| Author |
 Topic  |
|
Dwight Ennis
Administrator
1st Class Member
USA
|
 Posted - 09 Apr 2007 : 13:30:35
|
Back on Feb 13, I posted that I'd bought a Sherline CNC milling machine. It was with this project in mind that I bought it. 
The Prototype
First a little background... I've been a fan of the North Pacific Coast Railroad for over 25 years - ever since I got into narrow gauge in fact. This is partly because I grew up in Marin County and the NPC existed in my old stomping grounds, and partly because it was a unique California narrow gauge railroad that, like the SPC, was built "first class all the way." With the exception of the SCLCo, all of my NG layouts, both real and imagined, have been based on the idea of the NPC surviving into the 30's or 40's and never having been broadgauged.
I've also been a big fan of NPC #21, the "Thomas Stetson", ever since I first saw a photo of her.
Known up and down the line as The Freak, this "ugly duckling" was unique in many ways. Built by the NPC's Master Mechanic William J. Thomas in the Sausalito shops in 1901, she was one of (if not the) world's first cab forwards. She was also one of the (if not the) first oil fired locomotives on the west coast (if not in the world), and set the standard on the NPC. All future locomotives on the NPC would be oil fired. She was the first loco with automatic air and with 200lbs boiler pressure.
She used a marine water-tube boiler instead of the conventional fire-tube boiler design, a fact that contributed to her odd appearance since the conventional steam dome was replaced by the "steam collection cylinder" (for lack of a better name).
Rather than a conventional rectangular tender, her tender was essentially a flat car frame upon which rode two vertical cylindrical tanks - the front one carrying fuel oil and the rear one carrying water. The tender also sported a toolbox on the fireman's side and two vertical air tanks on the engineer's side.
Only in existence for a few years (1901 to 1905), there are only a few photos of her known to exist, and these show her evolving over her short life with the headlight being raised to allow for an oval cutout in the cab front (possibly to get more air to the firebox). She sported at least two different stacks, and had a sand dome installed later in life. The whistle also was moved.
The prototype photos...
A few drawings of her have been made, the best of which is probably the HO scale drawing appearing in Narrow Gauge to the Redwoods by Bray Dickinson. 1:48 drawings also appeared in the Gazette, but I've found a number of errors in this set, and have used them only as a rough guide.
Bob Baxter has already built an electric version of this locomotive. During this time, Tom Farin (also an NPC fan) put together some web pages with such prototype information as could be found, along with some speculation and educated guesses on different aspects of this locomotive. Finally, Keith Wiseman produced a white metal kit (sans motor) in HOn3. |
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 13:42:05
|
I've chosen to build #21 in her earliest version, as shown in the first photo of the previous post. Things have progressed, as will be shown, and I've kept it more or less a secret until now. I've never done anything like this in my life, having very little machining experience, and no CNC experience. I've done a little work with brass, usually in the form of making one or two parts for a primarily wood model, but nothing extensive. I haven't done much silver soldering in 25 years other than to assemble one Bangham whistle for my two-cylinder Shay. In fact, the closest I've come prior to starting this project was to build an On30 logging Mallet from Backwoods Miniatures, and that consisted mostly of folding parts along their photo-etched fold lines and soldering them with small iron.
So, I wanted to make sure this was something I could reasonably hope to do before going public. 
I generated my own set of CAD drawings in AutoCAD, based mostly upon the HO drawings in Dickinson's book, but with a few changes of my own made after studying the prototype photos. So far I've had little success generating the g-code using one of the "translation" programs that I've tried (and I've tried several), so I've ended up writing it all by hand, using AutoCAD's "ID" command to get the X-Y coordinates.
A few people have seen this stuff in the form of email updates. What I'm going to do now is to post these email updates in (more or less) their original form, and in chronological order, along with the dates they were sent out. Future updates will be made here directly.
Going in, I want to say a big public thank you to Dave Hottmann and to Bob Starr, both of whom have been an unending source of information, advice, techniques, and encouragement to this metal-working newbie, and without whom this project would never have gotten off the ground!! |
 |
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 13:51:55
|
Update #1 - First Mockup - 2/7/07
I have created full AutoCAD drawings of the cab, running boards, and other parts I'll need to fabricate. I plotted the drawings, cut them out, and made a mock-up from them to make sure things went together correctly. I should have printed it on stiffer paper.
I already have a working live steam 4-4-0 (one of Dave Hottmann's Accucraft Mogul conversions) so the running gear, boiler, etc. is already built. Most of what I'll be doing is cosmetic, though I do have to reroute the steam and exhaust lines. You can see the Mogul/American conversion behind the mock-up.
I wanted to make sure that my perimeter calculations for the curved area of the roof and the resultant length of the piece were correct. I'm starting to formulate ideas on how to actually fab the brass cab with internal bracing and reinforcement.
This is actually starting to take shape!! My CNC milling machine should arrive Friday, and the CAD drawings will be used to generate the machine path.
As you can tell, I'm getting really jazzed!!!
 |
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 14:26:53
|
Update #2 - CNC Mill - 2/21/07
Two weekends ago, I got the work bench assembled and the mill set up.
Last week, I got the computer hooked up to the mill and I started going through the instructions and learning g-code so I can program the thing to make the parts I want. I also watched the mill move under computer control for the first time. COOL!!!
This evening, I created the first g-code file which will cut the front cab walls. The CAM software that came with the mill has, among other things, a g-code editor for creating/editing the g-code, and a window called "Backplot" where you can run the g-code and watch the tool path, feed rate, etc. The software also controls the mill itself.
Here's a couple of shots of the screen showing the cab tool paths generated by the g-code I wrote. The first is looking down on the X-Y axis...
and the second is a 3D view showing cutting tool travel in the Z axis...
A small portion of the actual g-code is visible below the tool path image window. The tool starts from home and moves to start cutting in the lower left corner of the left window, then cuts the center and right hand windows. It then returns home and pauses for a tool change and possible changes to the fixtures holding the brass to the table, then cuts the outline and returns home, at which time the cutting of the part is finished.
I also watched the mill (minus cutting tool and material) go through the motions to cut out this part. In case you can't tell, I'm really JAZZED!!!
My brass sheet came in today, and I've ordered some 1/8 acetyl plastic which will mount to the milling table and back the brass sheet. It will also serve as a clamping fixture. I should have it by the end of the week. All that remains is to buy a 1/16" diameter end mill to cut the windows. The outline will be cut using a 1/4" diameter cutter. Offsets for these tools is already built into the g-code. |
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 14:28:32
|
Update #3 - First Metal has been Cut - 2/25/07
I finally got some metal cut this evening. I cut the roof/side cab walls for #21, the first part I've actually cut out using the new CNC.
Sherline sells 1/2" aluminum tooling plates designed to be bolted to the mill table. These have 10-32 tapped thru holes every 1.6", and I used this first to give me something to screw into. On top of the tooling plate, I added a 12" x 6" sheet of 1/4" acetyl plastic as a backing plate. I did that for two reasons... (1) it machines very easily and wouldn't dull my cutting bit, and (2) my part is slightly wider than the 4" tooling plate and I wanted to support it everywhere a cut was to be made.
I drilled 5 holes in the acetyl plastic that aligned with five holes in the tooling plate, and drilled matching holes in the #22 gauge brass sheet I was going to cut (also 12" x 6"). These holes were positioned so the cutter would clear the screw heads. I bolted both of these to the tooling plate.
I taped a plot of the drawing to the workpiece to align the cutter's "home point" for the cutting process, and to be sure the cutter would clear the screw heads.
I did a few "dry runs" (everything except the actual cutting) with my finger on the "Abort" button to make sure everything would work as planned, and made necessary adjustments.
It was time to run it for real. The cutting process went smoothly and tool a little while as I was using a 1/8" diameter end mill and had the feed set slow.
Here's the finished cut...
and here's the completed part...
I absolutely LOVE this tool!!!  |
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 14:31:07
|
Update #4 - Cab Front Walls - 2/27/07
I got the cab front walls cut out last night. There was a little cleanup to do around the insides of the windows with a file. BTW, I've modified the window profiles since the mockup to more closely match the prototype.
One of my concerns has been trying to fold the angled front walls after cutting them out. The cross section here is pretty small... 0.090" to be exact, and the 14° bend goes right in the middle. I bought a Backwoods Miniatures logging mallet kit a few years back (another project half-finished) and I remembered they had etched fold lines included in their flat parts. So I got the bright idea of using a small Dremel ball cutter in the mill chuck to cut a similar fold notches into my front wall.
Here's the part (inside view)...
 |
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 14:32:40
|
Update #5 - And the Cab Lower Sides - 2/25/07
 |
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 14:41:48
|
Update #6 - [b]Cab Sub-Floor[b] - 3/2/07
Tonight I cut the cab sub-floor. This will (a) act as a bottom former and place to solder the front and lower side walls to, and (b) provide a place to attach the front hinge so the cab can open forward like a semi.
Mark Scrivener had warned me that this 260 brass can me miserable to machine. So far I've had no problems because I've been cutting 0.025" stock. This piece however was cut from 0.064" stock, and the problems became apparent immediately. The above piece was the second piece I cut.
Here's the first piece...
I guess I tried taking too much material off in a single cut. On the first piece, I was taking of 0.032, or half the thickness. The cut came out terrible, with what on wood I would call "splintering of the edges" - is this called "galling" in machining? Also, I didn't have the part adequately clamped and it broke loose, allowing the end mill to chew into the upper inside right-hand corner. Here's another shot of the edges of both parts...
The part on the left is the second part, with the first one on the right. Quite a difference.
On the second part, I reduced the depth of the cut to 0.016" and increased the total number of passes to four.
At least I learned something.  |
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 14:43:35
|
Update #6 - 3/6/07
Crap!!! I went to solder the cab front to the floor support (the last part I made) and it didn't fit!  Further investigation revealed that I used the wrong tool offset when I wrote the g-code for the floor support, so I'm going to have to make another one. I wrote the g-code for a 1/4" end mill, but ended up using a 1/8" mill to reduce galling. Unfortunately, I forgot to correct the g-code for the change. My first "mistake" on this project.
At least it's a simple piece with no rivet detail, etc. Far better to have to remake it than the cab front or roof.
I picked up my Little Torch oxy-acetylene set today... torch, regulators, tanks, caddy... the works. I made another test tonight silver soldering a couple of scraps - a piece of 0.025 to a piece of 0.062 - the same thickness I'll be silver soldering on the cab itself. It turned out far better and was a lot easier with the new torch than the previous test with the pistol grip butane torch. Here's the unit I got...
http://www.littletorch.com/caddy.html
Bottom of the page # 23-1004A
BTW, I've gotten all the rivets embossed on the front cab wall. Still have to do the roof. There are a couple of other issues I also need to work out with the lower side panel design so the cab can properly hinge forward. I may end up making these lower panels in two pieces instead of one so the curved handrail at cab rear can stay put when the cab swings up. I'll know more after I get support piece remade and the the front wall soldered to it. |
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 14:48:12
|
Reply to email response to Update #6 - 3/7/07
> Looks almost like a miniature version of my full size oxy acetylene torch.
Yeah, that's pretty much it exactly.
> How hot do the propane ones get versus the new acetylene version?
From their home page...
Produces a perfectly controlled, stable flame hot enough to melt commercially weldable metals, glass and ceramics with thread-thin flames up to 6000° F.
Smith Equipment's Little Torch welds an unlimited range of materials, brazes and solders the tiniest elements with ease, yet produces enough heat to weld steel up to 1/8" thick or melt 3 oz. of silver.
Not sure about the propane version but I'm sure it's substantially less. Also, the smallest two of the tips aren't recommended to be used with anything other than acetylene or hydrogen.
> And what melting point silver solder are you using?
Stay-Brite 4% 430 degrees. I also have some 0.032 stuff from OSH that melts at 1250 degrees F,
> Your first mistake :) - I have a feeling that I would make lots of those kinds of errors if I was doing it!
Yeah, I suppose it had to happen. [>:(] I'm just glad it was a simple piece as opposed to one of the others. I can rewrite the g-code for this piece in 5 minutes and have another cut in 1/2 hour (including setup).
BTW, here's a couple of pics of the front wall riveting...
 |
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 15:00:56
|
Update #7 - 3/12/07
Things have been moving right along and the cab is roughly 80% complete. I spent the day today assembling the parts I've made.
The new little torch is fabulous and as I gain experience, things are getting easier with every joint (please forgive the grungy finish... paint will make a world of difference). This isn't to say I haven't made mistakes. 
There are also a few imperfections, as with the left window stanchion here. I plan to bend these all back into position once I'm done. I stupidly tried to hold the roof in position by clamping from the bottom of the cab to the roof, and those narrow little stanchions bent.  They have been the biggest pain in the ass of the whole project. As I feared, they didn't bend in the brake and stubbornly remained straight.  I finally ended up filing the profile into them.
Here's a shot of the inside. The middle and rear braces with the half-loop to support the top cylindrical whatever-it-is each took about 50 minutes to mill out. Being made from 0.06" material, I had to do each in four passes to prevent galling. Listening to the shop vac run for 50 minutes straight is no fun at all. 
I made a discovery though... it's better to have the mill moving against it's normal cutting direction on the part side of the cut. It produces a much nicer finish on that side of the cut. Unfortunately, I already had the g-code written to do just the opposite by cutting from the other direction, so I had to do some filing. Next time.
Lastly, here's a shot with one of the side panels taped in place for effect.
I'm getting there.
|
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 15:12:57
|
Update #8 - 3/29/07
It's been a little over two weeks since my last update, partly because (a) other things (like a St. Paddy's Day party on the resultant hangover) took one of my weekends, and partly because I was waiting for a tool. I ordered this Grizzly combination 12" shear/brake/slip rollers...
and getting it delivered was a minor nightmare. The first one was four days late since Fed-Ex. Freight didn't realize it was being delivered to a business address (I have things shipped to work). When it finally arrived, the shipping carton (a simple cardboard box for a 120 lb. machine) was damaged and so was the tool (it wasn't secured within the box or padded in any way), so I refused delivery and sent it back. A phone call to Grizzly resulted in them telling me they "had to verify with the carrier" that I'd refused delivery before they would ship a replacement, and that took another 48 hours. Then Fed-Ex Freight forgot to complete the paperwork, so that delayed it another day. However, I finally got it last Tuesday and it was in good shape.
Okay, on to the cab. I had to remake the lower side panels. First off, trying to get the rivet pattern accurately placed was difficult at first. Rivets can't be embossed on the milling machine because the punch and die need to remain stationary while the work moves, and with the work secured to the milling table, there's no place to affix the die that won't move with the work. As a result, I used my Northwest Shortline Riveter. However, the fence I clamped to the table slipped and I screwed up one lower panel.
Another reason for remaking them is that the more I looked at the photos, the more apparent it became that I had the curved section behind the cab wrong. I'd followed other drawings (like the ones in the Gazette) and whoever drew it used a simple quarter-circle. But if you look closely at the prototype, you'll see that the panel is about 1/3 taller the curve is wide, and there's a vertical section at the bottom of the curve.
So, I drew up new plans in AutoCAD and wrote new g-code to cut the parts - in four pieces this time instead of two, because I want the curved section to remain attached to the running boards when the cab itself hinges forward.
The NWSL rivet set includes five sets of a punch and die each, plus an additional large punch with no corresponding die. I solved my rivet problem by chucking this extra punch in the mill's drill chuck, unplugging the Z axis stepper from the controller, and writing g-code that would go to each rivet location and then pause until I clicked the "Resume" button on the software. At each location, I'd manually lower the spinning punch into the piece to create a "center punch dimple" for the riveting punch later. After all the rivet locations were dimpled, I swapped the punch for a cutter and cut out the piece. Later, when actually embossing the rivets, it was pretty easy to manually locate the dimple by moving the piece around until the punch dropped into the dimple and then press in the rivet.
Here's one of the resulting curved walls...
Next I made the front piece for the steam collection cylinder shroud, and then the shroud cover itself.
This last was somewhat of a bear, and I had to make it four times before I got one I was happy with. Length is critical since it had to form somewhat a half-cylinder and there are flanges that stick out each side which attach to the cab roof. Getting a sharp bend at these flanges was also difficult. After two tries, I called Bob Starr, and he advised me to file a groove at the bend line. Never being one to use muscle when there's an electric tool that will do the same job (hehehe), I went to OSH and bought an engraving cutter bit from Dremel...
Again, I chucked it up in the mill, wrote some g-code that would go to the beginning of each fold line, pause while I lowered the bit into the work, and then cut the line. It's actually just like what I did on with fold lines on the cab front (I don't know why I didn't think of it before Bob suggested it). I also used it to center-dimple the rivet locations on this piece and it worked well.
So without further adieu, here's the cab so far (again the curved sides are only taped in place for the photos)...
All that remains is to make the window inner frames and solder on the 0.050 and 0.025 half-round beading around the windows, and the cab will be essentially finished.
|
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 15:19:00
|
Update #10 - 4/2/07
I essentially finished the cab yesterday. Now it's on to the cab deck/running boards, and then I'll be rolling the boiler jacket and actually starting the disassembly of the Mogul/American conversion.
I have done some tedious and difficult things in my modeling career, but soldering the 0.025" half-round beading around the windows is right up there in the top five! What a total PITA!!! It took me most of the day yesterday to do the five windows. If I'd known going in what was involved, I'd have been sorely tempted to leave it off.
|
|
|
|
Dwight Ennis
Administrator
1st Class Member
USA
|
Posted - 09 Apr 2007 : 15:28:18
|
That does it for the previous updates. Last weekend I took the Mogul/American over to Mark Scrivener's house and ran her for the first time. As I expected, having come from Dave Hottmann, she ran like a jewel. When I got her home, I completely stripped her down to the frame and running gear.
I made the main cab deck/running boards cut out at least in its basic outline, and I made a test boiler wrapper to get the piece length correct so there's no gap along the bottom once rolled. No photos of any of that yet, but I'll get some taken eventually. For the deck, I ordered a piece of 0.064" #353 Engraver's Brass. It machines like a dream compared to the 0.064" #260 stuff I used for the internal cab supports. I cut it out in one pass with no galling at all.
The boiler intersects the cab deck/running boards at a three-degree angle, so there is a conic section cutout in the deck where it meets the boiler. Before I can lay this out, I need to get the basic boiler wrapper made so I can determine the exact O.D.
That brings everything completely up to date. If you've gotten this far, thanks for your interest. |
|
|
|
Charles
Engineer
1st Class Member
USA
|
Posted - 09 Apr 2007 : 15:58:59
|
Dwight
This project will do into the file of mastery of live steam excellence and uniqueness.
Will be following the post, outstanding CNC work combined with craftsmanship putting it together.
Congratulations thus far.
|
|
|
|
|
Bruce Chandler
Engineer
1st Class Member
USA
|
Posted - 09 Apr 2007 : 16:10:20
|
Wow! That is one heck of a first project!
Very impressive and very nice looking. |
Bruce
|
|
|
|
Topic |
|
|
|
All Forums
Forums
NPC #21 Live Steam Builder's Log - Update 12/31/07
Forum Locked
Printer Friendly