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

[TomcatViP]

Regarding the discussion on bolts and torques for the Merlin engine, Packard replaced the chaos of English Imperial bolts made of various standards and quality with the American one, a much more rational choice. Hence, all the tightening torques had to be calculated again (different thread, different diameter, dissimilar materials...).

Also, @Calum, I would not recommend you give advice for bolts installation based on what seems to be your F1 racing experience. Bolts have to be used with much more relaxed constraints in the normal usage (for example, a 50% load margin is a reasonable and efficient one in most case).
Some bolt manufacturers have published very well done and comprehensive bolts datasheet that summarize most user case with practical indications.
In most cases, readers will face a problem around a design built according to a norm. Datasheet like this are then the perfect and safe source to search for such information.
Calcuting specific bolts tightening pattern is the work of an Engineer.

 

[Scott Kenny]

[...]

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.

Simple mechanical direct injection works the same way in gas and diesel engines.

You have a high-pressure fuel pump that has to exceed the pressure in the cylinder to push fuel past a check valve. Said fuel pump usually has a variable displacement chamber per cylinder, and at the right time in the compression stroke the fuel pump quickly slams the appropriate load of fuel into the line to the injector, through the check valve.

I mean, the inside of a naturally aspirated diesel engine has ~20bar just from the compression ratio, not counting the effects of the heat due to compression. IIRC the big R3350 gas engines used a 3500psi injector pump, and the (really crappy) GM 350 Diesel used a similar setup.

You can make a fancier direct injection system that uses a constant volume fuel pump and has the injector triggered by a camshaft similar to the one used to open a poppet valve in the cylinder. That's the method that the old Detroit Diesels used (-53, -71, and -92 series engines). Works well, needs a special dial indicator base to perfectly time. I'm not sure how exactly they controlled how much fuel went in unless the whole fuel cam moved.

 

[Calum Douglas]

Regarding the discussion on bolts and torques for the Merlin engine, Packard replaced the chaos of English Imperial bolts made of various standards and quality with the American one, a much more rational choice. Hence, all the tightening torques had to be calculated again (different thread, different diameter, dissimilar materials...).

Also, @Calum, I would not recommend you give advice for bolts installation based on what seems to be your F1 racing experience. Bolts have to be used with much more relaxed constraints in the normal usage (for example, a 50% load margin is a reasonable and efficient one in most case).
Some bolt manufacturers have published very well done and comprehensive bolts datasheet that summarize most user case with practical indications.
In most cases, readers will face a problem around a design built according to a norm. Datasheet like this are then the perfect and safe source to search for such information.
Calcuting specific bolts tightening pattern is the work of an Engineer.

Pretension to at least 75% yield is a perfectly normal target in any sensible design which has nothing to do with racing. Oh and I`ll say whatever I like thank you very much.

There is no scenario in which torque to 50% is quote "efficient". Provide sources and data for this >


HJ

 

[TomcatViP]

Baah

 

[F119Doctor]

Pretension to at least 75% yield is a perfectly normal target in any sensible design which has nothing to do with racing. Oh and I`ll say whatever I like thank you very much.

There is no scenario in which torque to 50% is quote "efficient". Provide sources and data for this >

View attachment 754063
HJ

I’ve never gotten into the design weeds of any specific bolted joint, but engineering school it was noted that a properly designed joint should always have the threaded fasteners clamping load higher than the operating loads. Of this is the case, the fastener would never see any alternating loads and thus would never fatigue. In addition, I know that several auto manufacturers torque their piston rod big ends to the yield point, which would be approximately 100% of preload capability while negating preload variability due torque measurement and friction variations.

 

[Pasoleati]

There is one puzzling question: Why didn't Wright go totally to direct injection with the R-3350 post-war as at least some strictly post-war models (according to Jane's CA models; BA and BD models had direct injection) had single-point injection to the impeller? It is very annoying that there is no thorough book on the R-3350.

 

[Nicknick]

I’ve never gotten into the design weeds of any specific bolted joint, but engineering school it was noted that a properly designed joint should always have the threaded fasteners clamping load higher than the operating loads. Of this is the case, the fastener would never see any alternating loads and thus would never fatigue. In addition, I know that several auto manufacturers torque their piston rod big ends to the yield point, which would be approximately 100% of preload capability while negating preload variability due torque measurement and friction variations.

Tightening the bolts over the yield point is done to minimize tolerances in the bolt forces. Under max. operational load, the bolts will stretch even a little further, but than the stretching will come to an end. The fasteners still don't fatique because the load variation is small enough.

Keep in mind, that most of the tightening moment is used to overcome friction and not to produce the desired axial force. Friction is allways hard to predict, so tightening bolts by a given moment will cause significant variations in bolt forces. If roundness matters (cylinder and conrod big ends), overstretching the bolts is the common method. Of course, there are alternative solutions (like measuring the stretching of a conrod bold with a micometer) but those need much more effort.

 

[Nicknick]

@Scott Kenny : Diesel injection pumps usually have a constant pumping cylinder volume and a variable opening timing. The Injection ends, when the pressure inside the pumping cylinder is relieved. This method enables a constant injection begin with a variable ending, which was good enough in the old days.

In modern Common rail pumps, it is slightly different, the plunger moves upwards and the built up of pressure starts when the intake valve is beeing closed. This method allows the use of an electromagnetic valve with very little valve actuating forces, since it it pressed in the seat by the chamber pressure.

I think, the Detroit Diesels had a system with pressure increase by an differential piston and were fed with low pressure oil.

 

[F119Doctor]

Tightening the bolts over the yield point is done to minimize tolerances in the bolt forces. Under max. operational load, the bolts will stretch even a little further, but than the stretching will come to an end. The fasteners still don't fatique because the load variation is small enough.

Keep in mind, that most of the tightening moment is used to overcome friction and not to produce the desired axial force. Friction is allways hard to predict, so tightening bolts by a given moment will cause significant variations in bolt forces. If roundness matters (cylinder and conrod big ends), overstretching the bolts is the common method. Of course, there are alternative solutions (like measuring the stretching of a conrod bold with a micometer) but those need much more effort.

In modern P&W turbine engines, the most critical threaded fastener joints have the nut torqued to a lower pre-torque level, followed by an angle of turn. This is usually followed by loosening the fastener, then reapplying the pre-torque and verifying it returns the the previous position. If it does, then the final angle of turn is applied.

If you have access to both ends of a bolt, physically measuring stretch is the most accurate method for determining the preload applied.