High performance engines
Discussion
Just watching Nat Geo+1 supercars GT-R one section is about the engine and they state high performance engines cannot be mass produced they are all hand made.
Now this is the same for every make but why is it?
What is the key aspect that as yet we cannot make a machine be able to create a perfect sportscar engine?
Is it the shim size being so accurate that a mass production line cannot be matched.
The mechanics o the production line claim building by hand reduce tolerances by 50% over mass produced units.
(Also amazingly once built they spend over one hour tuning every engine whilst running on a test bed so that they know power and torque outputs).
Incredible.
Now this is the same for every make but why is it?
What is the key aspect that as yet we cannot make a machine be able to create a perfect sportscar engine?
Is it the shim size being so accurate that a mass production line cannot be matched.
The mechanics o the production line claim building by hand reduce tolerances by 50% over mass produced units.
(Also amazingly once built they spend over one hour tuning every engine whilst running on a test bed so that they know power and torque outputs).
Incredible.
I imagine that it's the economies of scale preclude the economics of it. I.e. we build a lot of normal engines, and so it's "worth" designing the tool and robots to assemble them. But we don't make many high performance engines (by comparison) so it's not worth all the investment in automatic large parts of the process. Particularly because performance engines are more likely to have frequent spec changes as more power is eeked out over their life. Humans are much easier to adapt to changes in the process, whereas a robot will require reprogramming etc.
ETA - Obv a lot of the stuff on high performance engines is created by CAD/CAM systems, but again these are often better suited to low volume applications rather than high volume.
I guess it's similar to art works; one off commissions are hand painted, but cheap reproductions are screen printed. Volume vs. Price.
ETA - Obv a lot of the stuff on high performance engines is created by CAD/CAM systems, but again these are often better suited to low volume applications rather than high volume.
I guess it's similar to art works; one off commissions are hand painted, but cheap reproductions are screen printed. Volume vs. Price.
I can't imagine that it's impossible to mechanically produce them - machines can generally manage much tighter tolerances than people can - in many cases machines be be accurate to within a few atoms (look at the machine they used to make the Hubble mirrors if you want an extreme example). I suppose the kinds of machines required to produce such fine tolerances are just too expensive, though.
kambites said:
I can't imagine that it's impossible to mechanically produce them - machines can generally manage much tighter tolerances than people can - in many cases machines be be accurate to within a few atoms (look at the machine they used to make the Hubble mirrors if you want an extreme example). I suppose the kinds of machines required to produce such fine tolerances are just too expensive, though.
Right, but it's not about a human producing something to a tight tolerance; a lot of it is about humans taking a range of machine produced parts (with their inevitable variation), and selecting which ones to put together to get the best overall results.Be intereresting to know though whether cylinder heads could be ported on a 5 axis machine, with an acceptable surface finish?
Mave said:
Right, but it's not about a human producing something to a tight tolerance; a lot of it is about humans taking a range of machine produced parts (with their inevitable variation), and selecting which ones to put together to get the best overall results.
Be intereresting to know though whether cylinder heads could be ported on a 5 axis machine, with an acceptable surface finish?
Well yes, but the point is that with suitably accurate machining in the first place, there will be no significant variation in the machine produced parts; certainly not variations big enough for a human to detect. Be intereresting to know though whether cylinder heads could be ported on a 5 axis machine, with an acceptable surface finish?
The problem is simply that this level of accuracy of mechanisation is too expensive.
Edited by kambites on Monday 26th November 21:10
kambites said:
Well yes, but the point is that with suitably accurate machining in the first place, there will be no significant variation in the machine produced parts; certainly not variations big enough for a human to detect.
The problem is simply that this level of accuracy of mechanisation is too expensive.
...but that fundamentally assumes a machined finish. Not everything has a machined finish.The problem is simply that this level of accuracy of mechanisation is too expensive.
Edited by kambites on Monday 26th November 21:10
There are also some areas of even machined finish where machining access / sequence means you can't get the tolerance you might want for a performance engine
in 2012, it's b*ll*cks!
These days, any machining / build process by a machine is to a higher tollerance than by a human. (and will be 300% more repeatable too)
The issue is cost. To "tool up" to make ultra tight tollerance components, costs a lot of money, so if you are going to make only a few, it's cheaper to just pay someone to build them slowly and carefully. Plenty of manufacturers run very very tight tollerances (tighter than for the average race engine that can be "hand fettled" during assembly to make it work properly (for example ferrari, BMW etc)
The real reason most mass produced engines are not it a very high state of tune is twofold:
1) cost - most mass produced things simply aren't required to be ultra-high performance, or the cost of making them so is not tollerable in the end product
2) Worldwide compiliance with conditions and durability. For example, Toyota aren't going to do a specific engine calibration to run on the particularly good fuel you get in, say germany (102RON) for there avensis. You make high volume things cheap (= more profit) by making things common and simple.
These days, any machining / build process by a machine is to a higher tollerance than by a human. (and will be 300% more repeatable too)
The issue is cost. To "tool up" to make ultra tight tollerance components, costs a lot of money, so if you are going to make only a few, it's cheaper to just pay someone to build them slowly and carefully. Plenty of manufacturers run very very tight tollerances (tighter than for the average race engine that can be "hand fettled" during assembly to make it work properly (for example ferrari, BMW etc)
The real reason most mass produced engines are not it a very high state of tune is twofold:
1) cost - most mass produced things simply aren't required to be ultra-high performance, or the cost of making them so is not tollerable in the end product
2) Worldwide compiliance with conditions and durability. For example, Toyota aren't going to do a specific engine calibration to run on the particularly good fuel you get in, say germany (102RON) for there avensis. You make high volume things cheap (= more profit) by making things common and simple.
I saw a bit on TV about corvettes, they make it a selling point of having the "hand assembled engine", the assembler signs each unit he makes.
What they showed was a computerised system that told the assembler exactly what to do, the crank was lowered in via a computer controlled arm, the system sensed when the crank was in place correctly.
When it came to tightening any bolts the machine told the assembler which bolt to put the computerised torque wrench on via an on screen display. When in place he pressed a button, which did the bolt up to the specified torque, computer then showed which bolt was next.
Seemed to me the assembler was the lackey, the computer was doing the important bits.
What they showed was a computerised system that told the assembler exactly what to do, the crank was lowered in via a computer controlled arm, the system sensed when the crank was in place correctly.
When it came to tightening any bolts the machine told the assembler which bolt to put the computerised torque wrench on via an on screen display. When in place he pressed a button, which did the bolt up to the specified torque, computer then showed which bolt was next.
Seemed to me the assembler was the lackey, the computer was doing the important bits.
Max_Torque said:
in 2012, it's b*ll*cks!
These days, any machining / build process by a machine is to a higher tollerance than by a human. (and will be 300% more repeatable too)
b*ll*cks.These days, any machining / build process by a machine is to a higher tollerance than by a human. (and will be 300% more repeatable too)
I've seen plenty of components where the machining access means you just can't get enough stiffness into the tooling to get a repeatable position, eg failing to achieve a 0.5mm positional tolerance on a series of holes on a PCD. I've seen a shaft where we spent £360K on component specific machining tooling before we could actually machine it to a 10 thou bore profile. That's before we talk about cast components.
I'd agree that generally machine processes are better, but to say that they are always better is a bit of over simplification (and don't get me going on how CMM results are interpreted on on automated system...)
Mave said:
Max_Torque said:
in 2012, it's b*ll*cks!
These days, any machining / build process by a machine is to a higher tollerance than by a human. (and will be 300% more repeatable too)
b*ll*cks.These days, any machining / build process by a machine is to a higher tollerance than by a human. (and will be 300% more repeatable too)
I've seen plenty of components where the machining access means you just can't get enough stiffness into the tooling to get a repeatable position, eg failing to achieve a 0.5mm positional tolerance on a series of holes on a PCD. I've seen a shaft where we spent £360K on component specific machining tooling before we could actually machine it to a 10 thou bore profile. That's before we talk about cast components.
I'd agree that generally machine processes are better, but to say that they are always better is a bit of over simplification (and don't get me going on how CMM results are interpreted on on automated system...)
Meldonte said:
Interesting question and pretty sure mave has hit the nail on the head! Sometimes machines are far less accurate as they have limited ways of knowing when they are going out of tolerance with wear or compliamce or other external factors unless you have in process inspection set up. It's not that the parts are hand made - they aren't. But they are hand assembled to ensure that you can pair up the right sets of mating components with each other to give you the correct fits - it's called blue printing. Cost does come into it of course, blue printing allows you to open up tolerances on lots of parts and as long as there is significant variance in a batch you will ensure you will always be able to marry up components with minimal rework. It's cheaper to measure something precisely than it is to manufacture it to the same precision.
Are you suggesting that this too couldn't be automated? Surely the man doing the assembly will be measuring a component, and then choosing one from a set of other components which fits it? That could easily be done by machine.SO if your machine can't get a 0.5mm tollerance (which is a country mile when it comes to tollerances) due to a tooling limitation, how is a bloke with a hand drill going to do any better??
No, the limitation is still cost. If you can design tooling access by hand, you can design for it by machine.
No, the limitation is still cost. If you can design tooling access by hand, you can design for it by machine.
Mave said:
kambites said:
I can't imagine that it's impossible to mechanically produce them - machines can generally manage much tighter tolerances than people can - in many cases machines be be accurate to within a few atoms (look at the machine they used to make the Hubble mirrors if you want an extreme example). I suppose the kinds of machines required to produce such fine tolerances are just too expensive, though.
Right, but it's not about a human producing something to a tight tolerance; a lot of it is about humans taking a range of machine produced parts (with their inevitable variation), and selecting which ones to put together to get the best overall results.Be intereresting to know though whether cylinder heads could be ported on a 5 axis machine, with an acceptable surface finish?
Mave said:
...but that fundamentally assumes a machined finish. Not everything has a machined finish.
There are also some areas of even machined finish where machining access / sequence means you can't get the tolerance you might want for a performance engine
Machines can produce any finish that people can, in theory. There are also some areas of even machined finish where machining access / sequence means you can't get the tolerance you might want for a performance engine
I don't believe there is any element of manufacturer where a machine cannot produce finer tolerances that a human can detect by eye or feel. You just need a good enough machine, which can be rather expensive.
kambites said:
Machines can produce any finish that people can, in theory.
I don't believe there is any element of manufacturer where a machine cannot produce finer tolerances that a human can detect by eye or feel. You just need a good enough machine, which can be rather expensive.
Can you get 5-axis adaptive grinding stations yet?I don't believe there is any element of manufacturer where a machine cannot produce finer tolerances that a human can detect by eye or feel. You just need a good enough machine, which can be rather expensive.
As an aside, I also don't think we are necessarily capable of defining the tolerances of a part in a way which machines can measure and interpret correctly.
kambites said:
Well yes, but the point is that with suitably accurate machining in the first place, there will be no significant variation in the machine produced parts; certainly not variations big enough for a human to detect.
The problem is simply that this level of accuracy of mechanisation is too expensive.
What sort of variation do you think a human can detect?The problem is simply that this level of accuracy of mechanisation is too expensive.
Edited by kambites on Monday 26th November 21:10
If you assemble a stack of components, even if each has a tight tolerance, you quickly get something that's detectable to humans.
Gassing Station | General Gassing | Top of Page | What's New | My Stuff