/intercooling capacity requirements. (say targeting a critical altitude in the 20,000-25,000 ft range) Or even having turbos small enough to allow reasonable use with no intercooling at all, more akin to Allison's non-intercooled auxiliary superchargers.

This issue can be drastically reduced by adopting a liquid intercooler, and Allison could (with enough funding) have tackled this engineering project themselves, optimizing around GE's turbochargers. (this would have been a critical development all around, and really the biggest single achievement for extending the V-1710 beyond single-stage operation) Given the coolant loop would be shared with the engine's own cooling system, this seems a reasonable thing for Allison to invest in developing rather than GE. (and turbocharger relevant, thus more likely to gain Army support/acceptance) If this work had been done early enough, the P-38 could have abandoned its wing intercoolers sooner and gained both power and (potentially) leading-edge fuel tanks in the wings for longer range operations. (granted, still needed cockpit heating, dive flaps and hydraulic -or tab- boosted ailerons to make a really potent escort fighter ... just the boosted ailerons to make a good warm-climate/low alt fighter though)



One possible workaround for the funding difficulties would have been licensing/adapting existing auxiliary supercharger technology from P&W, or at least using it to accelerate Allison's own development (and avoid delays with half-hearted 'cheap' attempts at repurposing the same integral supercharger housing+impeller into an auxiliary stage, among other things Allison tried initially). Particularly the pre-war developments for the R-1830 (which should have fit the mass flow and pressure requirements of the early V-1710 rather well, likely matched with the 7.48:1 integeral stage speed intended for low-alt and turbocharged installations). Work to allow either rear or side-mounted auxiliary stage orientation (or perhaps even under or top-mounted) would improve flexibility of installing such engines on existing designs using single-stage engines. The V-1710 was specifically designed for modular accessories, so having that sort of flexibility really should have come with the territory.


ADI/water injection is useful, but given the research and testing necessary for addressing concerns over corrosion (especially when used in combat -ie relatively recently before landing, not just on take-off) and required bulk/weight of an ADI fluid tank (again, reduced for use on-take-off only) makes this less attractive than intercooler development. I'm not sure if the R-2800 had operating limits for water-injection use to this effect or not, but the relatively early-ear introduction of that mechanism seems like it might have been initially limited for take-off purposes (akin to that on the Bramo 323R-2). For carrier or short airfield use, that added boost on take-off could be critical, especially for long-range or fighter-bomber missions with heavy external loads.


Edit:
Found the discussion on crankshaft variations
Curtiss-Wright: Loss of Don Berlin and downfall

According to "Vee's for Victory" there were 4 'different' crankshafts used in the Allisons. The first 3 look identical, at least from a distance. I don't work on them so there may be minor visual clues. The first were 'plain' crankshafts which I believe the C-33s got. The next version was shot peened, different surface texture? much improved fatigue life. I don't know when it was introduced. Then they nitrided the crankshafts in addition to the shot peening. This allowed for another major increase in fatigue life.

Each step allowed for roughly an unlimited life at a stress level that the preceeding step would only tolerate for a very short period of time. Nitriding was introduced in early 1942 and allowed about a 70% increase over the old plain steel (not shot peened) Crankshaft in stress levels for the crankshaft with both cranks operating at a level that they could sustain for an unlimited duration. Also in 1942 the casting method for the engine blocks changed. The new method required about 10% fewer operations to manufacture ( casting was closer to finished dimensions), weighed a bit less and was stronger. There may have been changes in the vibration damping system between the "C"s and the later engines. Or between certain models of the later engines.

The "C" series engines, according to the book, were rated at an overspeed of 3600rpm. The "E" and "F" engines were rated at 4100rpm for overspeed when introduced and the "G" series with the 12 counter weight crank was rated was rated in excess of 4400rpm.
This was not theory. As part of the engine type test the engine had to survive running at that speed for 30 seconds and do it a number of times during the duration of the test, usually a minimum of 10 times, depending on contract.

Now what happened in the field could be way different and what an individual pilot did either in pursuit of an enemy nearly in his sights or when trying to save his own life could be different also.
However, trying to operate "C" series engines at power levels used by "E" "F" engines, while possible short term, was at a lot higher risk and definitely shorter engine life, let alone the reduction gear problem.

The US Air Corp had the problem of rating engines for combat use with the factory 3000-8,000 miles away from the front lines. Spare engines and spare parts for even an in theater overhaul shop had to be transported those distances. They had to trade off short term performance gains of the aircraft vs blown engines, making men fly planes with engines in questionable condition, not having enough planes in service to fly the desired number of missions in a day and so on. Which more hazardous to a pilots life, not being allowed to use WEP settings and flying in a 12 plane formation to meet the enemy or being allowed to use WEP settings and having an 8-9 plane formation to meet the same number of enemy aircraft?
Maybe they did get it little wrong, maybe they got it a lot wrong.

 

[Shortround6]

Part of the problem with pre-war development was that Allison was basically an engine experimental shop attached to a bearing factory.
At some point in 1938 if I remember right, Allison had 25 people in the "engine" engineering section including the guy/s who ran the blueprint machine. The Army was late in making payments and at one point had requested Allison investigate fuel injection, This was while Allison was working on the basic engine, the remote engine for the P-39, the V-3420 double engine, the pusher engine for the Airacuda (which wound up with a turbo). Allison told them they didn't have the engineering staff and and if they wanted fuel injection pick one of the other projects to drop.
Allison was making a couple of engines a month in 1938. Much of the engineering staff was diverted to designing production tooling when the big contracts came until more staff could be hired (and a new, much larger factory built, which had to be designed ).
At one point in 1940 or early 1941 ( I forget which) Allison was short over 800 machine tools despite having one of the highest priority ratings that existed. Allison wound up with over 800 sub contractors supplying parts. By this time supply was certainly not the problem of the design staff but in the early years? Who was deciding what was built in house and what was contracted out?

Another big problem is that engines were often not delivered until months or even a year after they had been contracted for. Granted contracts could be amended so that new improvements might be implemented during a production run but sometimes "new" models of engine were being worked on and promised so as to dove tail with with aircraft production a year in the future. The P-40D was ordered with the -39 Allison almost a year before the prototype flew. The Air Corp had ordered 5 "F" series engines in 1939 using FY (fiscal year) 1940 funds for a model test and one each for the XP-46/46A and XP-47/47A. The XP-47s were canceled or morphed into the Thunderbolt (kept a piece of canopy framing?) and the XP-46 didn't fly until Feb 1941. P-40D flying with the extra engines from the XP--47S???

This confusion as to what the army wanted was one reason the Allison was designed the way it was, a central power section with the reduction gears as a separate section and the accessories as a separate section so the maximum use could be made of the production tooling for the different versions the Army was dreaming up. They would sometimes make p-40 engines for 1-2 weeks and then make p-39 engines for a few weeks and then switch back to P-40 engines and so on.

Sometimes money helps and sometimes it is luck. The improved casting technique for the engine blocks had actually been developed by a husband and wife artist team for casting aluminum sculptures. Totally outside the normal foundry industry. How much extra money had to do with finding them I have no idea

Part of the problem with the turbos was that the basic design may not scale well. And what you are trying to accomplish.

An Allison needs 10,000lbs of air per hour to make 1672 hp in the cylinders which translates to 1285hp to the prop for an engine with 9.60 gears after you take out the friction and power to drive the supercharger. You could get 1415hp from an engine with 6.44 gears but that maxed out the supercharger, it can't supply more than 10,000lb of air with a crank speed of 3000rpm.

You need both mass flow and pressure. At 25,000ft the turbo only has to compress the air 2.6 times to get sea level air pressure. At 20,000ft the turbo still has to compress it about 2.2 times (or just under) while flowing the same mass of air. You aren't going to save much of anything trying to design and build a slightly smaller turbo-supercharger.

The need for inter-coolers is can be shown by charts in standard text books of the day. Temperature rise though an auxiliary supercharger (regardless of how driven) to deliver sea level pressure with a compressor of 65% efficiency (and granted some superchargers could get a bit over 70%) was just under 50 degrees at 5000ft, around 90 degrees at 10,000ft, about 125 degrees at 15,000ft, 175 degrees at 20,000ft and 210/215 degrees at 25,000ft according to one chart. Granted the air does get colder as you climb so that at 20,000ft the air going to the cab may be only 100 degrees or so compared to the 59 degrees at sea level for standard conditions.

At under 20,000ft you may not need an inter-cooler but at under 20,000ft you don't need the turbo either.
At 20,000ft the exhaust thrust from a Melrin XX on a Hurricane was worth 126.8hp at 340mph. Mass flow was 151lbs per minute and exhaust gas velocity was 1788fpm from the exhaust nozzles. Running the exhaust through 5-7 feet of pipe and then through a turbine is going to kill a fair amount of the velocity even if you build a trick variable nozzle after the turbo.

 

[tomo pauk]

The V-1710 does not need the extra supercharger gear, for low level (thus becoming a 2-speed supercharged) until/unless USAF wants a bomber powered by such an engine. And then, if USAF wants a bomber powered by V-1710, using the turbo might make a lot of sense, the turbo engines were with low-geared integral supercharger already available.

 

[wuzak]

The V-1710 does not need the extra supercharger gear, for low level (thus becoming a 2-speed supercharged) until/unless USAF wants a bomber powered by such an engine. And then, if USAF wants a bomber powered by V-1710, using the turbo might make a lot of sense, the turbo engines were with low-geared integral supercharger already available.

A two speed drive for the V-1710 fitted to the P-40 and P-39 may have helped their performance.

Improved altitude performance could be chased, while maintaining low altitude performance at a similar level for not much extra weight (<100lb).

Single speed superchargers are compromises. Gear for high altitude performance, lose out at low altitudes.

 

[Shortround6]

Two speed superchargers and better fuel over lapped. The first few engines to use 2 speed superchargers were running on 87 or 91 octane fuel depending on country. The fuel limited the amount of boost and therefore the power an engine could develop. If you wanted power at 12-15,000ft you sacrificed power at low altitudes, sometimes a lot of power. The two speed supercharger was a way to get it back given the fuel limits. A Merlin X picked up almost 200hp for take-off over the single speed Merlin III with both running 87 octane fuel. About a 22% increase.
With 100 octane fuel more boost could be used at low altitude and more power developed which lessened the need the for 2 speed superchargers (although they were still nice to have) Fighters, before they started hanging 1000-2000lbs worth of bombs/drop tanks on them, could usually get by with single speed superchargers for take-off/low altitude. The Allison and Merlin benefiting from the move/s from 87 octane to 100 (or 100/115?) to 100/125 to 100/130 and finally 100/150.

Had fuel development stagnated or stalled for a long period of time perhaps more effort would have been put into 2 speed superchargers.
Please note that air cooled engines could not be pushed using higher boost to the extent that the liquid cooled engines could until the war was more than 1/2 over.

And please note that the long nose Allison running on US 100 octane fuel (100/100) offered just 35 less hp for take-off than a Merlin X running on 87 octane while still offering 1090hp at 13,200ft. Merlin III being good for 60hp less 3,000ft higher.

The ideal was better fuel, 2 speed supercharger drives and better compressor designs. But often one or two could fill in in for the other/s to some extent.

 

[tomo pauk]

A two speed drive for the V-1710 fitted to the P-40 and P-39 may have helped their performance.

Improved altitude performance could be chased, while maintaining low altitude performance at a similar level for not much extra weight (<100lb).

Single speed superchargers are compromises. Gear for high altitude performance, lose out at low altitudes.

The P-39 and P-40 have had problems with hi-alt performance. Installing the 2-speed drive with an extra low gear does not help their performance above 12-15 kft. What helped, albeit belatedly, was installation of higher gearing for the, still, 1-speed S/C. We can recall that P-39N was not lacking performance at low level.
Granted, with 2-stage supercharger, more than one S/C speed was necessarry.

 

[kool kitty89]

Aftercooling or water injection would have helped single-stage performance at all altitudes, especially for the 9.6:1 supercharger (or overreved 8.8) due to the charge cooling and improved density.

Additionally, I'd failed to consider the possibility of water injection (or for that matter, fuel vapor) increasing the speed of sound further, thus allowing impeller tip-speeds to be well in excess of the speed of sound at ~100*C (or there about) as humidity greatly increases this value. So even at the 9.6:1 ratio, overreving quite a bit might not hit mach 1 tip speeds.


Given the small size of Allison's staff, it's pretty impressive they managed to engineer what they did ... the Army probably should've taken more notice of that talent though. Aside from just diverting more funding to Allison directly, there's possibilities like second sourcing engine production to Continental and/or possibly outsourcing some of the supplemental engineering to Continental (or perhaps Chrysler, but it really seems like Continental got the brunt of the Hyper-Engine funding). Outsourcing to expand testing (especially for maximum WER limits) would have been a big thing to offload from Allison's limited internal staff/engineering resources too.


They did make quite a lot out of the engine in the end, though, even if it never got used as such operationally:
http://www.enginehistory.org/Convention/2009/Presentations/SuperchargingAllison.pdf

That final turbo-compound variant featured multi-port fuel injection and aftercooling (though the 3200 RPM WER limit was still rather conservative, particularly above its critical altitude).

Interesting that they went port-injected rather than direct injected. That's the only WWII era example of multi-port fuel injection I'm aware of. Interesting that both single-port (pressure injection carb) and multi-port fuel injection both made a reappearance with GM in the 1980s when engine computers were coming on the scene. (single-port fuel injection was called TBI -throttle body injection- in GM literature) Strange that port injection wasn't more common for WWII engines though given it offers many of the advantages of direct injection without the high pressure fuel pump requirements. (most of the timing and precise per-cylinder mixture control, though not as precise timing as direct injection it's still close and offers some degree of charge cooling where high pressure direct injection does not)



I suppose a non-structural reason for avoiding super-high RPM ranges was the ideal cruising performance at 1600 RPM making propeller and reduction gear design rather difficult. (to perform efficiently from 1600 to 3400 or perhaps close to 4000 RPM) And it's not like multiple reduction gears would have been a very practical option ... or at least, clutched prop reduction gearing doesn't seem to have been appealing enough to implement during the war. OTOH, Jumo pushed the 213 to 3250 RPM and its optimal cruise speed was proportionally lower to the smaller capacity V-1710, so perhaps 3400 RPM wouldn't have been unreasonable.




On the note of reduction gearing, though, I got to thinking about the transition from the C to F model (long nose to short simplified/stronger spur-gear arrangement) changes and whether it might have been easier or just all around more useful to avoid the integral gearing of the F series and standardize with the E series alone. If a very short extension shaft+remote gear box could be implemented, it would mean using the exact same engines in the P-40/P-38/P-39/P-51, simplifying production while making propeller and reduction gear changes much simpler and even possible as retrofits. Plus it might retain some of the (at least theoretical) cleaner long/tapered nose that the Tomahawk and XP-38 sported (I believe the YP-38 as well). The thrust-line would still be raised compared to the C series though, so the stub-shaft + separate gearbox wouldn't be able to perfectly mimic the P-40B/C installation.


On that note, I'm unsure whether the XP-46 really used an F series engine or not. Some articles mention using the -39, but Joe Baugher's page mention the -39 being planned for installation, but the -29 actually installed (possibly a typo). The -29 was one of the low-altitude (or turbocharger-oriented) engines the YP-38 used, which should be a C series engine with low-altitude 7.48:1 supercharger ratio. The XP-46 really looks like it has the P-40B/C/Tomahaek (H81) nose and propeller profile and carb intake placement. If the V-1710-29 really was used, then the performance testing of that aircraft would be nowhere close to what the -39 could have offered. (granted, actual info on the XP-46's performance is somewhat vague and contradictory from what I've seen and might be using a standard C15/V-1710-33 of the contemporary P-40 for all I know, but the cited 'poorer performance' than the P-40D prototype seems unusual in any case given how much more stripped down the thing was and the lower drag area -smaller wing and radiator area) The V-1710-29 WOULD be rated for 1150 hp take-off/military like the -39, just with much lower rated altitude. (similar to the A-36 vs Mustang I/IA ... except limiting boost pressure on the A-36 and 1150 rather than 1325 HP military -albeit at the same altitude the A-36 could do 1150 hp, but several thousand feet lower than the -39's ~12,000 ft)

I believe the XP-38 used early 'sea level' rated 6.44:1 gear ratio C series engines, while the YP-38 adopted 7.48:1 engines. (I don't think the 6.44:1 ratio was used in any operational aircraft) The -27 and -29 engines appear to have been used on the P-38D and E, with the F being the first model to use F-series engines, ironically enough. (and hence take-off power rising from 1150 to 1325 HP -like the A-36, and parallel to the 8.8:1 counterparts rising from 1040 to 1150 hp take-off/military -prior to later up-rating and WER)



Edit:
Looking at that above-linked article more (and the claims of placing the carb between the aux and integral supercharger stages improving altitude performance) I realized the inlet to the aux stage in the intermediate carburetor configuration has inlet guide vanes that might be acting as swirl throttle if they're not static. This might explain why they delayed implementation of an intermediate carburetor location given the added testing involved and superior performance. Using a conventional throttle plate to restrict air flow between 2 mechanical supercharger stages might also be problematic, so having a wide-open carburetor with the aux supercharger inlet acting as the throttle would make plenty of sense. (the text and drawings don't make it clear whether this is the case, though) A swirl throttle arrangement would also make the hydraulic coupling less necessary given the efficiency and improved low altitude performance of the variable inlet guide vane arrangement. (though it might also help compensate for the hydraulic coupling having only 1 gear, thus having the best of both the features of the Jumo 213 and DB-601/605/603, and also avoiding excessive oil heating at high engine RPM low-boost conditions, though a declutched/neutral position would help with that too, if it has one, thus avoiding the accessory drive end of the fluid coupling's turbine spinning at high speed, doing little work and generating lots of waste heat)

 

[wuzak]

Aftercooling or water injection would have helped single-stage performance at all altitudes, especially for the 9.6:1 supercharger (or overreved 8.8) due to the charge cooling and improved density.

Additionally, I'd failed to consider the possibility of water injection (or for that matter, fuel vapor) increasing the speed of sound further, thus allowing impeller tip-speeds to be well in excess of the speed of sound at ~100*C (or there about) as humidity greatly increases this value. So even at the 9.6:1 ratio, overreving quite a bit might not hit mach 1 tip speeds.

Aftercooling and ADI allow the engine to develop more boost, so without improvements to the supercharger the extra performance of such systems will tend to be concentrated at lower altitudes.

 

[wuzak]

That final turbo-compound variant featured multi-port fuel injection and aftercooling (though the 3200 RPM WER limit was still rather conservative, particularly above its critical altitude).

Not sure that they did.

The turbo-compound used a standard G-series core engine.

The AAC had Allison investigating injection (not sure if it was port or direct - have to check) for the V-1710 and X-3420 in the mid 1930s.

On that note, I'm unsure whether the XP-46 really used an F series engine or not. Some articles mention using the -39, but Joe Baugher's page mention the -39 being planned for installation, but the -29 actually installed (possibly a typo). The -29 was one of the low-altitude (or turbocharger-oriented) engines the YP-38 used, which should be a C series engine with low-altitude 7.48:1 supercharger ratio. The XP-46 really looks like it has the P-40B/C/Tomahaek (H81) nose and propeller profile and carb intake placement. If the V-1710-29 really was used, then the performance testing of that aircraft would be nowhere close to what the -39 could have offered. (granted, actual info on the XP-46's performance is somewhat vague and contradictory from what I've seen and might be using a standard C15/V-1710-33 of the contemporary P-40 for all I know, but the cited 'poorer performance' than the P-40D prototype seems unusual in any case given how much more stripped down the thing was and the lower drag area -smaller wing and radiator area) The V-1710-29 WOULD be rated for 1150 hp take-off/military like the -39, just with much lower rated altitude. (similar to the A-36 vs Mustang I/IA ... except limiting boost pressure on the A-36 and 1150 rather than 1325 HP military -albeit at the same altitude the A-36 could do 1150 hp, but several thousand feet lower than the -39's ~12,000 ft)

The crime of the XP-46 was that the performance improvement, if there was any, was insufficient to warrant disrupting production of the P-40.

Looking at that above-linked article more (and the claims of placing the carb between the aux and integral supercharger stages improving altitude performance) I realized the inlet to the aux stage in the intermediate carburetor configuration has inlet guide vanes that might be acting as swirl throttle if they're not static. This might explain why they delayed implementation of an intermediate carburetor location given the added testing involved and superior performance. Using a conventional throttle plate to restrict air flow between 2 mechanical supercharger stages might also be problematic, so having a wide-open carburetor with the aux supercharger inlet acting as the throttle would make plenty of sense. (the text and drawings don't make it clear whether this is the case, though) A swirl throttle arrangement would also make the hydraulic coupling less necessary given the efficiency and improved low altitude performance of the variable inlet guide vane arrangement. (though it might also help compensate for the hydraulic coupling having only 1 gear, thus having the best of both the features of the Jumo 213 and DB-601/605/603, and also avoiding excessive oil heating at high engine RPM low-boost conditions, though a declutched/neutral position would help with that too, if it has one, thus avoiding the accessory drive end of the fluid coupling's turbine spinning at high speed, doing little work and generating lots of waste heat)

I'm not sure how much an improvement using the carburettor between the supercharger stages gave to altitude performance. I suspect it has a lot to do with the intake shape of the carburettor on the auxiliary supercharger.

Practically speaking, using the carburettor between compressor stages allowed the core engine to be, essentially, the same as used in a turbocharged installation. The downside is that charge cooling could only come using ADI, and not an aftercooler.
I believe it also shortened the length of the whole engin without that massive carburettor hanging off the back.

 

[tomo pauk]

The relocation of carb gained some 2500 ft to the rated altitude. It also allowed for 4000 ft gain in condition of full ram, vs. just 1500 ft in case the carb was at the entrance of the 1st stage. Net result - aircraft's (P-63C) max speed on military power was attained at ~29000 ft, vs. at ~24000 ft for the P-63A. Rate of climb at high altitudes was also improved, along with max speed.

 

[GregP]

The Turbo-Compound was the V1710-127 or was an E27 in Allison parlance and did not use a G-series core. The G-series started with the V-1710-97, but was not used in the -109. The -127 was essentially the same as a -109 engine except equipped with mechanical feedback exhaust turbine and a different propeller shaft gear ratio.

The prop gear ratio was 2.48 : 1.

No agenda here, just saying ...

 

[wuzak]

The Turbo-Compound was the V1710-127 or was an E27 in Allison parlance and did not use a G-series core. The G-series started with the V-1710-97, but was not used in the -109. The -127 was essentially the same as a -109 engine except equipped with mechanical feedback exhaust turbine and a different propeller shaft gear ratio.

The prop gear ratio was 2.48 : 1.

No agenda here, just saying ...

Vees for Victory stated that the -E27/-127 was based on components from the -E30/-133, which was the latest development standard at the time.

 

[wuzak]

The relocation of carb gained some 2500 ft to the rated altitude. It also allowed for 4000 ft gain in condition of full ram, vs. just 1500 ft in case the carb was at the entrance of the 1st stage. Net result - aircraft's (P-63C) max speed on military power was attained at ~29000 ft, vs. at ~24000 ft for the P-63A. Rate of climb at high altitudes was also improved, along with max speed.

The change improved the altitude performance not because of the carburettor position, but because of the reduction in losses in the air flow for 1st stage to 2nd stage.