Sunday, September 19, 2010

Concrete Tie Car

My friend Dean Daughenbaugh (who founded New Rail Models with me) is always coming to me with little projects for custom loads. The most recent project is a flatcar with concrete ties. This is actually a pretty cool project. You can see some prototype photos here:

The first link is photo he sent me when he asked me to do this project, and the flatcar is a UP flatcar. The second photo provides more details and is on a BNSF flatcar, showing that these flatcars and loads are used by multiple railroads.

This project will require two molds. The first mold will be for the concrete tie loads, and the second version will be for the metal brackets that help keep the ties in place (I’m assuming).

I started by creating a 3-D model of one layer of the ties:

Tie Strip

The idea is to make these as a single injection-molded piece and then stack them, like this:

Tie Stack

I’ve started to cut the mold for these, and we’ll see how these turn out. Next up will be creating the molds for the metal brackets that will be glued to the flatcar.

O-Scale Gas Truck

We had a number of people ask for an O-scale version of the gas truck kit that I made, so I decided to start working on it. Everything was going just fine until near the end of the milling. At this point, the end mill started to wander, gouging the mold, and then the cutter broke:


You can see the problem in the center of the photo. I was using a long reach 1/32” diameter cutter to sharpen the end of the bands on the tank. The band detail was originally cut with a 1/16” diameter cutter. So I was going back with a finer cutter to make the bands sharper. Unfortunately, the long reach 1/32” cutter was much too flexible, so it got off track (it curved to the right on one pass). And then on the next pass, it curved to the left and broke. Grump. The thin part of this cutter is a little more than .4” long so it can reach down to the bottom of the band at the very right. But that long, thin shaft is not stiff enough.

I could fix this mold by removing about .010” from the top of the mold and then milling the cavity again. However I also realized I should be user a thicker master. This block of aluminum is 1/2” thick, and the cavity is .04” deep, which doesn’t leave much room on the back to resist injection pressure. So I think I’ll order some 1” thick stock and start over.

Friday, September 10, 2010

Flash be Gone!

At this year’s NNGC, Jimmy Booth suggested that I could rescue molds with the “flash” problem I’ve been having by sanding down the face of the mold, so I thought I’d give it a try. He suggested using 500 grit sandpaper and said it would take about 5 minutes.

Well, after 5 minutes with 500 grit I switched to 220 grit because I just wasn’t making much progress. I think I spent about 1/2 hour sanding away (perhaps it wasn’t really that long, but it felt like it). And here is the result:


Success! There is still a little “flash” around the ears on the top, so I’ll probably do a little more sanding. But now the parts come out of the molds essentially clean. Thanks Jimmy!

Backlash on my Taig Mill

I’ve been fighting a small problem for a while and decided to tackle it today (since I’m on vacation until Monday). Small circles weren’t coming out quite round. The reason is what’s known as backlash. When the milling machine changes direction, such as moving to the right and then changing to moving to the left, the knob turns a little bit before the table actually starts to move again. This is usually a result of the lag nut not being completely tight around the screw.

I’m using a program called Mach 3 to drive my milling machine, and it has backlash compensation. But the trick is to figure out how much backlash compensation to enter into the program. A while ago I’d set my Y axis to .003”, but recently discovered that I also had a problem with the X axis. But this time I decided to be systematic about it.

I created a toolpath that cuts small .050” diameter circles with a 1/32” diameter cutter, and then tested different backlash numbers to see what I would get. Here are the results:


None of these are a really nice, clean circle, but after staring at these under a microscope, I decided that a backlash compensation of .002” was the closest to a nice circle. It may be that I can get a better circle after some more tweaking, but I’m going for improvement rather than perfection. This is certainly better than the 0” I had before, which you see in the center of this image.

Thursday, September 9, 2010

Making of a Conoco Gas Truck, Part IX

Some of you are reading this after you’ve already purchased the final kit. I was working on finishing the kit parts up to 9:00 PM the night before my morning flight to St. Louis, so I didn’t have time to take photos and post an update before it was off to National Narrow Gauge Convention. And of course, once there, it was all fun and no time to get on the computer… The kits actually sold better than we expected. In fact, Ragg sold all the kits I made for him at the show, so I’m making more parts.

I also neglected to remind Jimmy Booth to make sure he had plenty of 1934 Ford truck kits on hand, so he sold out on the first day, and it would be another two weeks before he would have new kits. Woops.

I made a new mold for the oil cans, and they turned out rather nice. Here is a photo showing one S-scale oil can with the “flash” removed (more on that next):


And here is a close-up of the parts still on the sprue:


Notice the “flash” around these parts. Once again, the tool path gouged a path for flash into the mold. I noticed this after making one half of the mold, so I was able to eliminate the problem on the second half. But unfortunately, this means I have the “flash” on the parts. Darn!

Fortunately, now I know to look for this issue, which I believe is a bug in the program I’m using, and I now know how to get around it. Even better, Jimmy Booth gave me a suggestion on how to fix my existing molds, which I’m going to try soon.

Next up, I redesigned the mold for the tank lids and handles. The previous version was too hard to build. Here is the improved version:


The handles are much finer than before (because I got rid of the “flash” problem on this mold). And because the holes are larger, it’s much easier to glue the handles in place. I’m pretty happy with how these turned out. By the way, this is blown up way above actual size, so you can see the tool marks.

Unfortunately, between making these new molds and making a set of parts, I ran out of time to make a separate mold for the bottom of the tank. As a result, the kits have two tops. You can easily shave off the tank openings for the bottom half if it bothers you. However, you would only be able to see that they’re there if you turn the finished model upside down.

With the strong sales at the show, and quite a few requests, I’ve decided to make an O-scale version that will fit the Berkshire Valley 1934 Truck kit. This kit is the O-scale version of the PBL kit, which were both part of the Wheel Works line of kits in the yellow boxes.

Monday, June 7, 2010

Making of a Conoco Gas Truck, Part VIII

It’s been a little while since I’ve done any work on this project because work and family kept by pretty busy. Ragg built up the parts I sent him, but he felt that the walls of the troughs were too thick. And they were. I made them .032” thick, but that didn’t look good. Additionally, I had to reduce the size of the oil cans to 80% of S-scale in order to fit, so they looked too small. And to add insult to injury, the reduced-size oil cans didn’t even fit into the troughs. Doh!

So, it was back to the drawing board for some adjustments, and then time to make a new mold. I made the walls thinner (about .015”) and also made the troughs about one scale inch wider so I would have more room for full-scale oil cans. However, I had problems filling the part, as shown with the left-most casting below.


After a little adjusting, I finally got the parts to fill. First I added a third gate and made the gates a little wider (the middle casting above). That go me closer, but not quite there. So next I increased the wall thickness to .019”, and that finally did it, as you can see on the right above.

Here is a comparison of the before and after troughs:


The new trough is sitting next to the assembled kit. The thinner walls look far better, so I’m very happy I made a new mold.

The next step is for me to polish the molds for the tank itself. You may have noticed that the curve on the top of the tank doesn’t look completely smooth. Polishing the mold will take care of this. I also have to make a new test mold with full-scale oil cans to make sure they fit, and then it will be time to start producing kits. The kits should be available at the NNGC in St. Louis this September.

Monday, April 19, 2010

Making a Conoco Gas Truck, Part VII

Well, partial success. I made some mistakes, and I need to make some adjustments. But first, the results. As you can see, I now have lids for the tanks, as well as oil cans.


I’ve included a close-up of the two oil cans. These have some flash, and other issues, but the corrugation is clearly visible. Even with the naked eye you can tell that the corrugation is there, so it was definitely worth the effort.


So what are the problems? First, the extra flash is a result of a small mistake I made creating the tool paths. The milling machine ended up making the first pass two large, which results in extra flash and obscuring the fine detail of the two “ears” on the top of the can. I figured out what I’d done wrong after making the test mold. Here is a side-by-side comparison that shows what I cut on the left, and what I should have cut on the right (these images show the simulation of what the mold will look like after a pass with a .015” diameter bull-nose cutter):

Bad ToolpathGood Toolpath

Second, the oil cans are a little bit too large. They don’t quite fit into the troughs. They go about half way down and then stick.

As to the tank lids, the hole I made for the handles is too small, so the handles don’t fit into the hole. I think I’m going to change this so you can use a standard size drill to open up the hole. Getting the holes just the right size with my milling machine is a little tricky because of backlash, which results in holes that aren’t precisely round.

I’m really glad I made a test mold. In fact, given the problems I ran into, I think I’ll make one more test mold of just the tank lids before making the production mold.

Saturday, April 17, 2010

Making a Conoco Gas Truck, Part VI

I’d forgotten about one part that Ragg wanted me to make for the tank truck add-on kit—namely, 5-gallon oil cans. I had no idea what they looked like back in the 1920s, so I did a quick search on the internet and found a Conoco 5-gallon oil can from the late 1920s (as you can see to the right). Perfect!

Conoco Oil Can 300 Conoco Name 300

I did some more research and learned that these cans are about 10” in diameter and about 23” tall. With that information I was ready to start drawing the can in the computer. I do all my drawing at prototype size and then scale the model down to S-scale when I’m all done. The first step is to draw a 10” diameter circle, and then extrude the circle into a cylinder (left-most picture):

Oil Can 1 Oil Can 2 Oil Can 3

Next I added a slightly larger diameter circle on the bottom and extruded it up (middle picture above). Finally, I added some fillets to make the bottom look more like the can above.

Next I repeated the same operation near the top (left picture below). Then I added a truncated cone (middle picture) and the neck of the can (right picture below):

Oil Can 4 Oil Can 5Oil Can 6

As you can see, building a model in the computer is a series of operations that either add or remove material. Most of the steps so far have been adding material, but the fillet operations knock off the sharp edges, so they’re actually removing some material.

The final steps are pretty simple. I made the top slightly domed, and then added the two “ears” that make it easier to unscrew the lid of the can:

Oil Can 7 Oil Can 8 Oil Can 9

The picture in the middle above shows the can after I added some small depressions like the ones on the prototype. These are fairly subtle, so I’m not sure how well they’ll show up on a model that is only about 1/4” tall. Also, the handle that I’ve added is actually solid, as you can see in the right-most picture. But again, since this is only about .01” thick, I’m not sure that will really be noticeable.

Ragg suggested that I not include the handle at all, but I wanted to try. The production mold will include some number of oil cans (I’m not sure how many just yet), as well as three sets of lids for the gas tank on the truck. I have some questions about how well these parts are going to turn out, so I’ve decided to make a test mold that includes just one of each part. I’ll probably start making that mold tomorrow, and with any luck will have test parts before the end of the day. We’ll see how the corrugations and the handle turn out.

Saturday, April 3, 2010

Making a Conoco Gas Truck, Part V

Since my last post, I’ve been mulling over how best to remove the tank end part from the mold. The “right” way is to add ejector pins. The ejector pins are attached to an ejector plate. And the ejector plate also has two push-back pins that move the ejector plate away from the mold when the mold halves are closed. I’ve never added ejector pins before, and it seemed like a lot of work.

Then an idea struck me last night. Since I’m only making a small number of these, all I really need is the ejector pin and nothing else. I have some 1/4” K&S brass rod that I measured and came out to precisely .250 inches. Plus I have a little bench-top cut-off saw I got from Harbor Freight with a 2” diameter blade. So, first it was back to SolidWorks to add a hole in the middle of the mold, and then back to the milling machine to add this hole.

Holes cut by a milling machine or a drill usually aren’t completely smooth, so I use a reamer from and over/under set. These are reamers that are either .001” over or under the target size. Using the over 1/4” reamer makes the hold nice and smooth, and the rod will slide easily, but still with a tight fit so plastic won’t work it’s way between the rod and the hole.

I carefully put the rod into the mold so it’s flush with the back, and the put the mold into the machine. Once I’ve injected the plastic, I pull the mold out of the machine, put it upside down between two blocks, and then tap with a hammer on another piece of rod:


It worked like a charm! Below you can see the end glued onto the tank truck, along with a casting still attached to the sprue:


Now I only have two more molds to make: the tank bottom, and a mold for the tank covers. Both of these will be straight-forward to make, so I’m getting to the home stretch.

Tuesday, March 30, 2010

Making a Conoco Gas Truck, Part IV

While I was waiting for my replacement cutter to arrive, I did some cleaning up in my workshop. I should do this more often because I found some cutters I didn’t realize I had, including another 1/32” long-reach cutter! After quite a few hours, I finally finished milling both halves of the mold.

Now for the moment of truth—injecting a part. And disaster! Well, maybe not disaster, but certainly not what I was hoping for. I shot plastic into the mold and then I couldn’t get the part out at all. Stuck solid. Grump. In looking at my mold design I discovered I only had a 2 degree draft on the inside walls and no draft angle on the outside walls. That’s way to little draft, especially for such deep walls. So, I’ve adjust the draft angle on the inside walls to be 8 degrees. Fortunately, this is just removing more metal, so I don’t have to start all over. And I can mill out the plastic that’s stuck in the mold while I’m at it. This, of course, will take a few more evenings…

Sunday, March 28, 2010

Making a Conoco Gas Truck, Part III

I started working on the molds for the tank end, which looks like this:

Tank End_300 Tank End Back_300

The back side of the part presents some issues in terms of both making a mold and extracting the part from the mold. Making the mold requires that I cut some fairly deep slots into the aluminum mold halves. These slots need to about about 1/4” deep, and they’re about .047” wide at the bottom.

After I’ve designed a mold, I need to create tool paths that my computer will send to my CNC milling machine. For this I use a program called VisualMill. It takes as input a “solid” model of the mold. I then tell it what size cutter to use, how deep to make each cut, and so on. After a few minutes, it returns a tool path like this:

Tank End Core Toolpath

This particular tool path was for a 1/16” diameter cutter, which I used for the first pass. However, to get all the way to the bottom of the slot, I created another tool path using a long-reach 1/32” diameter cutter. Long reach in this case mean the thin part of the cutter is about .350” long. But that makes the cutter both somewhat fragile, and expensive (about $30 each).

I only had one of these long-reach cutters left, having broken a few others while making mold for the top of the tank. So I put my last long-reach cutter in the mill and started cutting. Unfortunately, I was a little impatient, so I set the cutting speeds too high and broke the cutter. Doohh! I’ve ordered some more cutters, and they should be here on Monday, so I’ll be able to continue working on this mold. But the next time, I’m going to cut slower so I don’t break this cutter.

By speed, I mean how fast the mill moves while it’s cutting. I had it set to move down at 1 inch per minute, and cut at 3 inches per minute. I thought that was slow enough, but apparently not. The other factor here is that I was using a flat-end cutter for removing most of the material, but I think a ball-nose cutter would be a better choice. A flat-end cutter experiences some shock when it first touches and bites into the metal. A ball-nose cutter experiences a less sudden shock because the tip is round.

Sunday, March 21, 2010

Making a Conoco Gas Truck, Part II

I continued working on the core for the tank supports and side troughs. In my previous post, I showed the inserts I was planning on using for the hollow part of the troughs. Here you can see me milling out the pockets for the two inserts.


Once finished, I flipped this mold over and milled recessed holes for screws to hold the inserts in place. Then I used a 4-40 tap to tap the holes in the inserts. Or at least, that was the idea—I broke the tap. Darn. So instead I put the inserts into the pocket and then banged on them with a mallet to firmly seat them, hoping I wouldn’t need the screws to hold the inserts in place.

The next photo shows me starting the process of milling away the unwanted, extra material from the inserts. This will take about two hours, and then I have to make several other passes with smaller cutters.


Quite a few hours later, I finally had the core finished. For this I used both flat and bull-nose cutters, varying from 1/8” down to .015” in diameter. Here is what it looks like when all finished:


Now for the moment of truth. Will I be able to fill this completely with styrene, or will I have to adjust the gates and runners so I can fill the cavity? And the answer is:


There’s a little bit of flash in some places that I think I can probably eliminate or at least reduce with some changes to the gating, but the answer is most certainly yes! So being the person I am who likes instant gratification, I threw together the parts I’ve made so far (without doing a very good job of cleaning up the flash). Here’s what they look like mounted on a PBL truck:


Not bad, eh? I like it. But I just realized I forgot to make cavities for the tank lids. Woops. At this point I need to create the mold for the end of the truck and the tank lids, and then it will be ready to go.

Friday, March 19, 2010

Making a Conoco Gas Truck, Part I

Truck 4_400I’ve been making injection molds and parts for Ragg’s…To Riches? for a little over a year now. I’m pretty new to all of this, being self-taught. So there’s been lots of trial and error. But I’m slowly getting better and trying more difficult molds.

Right now I’m working on a new set of molds for an S-scale gas tank that will fit on the back of a PBL 1934 Ford Pickup. The PBL kit is very close in appearance to the truck shown in the Rio Grande Southern Story, Volume VII on page 123. And I wanted to design a tank that would look much like the one in this photo. It’s very hard to find prototype information about these trucks, so I used some conjecture to create a “close” match.

My first step is to use a program called SolidWorks that allows me to build 3-dimensional models, like the one on the right. I like to use a different color for each part to help me remember how everything will fit together when it’s all done.

Making the Mold

Tank Core

Once I have the model, I then have to design a mold, again in SolidWorks. Molds consist of two parts—a core and a cavity. Generally, but not always, the core is the mold half like the one on the right where there is metal projecting out of the mold. The design at the right is a core that I can use for both the top and bottom half of the gas tank itself.

Making a mold like this is a little tricky, as you’ll see in the sequence below. I could have started with a thick piece of metal and then cut away all the parts I didn’t need. But that’s a lot of work, so I used another approach—I created an insert, as I’ll describe and show below. Inserts were a technique I’d never tried before.

Tanks Support and Trough Molds

I’m going to switch now to using photographs of actual aluminum parts to show you how I’m making the mold for the tank supports (the green parts in the first photo) and the troughs (the red part in the first photo) that are on either side of the tank. First I milled the cavity out of a block of 1/2 inch thick aluminum:


Next I milled a pair of inserts for the core side of the mold. To create the inserts, I needed fairly small blocks of aluminum (these are 0.5 by 0.7 by 2.5 inches) and I had a long 1/2 by 2.5 inch bar. In the past I would have cut this with a hack saw, but that’s not much fun at all. So I broke down and purchased a metal-cutting band saw. Boy, am I glad I did—it makes quick work of cutting the smaller bars, which I then finished off in the milling machine.