The Secret Horsepower Race by Calum Douglas (and piston engine discussion)

[Calum Douglas]

And that was done by the Ford plant in the UK, who slapped RR upside the head and said "we cannot make an engine like this".

Do you have a copy of the letter from Ford US or Packard asking what the torque specifications were? That Rolls literally didn't have because they didn't use torque wrenches but a highly trained fitter to hand assemble all this stuff to "tight enough"?!?

You are absolutely right that RR did not use torque wrenches at all.

This is because in Britain at the time RR called them "torque slipping spanners", read through any Merlin manual and see for yourself.

This is from the 1938 Merlin-II handbook.

I dont really know what you mean by "that was done by Ford". Anyway, Ford Plants never had anything to do with the engineering of the Merlin, the "Ford" factory was built, funded and set up by the Air Ministry, it was never made at any existing Ford plant,
Ford were brought in to run and administer the plant which the Ministry set up.



Ford never made a single Merlin in any existing Ford factory.



The absurd nonsense talked about "hand built" all falls apart once you read the files, its nonsense and not a shred of primary source proof of it exists anywhere.

What is true, is that in the early to mid 30`s ALL Rolls-Royce engines were "hand built" because there was no mass production
set up yet at RR (Bristol started early and the shadow factory mass production setup started in 1935/6).

RR / Air Ministry then set up plants at Crewe and Glasgow where totally unskilled workers who had probably never even
repaired a lawn mower were brought in. This was enabled by American automatic machine tools (which is what the ACTUAL
US contribution to Merlin manufacture in Britain was, but nobody wants to talk about it as "who cares about machine tools").

Glasgow was mass producing Merlin`s with unskilled and semi-skilled labour before Packard had made a single one.

So yes, early Merlin`s were all "hand built" because there was NO mass production occurring, later both RR, Ford and Packard
all mass produced Merlin's.

Packard later released a sort of commemorative brochure about their achievements in which they declared that they
had mass produced the "hand built Merlin". This was entirely true, at one stage they WERE all hand made, but
some people decided that this meant that by magic the British had managed to make twice as many Merlins as anyone
else during the war with little men with files and hacksaws in sheds. When the initial discussions with Packard were initiated
was about the time when everyone was gearing up to turn highly skilled craft assembly into mass production.

The notion that this "hand built" situation carried on through the war is absolutely absurd, statistically and there is
no evidence for it whatsoever.

Hookers comment is likely just some half remembered hand-me-down quote from someone, at the time it occured
Hooker was just a mathematician/aero thermal specialist working on supercharger analysis and had only just started work
at RR. Why on earth anyone thinks he was setting up a Ford factory at the time I dont know. By the way, that book was
written in hospital at the end of his life and came out in 1984, the year he died. Its full of charming recollections
of things forty years earlier some of which are just half remembered anecdotes he was told by someone else.

See the real FORD letter at the bottom:



In the actual files, nothing of any supporting nature for it exists.

 

[HoHun]

Hi Calum,

This was enabled by American automatic machine tools (which is what the ACTUAL
US contribution to Merlin manufacture in Britain was, but nobody wants to talk about it as "who cares about machine tools").

I do, and I bet a lot of the other readers of this thread do too

To add some German perspective, my vague impression is that despite the overall reputation of German machine tools, there was a scarcity of these both in pre-war and in wartime Germany, and this scarcity impacted a lot of the aircraft programmes, including the Me 309. (Though "Vorrichtungsbau" is a somewhat broader term in describing all of the production implements, I believe a big part of it was integrating machine tools into a coherent production line.)

With the Germans often being unable to put new designs into production, that of course put an increased emphasis on improvement of their engines to keep the existing airframes competitive ... and your book covers how well (or not!) that went

Regards,

Henning (HoHun)

 

[Calum Douglas]

Hi Calum,



I do, and I bet a lot of the other readers of this thread do too

To add some German perspective, my vague impression is that despite the overall reputation of German machine tools, there was a scarcity of these both in pre-war and in wartime Germany, and this scarcity impacted a lot of the aircraft programmes, including the Me 309. (Though "Vorrichtungsbau" is a somewhat broader term in describing all of the production implements, I believe a big part of it was integrating machine tools into a coherent production line.)

With the Germans often being unable to put new designs into production, that of course put an increased emphasis on improvement of their engines to keep the existing airframes competitive ... and your book covers how well (or not!) that went

Regards,

Henning (HoHun)

Daimler-Benz sent a delegation to the USA in about 1936 to purchase large numbers of US made automatic machine tools,
they were not able to purchase anything like as much as they wanted because of limitations in the amount of US currency reserve
held in Germany at the time. I`m not an expert on international trade finance in the 30`s, so I`m not quite sure why they
couldn't exchange Gold for Dollars, possibly by then there were some restrictions in place on German monetary activities
which made life difficult.

 

[Nicknick]

It's not any different from your other ADI injection.

Also, IIRC it's better than 50% alcohol and a couple % corrosion-preventatives.

Sounds quite like good old GM1.... (Göhring Mixture 1)

 

[pathology_doc]

I do a little calculation

Excellent; thank you.

 

[Scott Kenny]

You are absolutely right that RR did not use torque wrenches at all.

This is because in Britain at the time RR called them "torque slipping spanners", read through any Merlin manual and see for yourself.

Okay, so now I need RR to make me a couple thousand of those for all the different factories.

How do I know that the Torque Slipping Spanner is set correctly after the trip across the ocean?

The specs sent to the US did not give a torque number, your own reference doesn't give it. You look in a modern shop manual, it'll tell you "torque head bolts to X ft-lbs" (or N-m if you're reading a non-US manual)

 

[Calum Douglas]

Okay, so now I need RR to make me a couple thousand of those for all the different factories.

How do I know that the Torque Slipping Spanner is set correctly after the trip across the ocean?

The specs sent to the US did not give a torque number, your own reference doesn't give it. You look in a modern shop manual, it'll tell you "torque head bolts to X ft-lbs" (or N-m if you're reading a non-US manual)

Thousands of gauges and other checking apparatus for Merlin part`s was sent to Packard, along with three of their top
engineers including the chief designer, who then lived in America for several years to be at Packard. The notion that
such information on tightening bolts was not provided before assembly began is patently insane.





One suspects the Americans probably knew how to tighten bolts with apparatus not requiring to be sent half way across the world,
and you would hardly need thousands anyway as the line only has something like 50 engines on it at at one time.

Its a photo from a 200 odd page manual. Amazingly enough
I didnt post you 200 screenshots. Do you think the torque slipping spanner is calibrated to a random number?

Even mechanics were sent over to train in assembly, the fact initial engineering drawings might not have had a torque
spec would be totally irrelevant, why would you need to know an assembly torque to give a quote for machining
a cylinder block ? That stuff wasnt needed until many months later. Do you think they sent over mechanics all the way
over the Atlantic for them to arrive at Packard and say "ah ok chaps well we never bother tightening anything properly
just sort of turn it until you think its ok". and then flew back across the Atlantic ?

 

[newsdeskdan]

Okay, so now I need RR to make me a couple thousand of those for all the different factories.

How do I know that the Torque Slipping Spanner is set correctly after the trip across the ocean?

The specs sent to the US did not give a torque number, your own reference doesn't give it. You look in a modern shop manual, it'll tell you "torque head bolts to X ft-lbs" (or N-m if you're reading a non-US manual)

I'm interested in the American set-up. What brand of torque wrenches were being used by Packard at this time, Scott, and how many did they have? Presumably the Americans needed those torque slipping spanners because they didn't have an equivalent tool - but you're saying that they actually did have those tools, and that sending the torque slipping spanners was unnecessary?

Tracing the history of the torque wrench in the US is difficult because there seem to be a lot of different claims as to who first gained a US patent on it. It appears to have first been made commercially available in... 1938? So did Packard buy theirs from Chrysler at that time or later? Or from Sturtevant? I don't understand why the Americans needed all those tools from RR if they already had everything they needed?

Also, if the Brits didn't provide a torque number, as you're saying, how did Packard get a torque number? Did they just work it out for themselves? Can you show some primary sources on this please, Scott? Ideally, could you show what tool Packard was using, what torque number they used for Merlins and how they arrived at that number? Thanks!

 

[Scott Kenny]

I'm interested in the American set-up. What brand of torque wrenches were being used by Packard at this time, Scott, and how many did they have? Presumably the Americans needed those torque slipping spanners because they didn't have an equivalent tool - but you're saying that they actually did have those tools, and that sending the torque slipping spanners was unnecessary?

They had torque wrenches, I'll need to see who got my Grandfather's. Not that much of my family was mechanics before me...

If the torque specs had been provided, other than "use this slipping torque spanner", you could use a torque wrench which is significantly cheaper because you only need ~3 torque wrenches per mechanic to cover a range of inch-lbs to hundreds of ft-lbs, and usually only one per station on the line versus needing a separate slipping spanner for each specific bolt use. One for the head bolts, another for the same size bolts but much higher torque because it's used on the connecting rods... etc.

Instead what they got told was "use this slipping spanner". Great. How do we know it's set correctly, versus a flex beam or better yet a dial torque wrench (assuming that the ratchet type wasn't available yet, all the ratchet types I've seen have plastic boxes not wood).

Packard: "Give me the name of the man certifying that each "spanner" is set correctly, we will hold him accountable for any failures due to improperly tightened bolts."

Also, if the Brits didn't provide a torque number, as you're saying, how did Packard get a torque number? Did they just work it out for themselves? Can you show some primary sources on this please, Scott? Ideally, could you show what tool Packard was using, what torque number they used for Merlins and how they arrived at that number? Thanks!

As was explained to me, that's exactly what they had to do in order to productionize the Merlin. They went back to the size of bolts and the Society of Automotive Engineer papers and charts for that size bolt and broke a few engines because the specs weren't tight enough.

Merlin parts were not originally 100% interchangeable (in the 1930s), the acceptable as-finished tolerances were wide enough that you could not just grab the next part off the line but had to swap around till you got a good set. Can't build an engine quickly that way, but it will be higher performance if you do. It's what we call blueprinting an engine in hot-rodding...

Once productionized, the parts were more or less interchangeable.

 

[newsdeskdan]

They had torque wrenches, I'll need to see who got my Grandfather's. Not that much of my family was mechanics before me...

If the torque specs had been provided, other than "use this slipping torque spanner", you could use a torque wrench which is significantly cheaper because you only need ~3 torque wrenches per mechanic to cover a range of inch-lbs to hundreds of ft-lbs, and usually only one per station on the line versus needing a separate slipping spanner for each specific bolt use. One for the head bolts, another for the same size bolts but much higher torque because it's used on the connecting rods... etc.

Instead what they got told was "use this slipping spanner". Great. How do we know it's set correctly, versus a flex beam or better yet a dial torque wrench (assuming that the ratchet type wasn't available yet, all the ratchet types I've seen have plastic boxes not wood).

Packard: "Give me the name of the man certifying that each "spanner" is set correctly, we will hold him accountable for any failures due to improperly tightened bolts."



As was explained to me, that's exactly what they had to do in order to productionize the Merlin. They went back to the size of bolts and the Society of Automotive Engineer papers and charts for that size bolt and broke a few engines because the specs weren't tight enough.

Merlin parts were not originally 100% interchangeable (in the 1930s), the acceptable as-finished tolerances were wide enough that you could not just grab the next part off the line but had to swap around till you got a good set. Can't build an engine quickly that way, but it will be higher performance if you do. It's what we call blueprinting an engine in hot-rodding...

Once productionized, the parts were more or less interchangeable.

The American torque wrench thing seems to be a bit of a mystery. Who sold Packard the torque wrenches? Chrysler seems to have been behind the development of the production model but Sturtevant seems to have been given the licence to make it, and they were a small company by all accounts. If the tool was patented in 1938 (patent granted in 1942) and Packard was tooling up for Merlin production in 1941, that gives three years for Sturtevant, if that's who it was, to mass produce enough torque wrenches to meet demand - and demand must have been absolutely monumental.

While, anecdotally, it seems like maybe Packard had some torque wrenches, as you say - it doesn't seem like many (any?) period examples of these have actually survived. You'd think there would be loads of survivors on eBay etc. from old toolkits gathering dust in the corner of garages, storage lockers etc. if they'd been available in their tens of thousands by 1941.

It just doesn't seem like there's a lot of hard evidence to show what Packard was using, what they thought of the RR materials and what they actually did as regards the correct degree of torque being applied to bolts. Do you know if there are any documents on it?

Also, is there any evidence that torque specs definitely were not provided, like a letter from Packard asking for them etc?

 

[Calum Douglas]

They had torque wrenches, I'll need to see who got my Grandfather's. Not that much of my family was mechanics before me...

If the torque specs had been provided, other than "use this slipping torque spanner", you could use a torque wrench which is significantly cheaper because you only need ~3 torque wrenches per mechanic to cover a range of inch-lbs to hundreds of ft-lbs, and usually only one per station on the line versus needing a separate slipping spanner for each specific bolt use. One for the head bolts, another for the same size bolts but much higher torque because it's used on the connecting rods... etc.

Instead what they got told was "use this slipping spanner". Great. How do we know it's set correctly, versus a flex beam or better yet a dial torque wrench (assuming that the ratchet type wasn't available yet, all the ratchet types I've seen have plastic boxes not wood).

Packard: "Give me the name of the man certifying that each "spanner" is set correctly, we will hold him accountable for any failures due to improperly tightened bolts."



As was explained to me, that's exactly what they had to do in order to productionize the Merlin. They went back to the size of bolts and the Society of Automotive Engineer papers and charts for that size bolt and broke a few engines because the specs weren't tight enough.

Merlin parts were not originally 100% interchangeable (in the 1930s), the acceptable as-finished tolerances were wide enough that you could not just grab the next part off the line but had to swap around till you got a good set. Can't build an engine quickly that way, but it will be higher performance if you do. It's what we call blueprinting an engine in hot-rodding...

Once productionized, the parts were more or less interchangeable.

Pg.22 of the 1938, Merlin II manual ^

The calibration / checking procedure is on page 146

 

[Apophenia]

Excuse my mechanical ignorance but am I right is assuming that the differences here are in the way that a fitter knows the torque has been correctly applied?

IIRC, a basic dial-type torque wrench relies upon the mechanic correctly reading said dial to avoid over-torquing. (The click-type being slightly more sophisticated in providing an audible warning.) In contrast, a slipping spanner (aka 'cam-over' style torque wrench) eliminates the possibility of exceeding the targeted torque. If that is correct, the slipper spanner seems the simpler method of staying within guidelines - assuming, as Scott noted, that the spanner is set correctly.

Or am I over-simplifying things?

 

[Calum Douglas]

Excuse my mechanical ignorance but am I right is assuming that the differences here are in the way that a fitter knows the torque has been correctly applied?

IIRC, a basic dial-type torque wrench relies upon the mechanic correctly reading said dial to avoid over-torquing. (The click-type being slightly more sophisticated in providing an audible warning.) In contrast, a slipping spanner (aka 'cam-over' style torque wrench) eliminates the possibility of exceeding the targeted torque. If that is correct, the slipper spanner seems the simpler method of staying within guidelines - assuming, as Scott noted, that the spanner is set correctly.

Or am I over-simplifying things?

More or less correct yes.

However life does get very complicated once you get into the detail, and in fact the very most accurate methods are done by hand without any "direct" torque measurement gauge. Often this was done on con-rod bolts where the preload accuracy is extremely critical, AND unusually you can access both ends of the bolt/stud. In that case the exact extension of the stud is calculated at the desired preload, and the operator measures the stretch of the stud as he turns the spanner. It is entirely possible to get that wrong as there is no "safety stop" but it enables the highest possible accuracy of all torque methods. This was done in several WW2 engines, including by Daimler-Benz and RR and Packard, although from the letters it looks like they flirted without it as it really slows down assembly, usually a few months later there was a spate of broken rods and then they brought the practice back.

A sort of half-way house was also quite often done which also doesn't have to involve any direct torque measurement at all, which is to "snug" the bolt/nut down by hand (probably several times on and off so it settles fully) then apply a pre determined ANGLE (without any torque measurement) for example 90 degrees (or whatever the bolt thread helix angle tells you gives the required extension).

Like the measurement of the stud length this almost completely eliminates the variables of oiling on the threads which can be thrown considerably off depending on the type of oil used, how much and if someone forgets to put it under the nut face and so on.

A torque wrench is actually one of the less accurate ways of doing it in fact for that reason.

Today if you want to get really expensive you can get ultrasonic methods which measure the tension in the bolt as you would in a guitar string my measuring the "note" it makes in resonance, and even instrumented washers.

Generally in something like an F1 engine today, the procedure for a cylinder head nut might be something like.

1) Bed down the nut 5 times on and off to "polish" the surfaces up and get rid of surface roughness
2) Apply 10N.m torque to the nut to snug it down
3) Turn the nut exactly 60 degrees
4) Leave it for a a set period of time
5) Undo it all and start again, go back to step 1 and repeat to ensure all the initial creep/bedding in has
been eliminated.

In mass production they cant do all that, so they "cheat" by using stretch bolts, which are actually designed so that when
you tighten them to a set angle or torque in one step they go past yield slightly. This basically guarantees the shank
has been tightened to yield, and you dont care. This is ok in a really well designed bolted joint as the stud will not
"see" much at all of the applied load on the joint from the engine running so you get away with the studs being
near failure. This is basically because the joint and stud are just springs in parallel, and so if the stud is very flexible
and the joint very stiff, most of the externally applied load wont do much to the bolt, IF the externally applied load
is less than the total preload. Hence the interest in applying the maximum possible preload.

The externally applied load here (which might be gas pressure) is the while alternative curve, and the proportion
of that which the bolt "sees" is the red wave, and the proportion the joint "sees" is the yellow wave"

The blue triangle represents the relative stiffness of the bolt to the joint. If the joint is infinitely stiff and the
bolt infinitely soft, NO external load at all would pass into the bolt no matter how hard you "pull" on the joint
UNTIL the gap separates as you apply the load equalling the preload. This does not happen a lot as a
typical head bolt might be tightened to give 8 tons preload EACH. The loads are usually determined to give the right
sealing forces, and are so far above the forces trying to fire the cylinder head off they don't do a lot in most engines !


The disadvantage of this is that each time you do it, the whole stud might get a mm longer, and so they need to be thrown
away after a couple of uses as they get long enough to hit the bottom of the thread and bottom-out.

Racers don't like that much as racing engines need to come apart a LOT and if you used stretch bolts you`d end up throwing
away a few thousand £ worth of MP35N studs every other week, so they invent complicated sets of tightening to get
to maybe 90% of yield without ever stretching it. The 10% remaining being judged to be more than enough to deal with
the actual running loads on top of the preload.

 

[Apophenia]

Wow! Thanks for that, Callum. Fascinating stuff ... and who knew that so many critical decisions were being made in how one tightened a bolt?

 

[Kanga]

My brother-in-law worked for a German pneumatics firm in the 70s. The nuts holding the tubes to the job were only supposed to be made finger tight. But old-school mechanics would check the joins, say "Hmm, they're a bit loose!" and apply the trusty shifting spanner, thus strpping the thread. Part of his job was to go around to clients to discourage this expensive practice, but he met plenty of resistance.

 

[Scott Kenny]

My brother-in-law worked for a German pneumatics firm in the 70s. The nuts holding the tubes to the job were only supposed to be made finger tight. But old-school mechanics would check the joins, say "Hmm, they're a bit loose!" and apply the trusty shifting spanner, thus strpping the thread. Part of his job was to go around to clients to discourage this expensive practice, but he met plenty of resistance.

My stepdad worked for the oil industry for a while, he had standing instructions that bolts needed to be at least X big (I think it was 1/2", but it's been a LONG time since I heard the story), anything smaller than that and the gorillas on the drill rigs would tighten the bolts till they broke. No torque wrenches in use, just a 5ft "cheater bar" extension for the ratchet handle.

However life does get very complicated once you get into the detail, and in fact the very most accurate methods are done by hand without any "direct" torque measurement gauge. Often this was done on con-rod bolts where the preload accuracy is extremely critical, AND unusually you can access both ends of the bolt/stud. In that case the exact extension of the stud is calculated at the desired preload, and the operator measures the stretch of the stud as he turns the spanner. It is entirely possible to get that wrong as there is no "safety stop" but it enables the highest possible accuracy of all torque methods. This was done in several WW2 engines, including by Daimler-Benz and RR and Packard, although from the letters it looks like they flirted without it as it really slows down assembly, usually a few months later there was a spate of broken rods and then they brought the practice back.

For IIRC Lycoming engines, the spec was torque to X and then measure stretch. Continue tightening until you reached Y stretch.

A sort of half-way house was also quite often done which also doesn't have to involve any direct torque measurement at all, which is to "snug" the bolt/nut down by hand (probably several times on and off so it settles fully) then apply a pre determined ANGLE (without any torque measurement) for example 90 degrees (or whatever the bolt thread helix angle tells you gives the required extension).

And IIRC Continentals used Torque to X ftlbs plus some angle. (I may have the makers backwards, it's been 25 years)

Like the measurement of the stud length this almost completely eliminates the variables of oiling on the threads which can be thrown considerably off depending on the type of oil used, how much and if someone forgets to put it under the nut face and so on.

The manufacturer's specs should say if they mean "wet torque", if not specified you torque dry.

If rebuilding a car engine, ARP studs are torqued wet, for example.

In mass production they cant do all that, so they "cheat" by using stretch bolts, which are actually designed so that when you tighten them to a set angle or torque in one step they go past yield slightly. This basically guarantees the shank has been tightened to yield, and you dont care. This is ok in a really well designed bolted joint as the stud will not "see" much at all of the applied load on the joint from the engine running so you get away with the studs being
near failure. This is basically because the joint and stud are just springs in parallel, and so if the stud is very flexible and the joint very stiff, most of the externally applied load wont do much to the bolt, IF the externally applied load is less than the total preload. Hence the interest in applying the maximum possible preload.

The externally applied load here (which might be gas pressure) is the while alternative curve, and the proportion of that which the bolt "sees" is the red wave, and the proportion the joint "sees" is the yellow wave"

The blue triangle represents the relative stiffness of the bolt to the joint. If the joint is infinitely stiff and the bolt infinitely soft, NO external load at all would pass into the bolt no matter how hard you "pull" on the joint UNTIL the gap separates as you apply the load equalling the preload. This does not happen a lot as a typical head bolt might be tightened to give 8 tons preload EACH. The loads are usually determined to give the right sealing forces, and are so far above the forces trying to fire the cylinder head off they don't do a lot in most engines !
[image deleted for space] The disadvantage of this is that each time you do it, the whole stud might get a mm longer, and so they need to be thrown away after a couple of uses as they get long enough to hit the bottom of the thread and bottom-out.

Aircraft types say torque-to-stretch bolts are one-time-use only.

 

[Calum Douglas]

The manufacturer's specs should say if they mean "wet torque", if not specified you torque dry

Sadly large discrepancies in the resulting friction coefficient can occur even through slightly different oils, and also how they`re applied,
so "wet" or "dry" is only the first step. Dry would be incredibly dangerous, because many bolts arrive with some oil residue
for corrosion protection, so some mechanics take them out the packet, dont apply oil and think "well thats DRY" then
apply about twice the torque necessary based on their assumption of the friction coefficient being "dry" and break the bolt.

Stuff like that is why a very small "snug" torque, followed by a calculated torque "by angle alone" is very popular (and safe)

See pg 410 THE SECRET HORSEPOWER RACE, "by stretch method" they mean measuring the length of the stud
with a precision dial gauge whilst tightening it and probably then checking with a micrometer. You can
buy special little gauges today which are (rightly) popular with small engine rebuilders.

 

[Nicknick]

@Calum Douglas : I've just had a disscussion about using four M10 bolts for a cylinder with 106 mm diameter and about 90 bar peak pressure for regular combustion. The four bolts were screwed in aluminium and I highly doubt that this would work reliable (not a race engine!) and asked if he had done such an analysis (nope).

It's not my project and I don't really have to care...

BTW, for each of you who is intrested in the bolt topic, you can find a very detailed and relative simle to understand explanation in this document, it is in German, but google will surly help.

Formeln für Schraubenverbindung bei verschiedenen Belastungsfällen

Formeln zur Berechnung einer Schraubenverbindung bei verschiedenen Belastungsfällen und Anwendungen nach VDI 2230.

www.schweizer-fn.de

 

[OliverSedlacek]

For the rest of us I recommend this book that dives deep into tightening bolts!Nuts, Bolts, Fasteners and Plumbing

 

[Calum Douglas]

For the rest of us I recommend this book that dives deep into tightening bolts!Nuts, Bolts, Fasteners and Plumbing

A very good book ! Recommended

 

[Hood]

Well I guess we can now truly say that Secret Projects Forum covers the nuts and bolts of aviation history.

 

[Calum Douglas]

@Calum Douglas : I've just had a disscussion about using four M10 bolts for a cylinder with 106 mm diameter and about 90 bar peak pressure for regular combustion. The four bolts were screwed in aluminium and I highly doubt that this would work reliable (not a race engine!) and asked if he had done such an analysis (nope).

It's not my project and I don't really have to care...

BTW, for each of you who is intrested in the bolt topic, you can find a very detailed and relative simle to understand explanation in this document, it is in German, but google will surly help.

Formeln für Schraubenverbindung bei verschiedenen Belastungsfällen

Formeln zur Berechnung einer Schraubenverbindung bei verschiedenen Belastungsfällen und Anwendungen nach VDI 2230.

www.schweizer-fn.de

Impossible to be certain without seeing the design but in most cases problems will be down to sealing line load distribution.

Assuming they are 12.9`s, then the core is 8.4mm dia at the thread root so 55mm^2 per bolt or 221mm^2 for 4 bolts

90 bar over 106mm dia gives you 80kN gas force which gives you a bolt thread min dia stress of 358Mpa
and the 12.9 grade bolt is 1200Mpa x 0.9 = 1080Mpa yield, divide by two for sensible fatigue loading =
540 Mpa giving you 1.5 load factor. This wont include stress concentration allowance for the threads, but
it sounds ok to me.

If the bolts are done up to 90% yield you get a force of 1080Mpa *0.9*55.4mm^2 = 54 kN
per bolt. So you could stop the head flying off easily with two bolts (although you`d never do it
as the sealing would never work as the clamping line load around the bore circumference will be
too uneven.

Now if they are not 12.9 grade bolts, and the joint is crappy, and they`re not done up very tight then life might get less pleasant...

 

[Nicknick]

All the bolt bosses are cut towars the cylinders, with a wall thikness at the thinner section of about 2.2 mm, given the combination of cast aluminium and 12.9 bosses, I regard it as highly critical.

The exel sheet calculations predicted 140 bar (with 95 RON, PFI) which is ridicolous and would have destroyed the engine anyway. However, there might be a pre ignition or knocking from time to time, so there should be a margin of safety.

 

[Calum Douglas]

All the bolt bosses are cut towars the cylinders, with a wall thikness at the thinner section of about 2.2 mm, given the combination of cast aluminium and 12.9 bosses, I regard it as highly critical.

The exel sheet calculations predicted 140 bar (with 95 RON, PFI) which is ridicolous and would have destroyed the engine anyway. However, there might be a pre ignition or knocking from time to time, so there should be a margin of safety.

Well if you`re very worried you`d have to send me images and all the data for analysis.

 

[Nicknick]

Nany thanks, I appriciate that but I can't do that.

i learned a traditional wisdom of production engine design which was, 8.8 for aluminium, 10.9 for GG25 and 12.9 for GJV 450....Of course, this is an oversimplyfication and thread txpes (formed or cut) and length play a role as well.

I said, this area is critical and should be analysed, but I guess the (consultand) designer doesn't have any possibilities to do so...

I might make some work for customer in future, we will see...

 

[Scott Kenny]

@Calum Douglas Your book is on the way. Arriving tomorrow, I hope.

Edit: arrived in the afternoon, so I have some reading to do.

 

[Pasoleati]

Is there any data on the power requirements to drive the injection pump in the DB 605, Jumo 211/213 and the SU/RR injection pumps in the Merlin 100 series at full power?

 

[Nicknick]

I would like to know, if anyone has considered ammonia for charge air cooling in fighter engines (or aero race engines). Ammonia would enable to cool the compressed air down to lower temperatures than water injection, because condensation will be by far less critical (depending on the humidity in the air) and evaporation will even take place when the air temperature drops below 0°C. The lower evaporation heat of ammonia (about half of that of water), would mean, that twice as much ammonia as water would have to be used for the same cooling effect (even more, if a lower temperature should be reached), but other than water, ammonia would also act as fuel, which would safe a gasoline equivalent of about half the ammonia mass.

Ammonia needs to be stored I pressurized tanks (about 10 bar), which is surly a drawback, but on the other hand, no pumps are needed and freezing is not a problem. It is also very hard to ignite, so the ammonia tanks could leak when hit by a bullet, but without creating a fire. For obvious reasons, the ammonia tank shouldn’t be in front of the pilot.

Ammonia can be properly burned when mixed with gasoline and if I remember it right, it also enhances the knock resistance of the mixture (ammonia itself is almost entirely knock free, but very slow burning).

Another great advantage at low altitude and hot temperatures (e.g. at Reno air races) is, that ammonia can be injected partially before the compressor to cool down the temperature significantly (let's say, from 30° C to estimated -15 °C), this will enable higher boost pressure by reducing the thermal load and reducing the required rpm for the compressor, in combination with an reduced compressor work.

 

[wasp850]

Is it possible to find in the book the history and development of the AJA 180 radiator, which was the annular cooler used on the Ju 88/ Fw 190C/D/Ta 152 ?

 

[HoHun]

Hi Calum,

Daimler-Benz sent a delegation to the USA in about 1936 to purchase large numbers of US made automatic machine tools,
they were not able to purchase anything like as much as they wanted because of limitations in the amount of US currency reserve
held in Germany at the time. I`m not an expert on international trade finance in the 30`s, so I`m not quite sure why they
couldn't exchange Gold for Dollars, possibly by then there were some restrictions in place on German monetary activities
which made life difficult.

On this site, I just stumbled about a comment on French machine tools, stating that a lot of the machines purchased by the French at great cost in the US in the 1938 - 40 period ended up being sent to Germany:

Essais d'avions bizarres

À la Libération, l'Arsenal reprit sa place normale mais la production industrielle était devenue comateuse, les Allemands avaient vidé nos usines des machines-outils achetées à grand frais aux USA de 1938 à 1940 et nos Alliés avaient systématiquement bombardé toutes nos usines...

Regards,

Henning (HoHun)

 

[Elan Vital]

Hi @Calum Douglas . During my archive run this week, I was very surprised to stumble upon this correspondance (in spoilers for ease of reading:

Spoiler: French archives

19th April 1940 from the Armament Minister to the Air Minister:

"I would be obliged if you let me know if you could lend the Atelier de Constructions d'Issy les Moulineaux (AMX) a Daimler Benz or Junkers engine with gasoline/petrol injection for 3 months. This establishment must indeed study a petrol injected engine for armored combat vehicles."

Reply on May 2nd:

"In response to your letter. I have the honor to let you know that the Air Technical Services possess no fuel injected Daimler Benz or Junckers engine susceptible to be immediately lent to the AMX. The only running DB 601s are:
- one engine equipping a Messerchmitt 109 for tactical trials
- 2 engines (under overhaul) meant to equip a special Dornier aircraft that we are finishing to repair.
- one engine which is completing bench trials at the Engine Trials Center (old engine of the above-mentionned Me 109 which will be used as a spare)

Total: 4

Moreover, there is a single Junkers 211 at the trials laboratory of Hispano-Suiza at Colombes, which has been repaired using (parts from other damaged engines of the same type)."

Letter from the Armament minister on the 13th of May to the AMX's director, Engineer Duréault:

"I have the honor to let you know that the Air Minister can currently not put a Daimler Benz or Junkers engine at your disposal. However, there exists a single Junkers 211 at the trials laboratory of Hispano-Suiza at Colombes, which has been repaired using parts from other defective engines of the same type. The Hispano-Suiza company must imminently proceed to the trials of this engine, after which it can be lent to your establishment.

It would be likely profitable if an engineer from the AMX attended these trials, and if you want, the Department of the Air will assist you."

---

I recall reading about that Jumo 211 in the Secret Horsepower Race, but nonetheless it is interesting to know about the whereabouts of all these engines so late before France's fall; and even moreso that a tank company (AMX) would be interested in these engines in particular.

What surprises me is that since 1937/38, AMX had been going all in on diesel engine development. So either there is a misunderstanding in the letters and AMX just wants to study the fuel injection system in general for application to their diesel developments, or they recently started work on another line of engines, this time using gasoline/petrol.

At the same time, the ARL design bureau was cooperating with the Edgar Brandt Establishment to develop a petrol tank engine with "direct fuel injection"*. A postwar account on clandestine work during the occupation said this engine used the Fabian injection process, but I have been unable to find what that meant. I am also not sure if they really mean direct fuel injection as in german aircraft engines, or if it just the way the French talked about another form of injection.

*The ARL noted that fuel injection for tank engines could decrease specific fuel consumption by as much as 15-20% while increasing power density by 20% on a weight basis.

All of this was part of a folder on tank engine development in 1939/1940, as part of a more general archive digging session on French AFV development.

 

[tomo pauk]

@Calum Douglas - would you be so kind to post the table of contents of the Italian expanded 'version' of your book?

 

[Calum Douglas]

@Calum Douglas - would you be so kind to post the table of contents of the Italian expanded 'version' of your book?

That wont tell you anything about what's in the book, its just one year per chapter as per the English version. It just says "Intro" .."1939"..."1945"

 

[tomo pauk]

That wont tell you anything about what's in the book, its just one year per chapter as per the English version. It just says "Intro" .."1939"..."1945"

Okay.
Hopefully you will understand me asking questions, since I'm willing to buy the book even though (or because ) I have the English language version. Is there an overview/index denoting the more obscure & interesting Italian engines, like the Fiat engines for the racing aircraft, or the Piaggio high-altitude engines?

 

[Calum Douglas]

The extra material in that version comes from: 1) The material I had to originally cut from the Mortons version in English (which isnt Italian material), and 2) Biographical material about Italian designers, and a few extra details on Italian licence building etc.

I do not think you will find more about the specific items you describe inside.

 

[tomo pauk]

The extra material in that version comes from: 1) The material I had to originally cut from the Mortons version in English (which isnt Italian material), and 2) Biographical material about Italian designers, and a few extra details on Italian licence building etc.

I do not think you will find more about the specific items you describe inside.

Okay, thank you anyway

 

[Elan Vital]

Calum, did you look into the Anglo-French coordination committee papers at the National Archives? The CAB 85 series are about all this, and while they tend to be mere memos or very short minutes with not much information, they might further guide research.

The CAB 85-9 and -10 in particular were talking about coop in the field of aviation, and this is where I learnt that the plans to provide Merlins to France were still on (and they were talking about the Merlins in France that should be sent back home to equip RAF aircrafts to send back to France given the risk of them falling into German hands otherwise).
It turns out that given the time it would take for the French Fordair and SIGMA companies to license produce British engines with normal support from Rolls-Royce and Bristol, it was decided to develop the production of british engines independently of british sources, which would delay the start of production to late 1941.
For Fordair, it was decided to abandon the Merlin since it might be outdated by then, and the Vulture was suggested instead.
For the SIGMA, it was suggested to go forward with the Hercules 6 SM instead of the 3 SM, for production starting in October 1941.

 

[Calum Douglas]

Calum, did you look into the Anglo-French coordination committee papers at the National Archives? The CAB 85 series are about all this, and while they tend to be mere memos or very short minutes with not much information, they might further guide research.

The CAB 85-9 and -10 in particular were talking about coop in the field of aviation, and this is where I learnt that the plans to provide Merlins to France were still on (and they were talking about the Merlins in France that should be sent back home to equip RAF aircrafts to send back to France given the risk of them falling into German hands otherwise).
It turns out that given the time it would take for the French Fordair and SIGMA companies to license produce British engines with normal support from Rolls-Royce and Bristol, it was decided to develop the production of british engines independently of british sources, which would delay the start of production to late 1941.
For Fordair, it was decided to abandon the Merlin since it might be outdated by then, and the Vulture was suggested instead.
For the SIGMA, it was suggested to go forward with the Hercules 6 SM instead of the 3 SM, for production starting in October 1941.

I have a little on that on page 132,

However I dont have copies of any CAB-85`s so I`ll have a look at those, thanks.

Generally French material has been almost as difficult as Russian to get materials on, French museums have been... lets politely say "unhelpful" (even when I had French friends IN France approach them for me!).

 

[Calum Douglas]

Hour long video interview with Physicist John Griffith (A A Griffiths son, who proposed the axial turbine layout rather than centrifugal compressor) is now done. Should set a lot of the Whittle vs Griffith controversy story straight.

Kind thanks to the Griffith family for inviting me to their home to help get the story of his work and the jet straight.

Video will appear on my YouTube channel in early January.

 

[Elan Vital]

Looking forward to seeing that video when it uploads.

Of potential interest regarding the French side of the horsepower race, a relatively detailed article on the late Gnome-Rhône radials: http://www.aerostories.org/LeMag/AB-Derniers-Gnome-Rhone.pdf

Dates from October 2024, so no wonder it took so long for this article to pop up!

Some aspects remain obscure but it is at least one of the more detailed pieces on that subject, going over some structural developments and the superchargers.

It is in pdf form so it should be a lot easier to run into a translator than other secondary sources.