CNC Your Own Parts

[MrAl]

I believe a quality bearing would required grinding to the final tolerances.
volvolugnut

Hi,

Yes that's what they did.
From memory, first the machining to the rough size. Then heat treating in a molten salt vat for some minutes, then quenching. Then there might have been another machining cycle. Then they did 'carbonizing' to the piece, then the final grinding to the final shape.
The carbon adding part I think they call it Carburization.
The heat treating schedule was fairly ridged. The piece went into the molten salt bath and the timer times it for so many minutes then it is removed and immediately quenched. The hardness is then checked on random samples.

Whoa, molten salt bath? That alone puts it out of reach for the DIY person. Doesn't it take several thousand degrees to melt a salt?

Hi,

Well regular salt probably around 1500 degrees, but I do not know if they had any other additives too.

There are other ways though you can look around the web for heat treating at home. I think you can even add carbon to the outer surface to get the main wear surface even harder. I've never done any of this only the blow torch heating and quenching method part myself. This isn't the main concern though I don't think. The main concern is getting the final part to the exact dimensions. Because of that, I think we would need a higher priced CNC machine than the ones being solve for engraving and stuff like that in order to make something as fine as a bearing. I don't know for sure though I've never had to do that except for very slow moving, low power machinery. For other simpler parts though I would think it would work. For example, parts that don't have to move or move slowly like for the gas pedal linkage, stuff like that. Big parts would have to be made in pieces, welded together and ground down, or bolted together.

I guess the main point this this thread is that if you can't get a part, this may provide a path to success. With my old Hyundai I had to make a part I could not get anywhere, but it did not require any machining, so it was much simpler.

[BlackBart]

This brings up a production machining question I've had as far as accurate dimensions.
If there is a solid shaft (not a bolt) that passes through a round hole machined in another piece, they both can't be "exactly" 19mm. Is one or the other just a micron smaller or larger diameter so they fit?

[volvolugnut]

This brings up a production machining question I've had as far as accurate dimensions.
If there is a solid shaft (not a bolt) that passes through a round hole machined in another piece, they both can't be "exactly" 19mm. Is one or the other just a micron smaller or larger diameter so they fit?

Very perceptive question!
The answer has been worked out by engineers for decades. There are standards for sliding and press fits of mating parts. Sliding fits, of course, have to have more clearance than press fits. Tables are made with a range of tolerances for each class of fit. When an engineer or designer is creating the part drawings, they look up the desired class of fit and find the tolerances for the internal and external parts. These tables are in Machinist Handbooks. These handbooks show up used occasionally.
In the last several decades ASME created a system of drawing notation called Geometric Dimensioning and Tolerancing (GD&T). It is detailed and tedious, but allows clear communication of the component dimensions if everyone follows the standard correctly.
My brain had to retrace some old, old file space to remember some of this.
volvolugnut

[MrAl]

This brings up a production machining question I've had as far as accurate dimensions.
If there is a solid shaft (not a bolt) that passes through a round hole machined in another piece, they both can't be "exactly" 19mm. Is one or the other just a micron smaller or larger diameter so they fit?

Very perceptive question!
The answer has been worked out by engineers for decades. There are standards for sliding and press fits of mating parts. Sliding fits, of course, have to have more clearance than press fits. Tables are made with a range of tolerances for each class of fit. When an engineer or designer is creating the part drawings, they look up the desired class of fit and find the tolerances for the internal and external parts. These tables are in Machinist Handbooks. These handbooks show up used occasionally.
In the last several decades ASME created a system of drawing notation called Geometric Dimensioning and Tolerancing (GD&T). It is detailed and tedious, but allows clear communication of the component dimensions if everyone follows the standard correctly.
My brain had to retrace some old, old file space to remember some of this.
volvolugnut

And I must say that's a very good answer!
I used to have a machinists handbook that was 2 inches thick with all kinds of dimensional info in it but not sure what happened to it over the years.
For sliding mechanisms, they often use linear bearings. These things are a wonder of machining and have some interesting characteristics. One type has some sort of hard rubber sleeve that goes over a steel sliding bearing and the rubber is mounted in a pillar mount, and the pillar mount is bolted down to something so the shaft can slide though it. The other interesting question is how does the shaft line up perfectly with the sleeve bearing, because any tilt of the shaft will mean the bearing will rub more in some places than others. The cylinder of the sleeve bearing has to be aligned with the shaft so that there is little angle between the axial centers. However, the rubber allows the tiniest misalignment as it can give a very small amount and allow the inner metal bearing to be perfectly aligned as the tiny shaft angle demands. Pretty nice.
There another interesting part. A lead screw and follower. How can you get a follower to follow the lead screw as it turns without any slop. Even a tiny bit of slop in the mechanism will mean when the lead screw changes direction, the follower threads will move slightly in a direction that messes up the position. The answer is to use a spring loaded thread follower. The spring keeps the lateral movement to an absolute minimum without causes any binding of the threads and the lead screw.
These kinds of mechanisms are what keeps a CNC movement to a very close tolerance. There is zero play no matter how you try to skew the spindle head. It's pretty amazing really.
They have different types of linear bearings too some have little balls inside that track the shaft. They would have some way to deal with small changes in dimensions in the steel shaft or bearing itself over temperature. I don't remember all of the types now though.
A lot of stuff these days use these bearings too, like line printers, ink jet printers, 3d printers, you name it.
One of my least favorite bearings was the self oiling bronze bearing. These were used in home fans and may still be used in them. If they get dirty from dust they get gunked up and it becomes hard to turn the fan blades, so the motor eventually seizes. The fix is to take the armature out and clean the bearings and shaft and use some oil. The best type seems to be synthetic motor oil.
 
BTW I may have mixed up two of the steps in the making of a regular bearing in a previous post it's been a very long time since I had to understand that kind of thing. I think they may have done the final machining before the carburization step. I seriously doubt any of us here would be doing that anytime soon, but you never know.

[volvolugnut]

More rolling bearing trivia:
Rolling bearings have a life rating expressed as L subscript 10. This is the expected life in millions of revolutions with the input to the equation the RPM. The rating is for 90% reliability as there is always variation in the life. There are other more detailed bearing life rating methods.
It used to be the very best bearings were used in printing presses for printing the best registration of the various colors. And the very best printing presses were used for the glossy, color magazines like Playboy.
I learned a lot about bearings while designing high pressure pumps with spherical roller bearings. SKF makes some of the best bearings in the world.
volvolugnut

[BlackBart]

The answer has been worked out by engineers for decades. There are standards for sliding and press fits of mating parts. Sliding fits, of course, have to have more clearance than press fits. Tables are made with a range of tolerances for each class of fit. When an engineer or designer is creating the part drawings, they look up the desired class of fit and find the tolerances for the internal and external parts. These tables are in Machinist Handbooks. These handbooks show up used occasionally.

This is fascinating! I've always wondered. Mathematically the 19mm shaft and the 19mm hole occupy the same space at the full radius.

Great explanation, thank you.

[BlackBart]

When I was 17, working on my '73 Ford Capri, I took some assembly apart and found a stripped bolt. Big bolt. Its a Sunday of course, no parts stores open in those days, and I need it Monday morning for work.

Our neighbor across the alley, Bud Hashburger, was the retired head mechanic of the big downtown Denver Cadillac dealer. He was a wizard. I showed him this bolt (metric of course). He put the cigarette back in his mouth, set his half tall Miller beer down, looked it over quickly, then walked over and grabbed a measurement tool. He grabbed a raw stock bolt (head with a smooth shaft) and popped it in his big mill lathe. (Who has metric bolt stock laying around?) A cutter traveled down the shaft at exactly the right pace as it spun to cut perfect threads. Handed it to me - "that should do it."

I was amazed. I'm still amazed.

[MrAl]

When I was 17, working on my '73 Ford Capri, I took some assembly apart and found a stripped bolt. Big bolt. Its a Sunday of course, no parts stores open in those days, and I need it Monday morning for work.

Our neighbor across the alley, Bud Hashburger, was the retired head mechanic of the big downtown Denver Cadillac dealer. He was a wizard. I showed him this bolt (metric of course). He put the cigarette back in his mouth, set his half tall Miller beer down, looked it over quickly, then walked over and grabbed a measurement tool. He grabbed a raw stock bolt (head with a smooth shaft) and popped it in his big mill lathe. (Who has metric bolt stock laying around?) A cutter traveled down the shaft at exactly the right pace as it spun to cut perfect threads. Handed it to me - "that should do it."

I was amazed. I'm still amazed.

Cool story
Nice to hear about this stuff.
I've done some lathe work in the distant past too.

[volvolugnut]

Do you want to hear more large pump bearing design information?
volvolugnut

[BlackBart]

Oo! Oo! Yes!