Rivet embossing machine

Most railroad locomotives and rolling stock as well are festooned with rivets. When scratch building these can duplicated in a variety of ways. Probably the easiest is just to emboss them onto a piece of thin styrene and overlay it onto the surface that needs rivets. Other options include gluing small diameter slices of styrene rod on the surface, placing small drops of glue on the surface or drilling holes and inserting pins or track nails. All of those are pretty labor intensive and while OK for applications where only a limited number are needed or they are located in areas that would be difficult to do using the embossing method. There are also water slide decals that come with glue dots applied to them that can be applied in strips and some modelers have had success using them but you do have the issue of the decal film which is best used over a glossy surface and it can be difficult to totally hide the clear decal film.

An example of using pin heads for rivets can be seen in the photo below. Because the rivets needed to be placed in a narrow area along a piece of 'Z' shaped styrene it was not possible to emboss them. Due to the number required adding the rivet detail consumed most of the time required to build this car.

When I started my standard gauge 0-6-0 project I knew one of the issues was going to be rivets. The number required on the locomotive was not all that excessive but the tender was another issue all together. The drawing showed row after row of closely spaced rivets. The thought of drilling and installing that many pins put me off big time and caused me to go in search of other possibilities. One of the first things that came to mind was a sewing machine. Here was device that had a part that moved up and down and a mechanism for moving material in incremental steps. The feed amount was variable to provide the required number of stitches per inch. I had an old machine that had originally been my wife's mother's and was passed down to her and then set aside when she got a newer model.

I started experimenting using a cut down and properly blunted needle. Initially results were moderately encouraging. Obviously there would need to be some sort of guide to keep the material going in a straight line and to set the position for a line of rivets. The first issue I ran into was that the styrene tended to slip resulting in unevenly spaced rivets. The advance mechanism was designed to move cloth not slippery plastic. Increasing the pressure on the hold down foot did not help much. I also tried laminating some card board to the back of the styrene or using different types of tape which helped some. Another issue was that the area below the needle was a hole to allow the needle to drop below the surface and pick up the thread on the bobbin. This meant there was not much in the way of back up when embossing and resulted in a less that distinct rivet shape. I ended up spending way too much time with this and came to the conclusion that to make it all work would require some serious modifications to the sewing machine with no guarantee it would work to my satisfaction in the end. A new plan was needed.

It was then I decided to go back to my roots and try something from my past. One of the first companies that I worked for after I mustered out of the Air Force built systems for scanning materials being checked using ultrasound. The accuracy for the positioning wasn't as critical as that used in CNC machines and all that was required was a simple stepper motor drive. The technology back in those days (early 70's) was all discrete components and logic circuits as micro processors were in their early stages and quite expensive. Stepper motors are still being used, generally not in high accuracy machinery but in less demanding applications and are commonly found in hobbyist level 3D printers, laser cutters, hobby craft cutters like the Silhouette and other things. The difference today versus the way we did it back then was now everything is modular and computer controlled. The stepper motor drivers are designed to receive inputs from computers. That was a bit of a stumbling point for what I had in mind. Software is available that can be run on a laptop to provide the necessary step increment to the stepper motor but one needs to spend time learning the program and I didn't want to tie up my laptop. Likewise one no doubt could use something like a Raspberry Pi or Arduino with the proper programming and accomplish the same thing but again there is that learning curve and at my age it was something I didn't want to invest the time in just for one project.

Fortunately some of the logic circuits from back in the day are still available and I decided I would return to my roots and build my own pulse sequencer using discrete components. Note: all of this at this point is only to build an indexer to move the material to be embossed in precise and consistent increments. I also realize this will be of limited interest to most modelers but I'm putting it out here anyway.

This is a precision device. It consists of an aluminum extrusion with a steel rod inserted in each side which protrudes far enough to engage the rollers on the movable portion. The screw which runs from a bearing plate at the left end to the ball bearing stepper motor at the other end is similar to a acme type thread and is referred to as a ball screw. The groove in the screw is sized to ball bearings which are enclosed in a spiral track inside the ball nut which is attached to the underneath side of the movable carriage. This ball nut is preloaded which means that when it is screwed on ball screw the ball bearings are pressed tightly into the grooves providing a very precision fit with very little if any backlash. The screw has five threads per inch, this will figure into the math later on. I've already mentioned the word stepper motor a couple of times and for those not familiar with them you can find a better definition here than I can offer and includes an animated figure to illustrate how it works as well. The stepper motor I'm using makes a 1.8º of rotation each time the coils are are pulsed and therefore it requires 200 pulses to make one rotation. (1.8º x 200 = 360º)

Now for the math part. 200 pulses will rotate the ball screw 1 turn and with 5 threads per inch the screw will move the slide 1/5 of an inch or .200". Therefore each pulse sent to the motor will index the slide .001". Now all that is required is a circuit that will send the correct number of pulses to move the slide the amount you want between rivets in thousands of an inch when triggered. Is any of this making sense ? I hope so, it's not rocket science.

The item on the left is a regulated 24V DC switching power supply. This will supply the power required by the motor and it's driver. It will also be regulated down to provide power to the logic circuits needed to generate the pulses. The device on the right is the modular motor control. It takes the pulses from the logic circuit and drives the coils in the motor to provide the stepping function. The unit is universal and designed to drive a variety of different size and step angle motors, the charts printed on the side show the positions for programming switches on the connector side of the module. The module also accepts commands to determine the direction of rotation, provides inputs for jogging the motor to the initial starting position and provides stop functions that can stop the motor when a limit has been reached.

To date these three pieces have been tested. A square wave generator was bread boarded to provide a constant stream of pulses to the driver module which allowed me to run the slide back and forth and everything so far is working as advertised. The next chore will be to bread board the logic circuit which will read a set of thumb wheel switches. These will be set for the index required in thousands of an inch and supply that number of pulses to the motor driver when triggered to do so.