3/4" Scale Coupler Update

[Carrdo]

Richard will fill you in on what is going on in his shop with regard to the 3/4" scale couplers he is making; all via CNC machining.

[BClemens]

Wish the photos had more detail...a close-up. What CAM software is used?

[RET]

Hi,

There is no way you can build these things to make money. With all the required steps, fixturing, program etc.; for any reasonable selling price you would be making about 10 cents an hour. Much of the following is cut and paste from a record I'm making for myself.


About twenty years ago I borrowed Don Carr's 3 1/2" gauge coupler patterns and had 14 sets (top & bottom halves) cast in bronze. This summer, I finally started working on them, mainly because Don had started a thread on Chaski about using the same castings to make couplers the traditional manual way. Both methods are a lot of work with many stages and lots of fixturing, but at least by using CNC, all the finished couplers should be the same. The knuckle is steel and there is no casting for it; it must be machined by hand.

For either method, the first step is to take all the castings and using either a file or belt sander, make the mating face on the top and bottom halves as flat as possible. Since the outside of the casting is chilled by the casting process, it is hard and dulls either the file or the abrasive paper fairly quickly. While the castings appear to be quite clean, there may also be a little left over sand on the casting surface which would also contribute to the dulling process. I used both methods, but I had to change belts on the belt sander to finish all the parts.

Tooling fixture for coupler halves. 0,0,0 in the program is the tapped hole at lower right.

I am lucky enough to have one of Al Grigg's completed couplers (Don's patterns are also his), so I measured it very carefully to create a 2D CAD drawing of the two coupler halves using a vernier caliper for the linear dimensions and radius gauges for the radii. The top and bottom pieces were shown on the drawing with the mating faces down, long axes horizontal and 0, 0 at the center of the top knuckle pin (bottom part in the drawing). In the drawing, the bottom piece is created by rotating the top view about a horizontal axis and then filling in the inner parts that are visible in each view. In the drawing (named Coupler in Anvil), the axes of the top and bottom parts are made exactly 1.500" apart.

Next, use the dimensions in the drawing to create the fixture shown in the picture. In the picture it is shown mounted on the CNC mill table. I used 5/16" x 2 1/2" wide aluminum flat bar stock and cut a piece 3 3/4" long. It is important to get the part mounting holes located accurately in the fixture (including the #10-32 table slot holes); I used the digital readout on the Bridgeport for this purpose. The holes in the parts are #41 drill which is just a slide fit for 3/32" stainless rod. Use #3-48 cap screws to hold the parts on the fixture (2 for each part). The screw threads on the capscrews have to be filed down just enough to go through the #41 drilled holes in the parts. There is still enough thread left for holding in the tapped holes.

In the drawing, offset a line 1/16" away from the edge of both the top and bottom parts because a 1/8"dia. end mill is used to do the outline. In Anvil, use the Interrogation function to get the co-ordinates of the intersections of all the line segments that make up this outline, also the radii of all the curved parts. Using these co-ordinates and radii, write the G code for the 1/8"dia. cutter path around the part outlines. Backplot allows you to check the code as it is written.

Locate the knuckle pin hole in the top casting very carefully since everything else depends on this. Once this hole is drilled, carefully line up one of the bottom castings, clamp the two parts together and drill the corresponding hole in it. Make sure both holes are square and true. Take the first pair and using the Digital readout on the Bridgeport and the drawing, locate the second hole on the top part and then drill through this hole to locate the hole in the bottom casting. Stamp #1 on the top and bottom parts and also stamp "M" for "Master" on each part. Use this master set as a drill jig to carefully locate the coupler shank holes on all the other part sets. Number stamp each pair as they are drilled so there is no chance of getting any parts mixed up.

Machining top face of coupler shank.

Mount the fixture on the CNC mill table and use a 1/8" shank stepped pin to make 0,0,0 exactly at the top coupler knuckle pin hole (lower front right of the fixture). See the picture on page #1 to confirm this location. Use a 1/8" dia. 2 flute end mill to do the cutting. If you have one, use a carbide cutter instead of HSS. since the chilled outer skin of the castings dulls ordinary cutters after several parts and the cutter will eventually break after about 3 pieces.

There is a lot more, but this will do for a start.

Richard Trounce.

[RET]

Hi BClemens,

At least 15 years ago now I bought a "turnkey" CNC mill from Sherline in California. All I had to supply was the monitor. I used that mill for a number of years; I even used it to make the expansion links for Dart from an oil hardening tool steel. The links turned out very well.

About 4 years ago now I decided to build a bigger, heavier and better mill but use the same software and stepper motors so the computer wouldn't know the difference. I spent $2,500.00 per axis for 1" dia.Thomson shafts with support rails and machine tool quality ball screws because I wanted an accurate, robust machine with absolutely no backlash. The finished machine is turning out to be even better than I hoped. I also made a rotary headstock for the mill (you can see it in some of the future pictures that show how the knuckle is made).

In summary, I use Anvil 1000 (a good intuitive 2D CAD program) to make the drawings and I use the Sherline software to run the mill and write the "G" code. For complex 3D parts I use Synergy which is a full blown 3D CAD program. The learning curve for that program is almost straight up. Synergy will make the "G" code for complex 3D parts.

Hope this helps a bit.

Richard Trounce.

[BClemens]

Appreciate the photos - great work BTW! It's mind boggling how many different directions and pursuits you can go to 'make progress' in this hobby. I too converted (with servo motors and ball screws with anti-backlash nuts) a brand new Enco mill a number of years ago to CNC (use Artsoft Mack3 operating system software) - the reason was to save time by multi tasking. While the mill was working, I could be working on something else. So far (about 12 years now), that has not been the case because I normally will stand and watch the mill work. (I use Keycreator CAD/CAM software.) You have accomplished quite a lot with your CAM software doing surfaces so well. I would never have thought about attempting a coupler.... but their detail would be super compared to a sand casting.

I nave used the mill to machine the smokebox saddle radius and had anticipated machining the Ø6.125 drivers out of steel for thos locomotive - even wrote the programs but found correct castings for the wheels - a massing time savings! Then I built a foundry to do some iron - that is a whole 'nother story!!

Keep your info coming - most enjoyable - might try a pair of couplers myself - in 2 1/2" scale....

Bill C

[Harold_V]

Then I built a foundry to do some iron - that is a whole 'nother story!!

I'd be interested in reading of your experience. And, please clarify. Building a foundry implies you set up a casting process (a foundry is a factory that produces metal castings). Building a furnace implies you built a use-specific melting device. Did you do both, or are you calling a furnace a foundry? How deeply did you delve in to this process? Casting iron is not easy unless you have the proper melting device. Pouring is reputed to be quite easy, however, as gray iron (or ductile iron) has very good pouring qualities.

Harold

[BClemens]

Then I built a foundry to do some iron - that is a whole 'nother story!!

I'd be interested in reading of your experience. And, please clarify. Building a foundry implies you set up a casting process (a foundry is a factory that produces metal castings). Building a furnace implies you built a use-specific melting device. Did you do both, or are you calling a furnace a foundry? How deeply did you delve in to this process? Casting iron is not easy unless you have the proper melting device. Pouring is reputed to be quite easy, however, as gray iron (or ductile iron) has very good pouring qualities.

Harold

Harold,

Yes I did in fact build a foundry. I fabricated a McEnglevan style furnace with a dual port propane burner and a 12" internal diameter - also made the high pressure blower but found a few dollars later on and bought a commercial high pressure blower from Grainger. If that blower was run with the 3" air gate valve fully open, molten aluminum would swirl in the crucible and be blown up through the lid hole. So, I tried iron - it will melt 40lbs of iron in just over one hour of blasting away propane. (I had to use a 100 gallon propane tank) Cobalt blue glasses were required to look into the white hot furnace interior when the iron was molten.

The other portion of building the foundry was finding a recipe and the components for iron foundry sand (also built a muller for it) with sea coal, wood flour and Albany sand. Made wood flasks (burnt a few) of various sizes. Made most of the tools for mold making and working - sprung and bought a few commercial ones too for sprue and gate cutting and etc.

Iron will definitely make some sweet castings but melting it this way (propane) is not very cost effective. I dropped back to aluminum and bronze. (bronze is probably the trickiest metal to cast) The other portion was the additives and fluxes needed for a good pour. I usually used graphite crucibles and made tools for them (can provide photos) which were somewhat crude but made to fit the crucibles for both lifting out and then pouring. Iron usually required a double shank ring loop to accurately pour - heavy. Some of the fluxes and additives of different kinds were bought fro Power Models - sort-of dates this venture. I bought a LOT from Power Models as well as Coles - when in California.

The propane mixer was a small CO2 (life raft inflate bottle) cut apart and welded into a mixing chamber with a venturi type mixer where the high pressure air was choked to blow into the venturi in a rich (adjustable with a regulator) volume of propane. I used an oil furnace buzz box and electrodes to ignite the flame.

The main portion of a good casting is a proper pattern - I made many! Made patterns and cast a small drill press (precision) of iron - still use it.

I began building a cupola furnace and bought a pallet box of coal coke for it. I have never finished it but that would be the way to melt iron or even steel. That is continuous so you can melt iron for hours - as long as the molds are ready.

Bill C

[RET]

Hi Bill,

Yes, I too often watch the CNC mill run, sometimes because I need to babysit it, but often just for the fun of seeing it do what its supposed to do and see "Wow, this actually works!" For me, CNC adds a whole new dimension to what I can do; things that are very difficult or even impossible to do with conventional machines are easy with CNC. For instance, I would never consider doing the couplers without CNC. Just my take on things.

Now we'll carry on with the couplers.

I didn't do things in the order I'm presenting them. I actually did the coupler knuckle first to see if I could do it. Using the rotary headstock, it turned out fine so that hurdle was OK.


Start running the "G" code program (Coupler outline - top & bottom) which uses climb milling to go around the outside of both the top and bottom coupler halves. After each pass, go down .050" per step and repeat the pass until you touch the fixture base on the last pass.

When all fourteen of the coupler outlines have been finished, the next step is to machine the top face of the coupler shanks to provide a "true" reference bearing surface (as seen in picture #2) when the coupler halves are turned over for further machining in the next fixture. Use the "G" code program "Machine coupler top (top and bottom)" to do this. I used a 3/16" dia. 2 flute carbide end mill for this purpose and it did all of the blanks with no sign of getting dull. Much better than HSS. for this application.

First and second fixtures with the first on the left

Next, build the second fixture from the same bar stock. As you can see, it is quite similar to the first fixture, but all the support and clamping is done on the coupler shanks. That"s why the machined face is required on the top surface of the shank. A short #3-48 screw extends from the bottom of the new fixture to locate the knuckle pin hole horizontally so it can't move while the machining is done to the coupler body (knuckle pin recess, knuckle tail pocket and locking mechanism vertical recess). Again, locate the #3-48 tapped holes and the 3/16" dia. mounting holes accurately with the Bridgeport readout.

Mount the completed fixture accurately on the CNC mill and start writing the "G" code program (knuckle pin recess) to sink the pocket for the coupler knuckle in each half. When the program is complete, clamp the first set of top & bottom castings (stamped #1) in the fixture and use the 1/8" dia. stepped pin to make sure that the co-ordinates 0,0,0 are located accurately over the pivot pin hole (lower right in the picture - where the cutter is). Use a sharp 5/16" dia. 2 flute cutter for this purpose. In the picture, note how securely the parts are clamped to the fixture. In addition to the steel clamping bar visible in the picture, the parts are also held down by #3-48 cap screws through the coupler shanks.

Sinking the knuckle pin recess

The cutting forces are quite high when sinking with the 5/16" dia. end mill. End mills don't sink smoothly, especially when one side of the hole is open. This is the main reason for the firm clamping shown in the fixture.

Cutting the knuckle tail recess

As with all the other CNC operations, the tail recess has its own dedicated CNC program (knuckle tail recess). Make the recess in each coupler half .125" deep. Use a 1/8" dia. x 1/8" shank carbide end mill to cut the recess. Make each pass 50 thousandths deep for a total of 125 thousandths.

Since we are using the Sherline headstock on this machine, it is better to use end mills with 1/8," 3/16" or 1/4" shanks because there are collets for these sizes that fit inside the headstock, thus reducing the cutter overhang. That is not possible with the 5/16" cutter, so a 3/8" shank end mill holder that screws onto the outside of the spindle must be used, thus making the cutter overhang much longer.

Next, use the Anvil drawing "Coupler1" to determine the location of the end holes for the locking bar slot in the top and bottom castings. Don't forget to make sure the drill holes in the bottom part go all the way through.

Drilling the locking bar end holes

Using the coordinates from the drawing, write the "G" code program to drill the 1/8" holes. As in all the other "G" code setups, cycle all 14 coupler pairs through the setup. You get very good at loading and unloading all the parts through each setup.

Finally, write the "G" code to slot the distance between the drilled holes. This is where I'm at right now (De. 28, 2016) but it won't take long to finish this bit. All I have to do is run all the parts through this setup too.

This pretty well finishes the CNC steps that need to be done on the coupler halves. Don't think for a minute that the parts are finished! The locking bars need to be made and fitted, the tail pieces have to be rough shaped and silver soldered onto the knuckles (at least one fixture is required for that) and final machining done on the tail piece which will require another CNC program.



Time for another break.

One small aside. After fighting with "Open Office" or Word to get them to put the pictures where I want them and leave them there, I find this website software to be much easier to use! At least it will do what I want the way I want it. Also, while the picture size restrictions have been removed on the website, it still makes sense to keep the picture file sizes down for the pictures you post to avoid taking too much storage space. Thanks.

Richard Trounce.

[Dick_Morris]

One small aside. After fighting with "Open Office" or Word to get them to put the pictures where I want them and leave them there

In Word I found that clicking on an image or text box after pasting it to a document, selecting "picture tools," then "format", then "wrap text," and "tight" goes a long way towards keeping images from having a mind of their own.

[BClemens]

Richard,

You have patience! What you're doing is sort-of 'mass' production which takes diligence. (I would get bored) Interesting process too. Sorry to get off topic with the foundry stuff - Harold needed clarification.

Yes, the CNC mill has become a necessity in the shop. When the 'CNC computer' crashed (hard drive failure) I basically panicked with the realization that I must go through all that set-up again. But put together another XP machine and remembered the .XMI file! I put a SSD in this one too... It's the most used machine in the shop! (with me standing around watching (baby sitting). This was back when we could still purchase XP from Microsoft (and SSD's were about 10 times more expensive than now). It's still operating.... (not bragging on a computer please!)

Thanks again for the info!

Bill C.

[RET]

Hi,

I finally finished slotting all 14 of the coupler halves. It is quite rewarding when you put the top and bottom halves together and see that both halves of the locking bar slot mate perfectly in ALL of the 14 pairs. I am coming to the conclusion that CNC is more precise than the traditional jig and fixture setups as well as being easier to do.

I find the combination of the intuitive precision of Anvil 1000 together with the ease of use and accuracy of the Sherline software makes for a very powerful combination. Hey this is just my opinion based only on the devil I know, but it seems to be a pretty powerful one. When you add to that a little CNC mill with no play at all, you can do a lot. Yes, I don't use high feed rates or try to push speeds and depths of cut, but I am very pleased with the end result and what I can accomplish with the combination. By the way, the Sherline software runs on Ubuntu Linux, not Windows. Synergy also runs on a Linux platform.

Next, I have to make the locking bar sticks (to cut the individual locking bars from). I also have to start rough cutting the tail blanks to silver solder onto the knuckles and make the parts holding fixture to allow accurate silver soldering. After that, I need to make another fixture to hold the knuckle assembly for finish CNC machining of the knuckle tail.

I also have to add the knuckle manufacture to this record.

So you can see there is still more to do and its going to take a little while yet to do it, but I'm pleased with my progress so far.

Richard Trounce.

[JBodenmann]

Hello My Friends
Here are some couplers you might like. 3/4 inch scale investment cast in 4140 steel. The main body, the knuckle and the latch, with a steel pivot pin and a tiny .020 cotter. When you lift the latch it pops open the knuckle, they work nice! Available in two shank styles, solid and loop shank for a draft gear. If anyone is interested let me know and I'll post them in the sales section.
Jack