Question for Atlas - cost of wheels

[Ed Kapuscinski]

Being an 'economic' modeller, I have to agree. If they work as is, no need to change.
Layout I used (not mine) to scenic, operate and test is gone now. So even less need for expense.
d

For me it's the annoyance of shorting things out all the time.

[peteski]

For me it's the annoyance of shorting things out all the time.

All the time? Interesting. I don't seem to notice that problem on the layouts I operate.

[nkalanaga]

Ed:  Plastic also has the advantage on my dual gauge turnouts.  If the wheels are even slightly out of gauge, shorts can be a problem, as many rails as there is in a small space.  Plastic wheels are short-proof!

As for the cost to change the axle length, it could be more complicated than it looks, depending on how they make the axles.  If the length is computer-controlled, then yes, it's just a matter of changing a file.  On the other hand, if the length is mechanically controlled, with a physical tool setting the length, it would require new tooling.  We have machines at the factory where I work that measure things both ways.  The computer/sensor versions are nice - as long as everything is in adjustment and the sensors work.  The mechanical stops are a pain to adjust, and sometimes to work around, but once set, they work every time.

[peteski]

The computer/sensor versions are nice - as long as everything is in adjustment and the sensors work.  The mechanical stops are a pain to adjust, and sometimes to work around, but once set, they work every time.

There are quite a few numerically-controlled (NC) lathes and mills being used all over the world. They are very reliable and widely used for repeatable tasks. They are used to make all sorts of parts for all sorts of items we use everyday (like cars). If the were as unreliable as you imply then our cars would have bunch of loose fitting parts. 

[nkalanaga]

Oh, I'm sure the lathes and mills are fine, but OUR machines aren't that reliable.  They also aren't lathes and mills, but assembly machines.

My real point was that, if the axle length is set programmatically, changing it would be simple, and almost cost-free.  But if the length is set mechanically, it could be more difficult. 

[peteski]

Oh, I'm sure the lathes and mills are fine, but OUR machines aren't that reliable.  They also aren't lathes and mills, but assembly machines.

But we were discussing high volume factory setup for doing wheels in this thread, not hobbyist's lathes.  Whatever those industrial machines are, numerical control is very reliable. I'm also not following you as to how an assembly machine would be responsible for controlling axle length. The axles are single pieces of metal, machined on some sort of lathes (probably numerically controlled).

[nkalanaga]

I KNOW how axles are made - you cut a piece of rod to length, and then taper the ends.  No, our assembly machines have nothing to do with the process.  But there has to be SOMETHING that determines the final length. 

Computer-controlled lathes can do that, as you said.  A very basic lathe, with a mechanical stop to limit the cutting distance, can also do it.  I wouldn't be surprised if there are other ways of doing it.  All I was saying is that, since we don't know how it's done at that factory, it could be very easy, or not so easy, to make a different length.

[narrowminded]

While it doesn't change a thing knowing this sometimes it's interesting to understand how things might be done.  So, for the curious, here goes with a cursory description of how that's likely done. 

To be done most efficiently and absurdly accurate you would use a small parts CNC lathe with sub spindle.  The small parts reference means a high speed spindle (20,000+/- RPM) for proper speed (SFPM), usually intended for parts around 1" diameter or less.   It does not necessarily mean a physically small machine as they are typically fully cabineted with tool changers, spindles, hydraulics, and all the necessary doo-dads (a highly technical term meaning "stuff").   The sub spindle means it can grip the part from both ends that can be independently controlled, similar to the way a tailstock center might be in conventional machines, but powered.  It's a second spindle placed opposing the first and is powered synchronous with the main spindle, as well as travelling in and out on ways. 

This allows loading a full length bar (or more with an automatic bar feeder) of material in the end of the machine which automatically advances as bar is consumed. You start by machining the first end in a conventional fashion with the bar projecting out of the main spindle so it can be machined on the sides and end.  When that operation is complete the powered sub spindle, which is matched in RPM to the first, advances and grips the part. Next release the grip in the main spindle, pull the part out to parting length, re-grip, part (cut) off leaving the part gripped and rotating in the sub spindle while the bar is left gripped in the main spindle.  Then, moving the sub spindle further out with the parted off piece gripped and rotating, projecting from the sub spindle, machine the second end as you did the first end.  The spindle(s) never shuts off.  The part is finished at both ends, extremely accurate due to single setup, and requiring no second operation or deburring.  Cleaning, maybe. 

This is how many small parts with two ends machined are done in production.  Often the hardest part with pieces this tiny is collecting the finished parts, not losing them in the chip tray, but there are methods for that, too.   Properly set up this can run almost unattended too, meaning an operator could probably keep several similar machines fed with material, occasional parts inspection, and reliably producing good parts.  Plus, once set up to run these specific parts, a part that will use the same tools but might be a different dimension, within .010" or so, it's literally a few edits and it's done, starts producing that new part. That's how you can get finished wheelsets at the prices you can when they are produced in the tens of thousands at a time and why I said that the cost to change that axle length would be so small as to be hard to calculate.

It could be done without the sub spindle but that requires a second operation.  That's more labor intensive, handling the part twice and not being able to leave the machine but with a proper set up this too could be just as reliably accurate, but more expensive.  And if there aren't too many to do (a few hundred as opposed to thousands) it can be viable but will cost more and is why short runs or customs DO cost more.  And often, a LOT more, and totally legitimate, not the manufacturer trying to take advantage.

Hope that helps.

[peteski]

Thanks! That was almost like watching the "How things are made" show on TV (which I think is my favorite contemporary TV show). Maybe I can find a video of such machine operating (which would make it even more fun to see than just reading about it).

[narrowminded]

Thanks! That was almost like watching the "How things are made" show on TV (which I think is my favorite contemporary TV show). Maybe I can find a video of such machine operating (which would make it even more fun to see than just reading about it).

Here's an example of one of those machines at work.  What's demonstrated is a larger, more involved part and steel so think of a smaller part, just a turn and taper, and at the speeds acceptable in brass.  Cycle time can be measured in seconds per part, not minutes.

[MichaelWinicki]

Hate me, but I prefer plastic wheels.

Atlas brown wheelsets are my preferred plastic wheelset.

[peteski]

Here's an example of one of those machines at work.  What's demonstrated is a larger, more involved part and steel so think of a smaller part, just a turn and taper, and at the speeds acceptable in brass.  Cycle time can be measured in seconds per part, not minutes.

I really enjoyed that - thanks!

[nkalanaga]

Fascinating!  My maternal grandfather was a machinist, but he never ran anything like that.  Of course, he retired in the late 60s, after a life of working on mostly diesel engines and drive trains, including many years on an ocean-going dredge in the western Pacific.  He liked "modern" stuff, and would have enjoyed learning to run today's machines.

I wouldn't be surprised if that's how some of the brass parts our plant buys are made.  I've been wondering just how they could be turned on a "traditional" lathe, and they were obviously turned, not cast or stamped.

I was still picturing a production machine from his era, probably starting with a long rod, turning the point on one end, then maybe cutting the rod to length with a tool that would turn both the second point on that piece and the first on the next in one pass.

[peteski]

Fascinating!  My maternal grandfather was a machinist, but he never ran anything like that.  Of course, he retired in the late 60s, after a life of working on mostly diesel engines and drive trains, including many years on an ocean-going dredge in the western Pacific.  He liked "modern" stuff, and would have enjoyed learning to run today's machines.

I wouldn't be surprised if that's how some of the brass parts our plant buys are made.  I've been wondering just how they could be turned on a "traditional" lathe, and they were obviously turned, not cast or stamped.

I was still picturing a production machine from his era, probably starting with a long rod, turning the point on one end, then maybe cutting the rod to length with a tool that would turn both the second point on that piece and the first on the next in one pass.

I also seem to recall either hearing or reading that many of the small metal model railroad model parts (like wheels, axles, and buffers) are made on Swiss Screw Machines.
https://www.thomasnet.com/articles/custom-manufacturing-fabricating/cnc-turning-swiss-machines/

[narrowminded]

I also seem to recall either hearing or reading that many of the small metal model railroad model parts (like wheels, axles, and buffers) are made on Swiss Screw Machines.

Very good chance because that's a term that's been around in high volume, small part machines, even when they weren't CNC.  One distinguishing feature was the headstock moving (feeding) rather than a carriage and use of "gang" tooling, meaning the cutting tools are lined up across the machine.  I'm sure they do have some of those tools as they'd be crazy not to with the volumes of some of those parts they make.  I also think some of the improved part quality that has been showing up confirms that there's some better design going on and then the quality machinery necessary to produce those designs.

If you notice, that machine in your video has the sub spindle option as I had described as well as live tooling, meaning tools that have their own powered spindle and, for example, can cross drill or mill from the side of a part with the normal spindle stopped, holding in place or used to index the part around or advance and retract with the live tool running.  With the computer controls you can interpolate any of those motions so in some fashion can machine a feature with straight lines (like a square) in the face of a part, or any complex variations that can be accomplished with those motions in sync.  Those types of combined motion aren't as easy to hold extreme tolerance performing but if the tolerance is a little loose or the machine is that good (expensive)... go for it!  That's also the types of things that have been improving in each generation of machine.  The features keep improving and the costs hold or get better.

The machine that I made my gears in is a swiss style lathe with gang tooling and live tools but does not have the sub spindle.  I wish it did and it's on that shop's wish list, waiting for the right job to come along.   It is the live tooling that made it possible to hob my own worm gears.  It took some specialized in house tooling to complete the setup but it couldn't be done without that basic live tool capability with all spindle motions and speeds controllable and synchronized.  It was very efficient on the wheels too but both needed a second operation, even if just a very quick off machine deburring.  A sub spindle would eliminate that second operation.  And I still need to work up a better rig for part catching with those gears.  They are tiny, light and as a result, easy to lose when parted off.