Fabrication Begins

After creating the computer model and ordering our parts and steel, we were ready to start building this thing. Since we were still waiting on parts, we could only start work on some of the stock we already had in our shop. So, we started by cutting the 2" rod that would be the sleeve as shown in the picture below. The Acme threaded nuts will be welded on to the ends, and there will be a sprocket welded on the outside. This whole assembly will be what pushes the Acme threaded rod and the crush block.

Click to enlarge

The first thing we did was to cut the 2" diameter rod into three 6" sections. While we only need about 3" for the final part, cutting it long will allow us some extra to put into the lathe chuck. This lets us do all of our lathe work for the piece at one time, instead of having to stop and flip the part around to work on the other side. This also further insures that everything will be concentric, where if we take the part out of the chuck to do the other side, there is no guarantee that the part will line up exactly as it was. We did the cuts on our "drop saw", which is a band saw that cuts by falling onto the part. We control the fall rate with a dial. Since steel takes awhile to cut, we will spend quite a bit of time in front of this machine before we finish this project.

Our "drop saw", a band saw connected to a hydraulic piston

After cutting, we put the rod into the lathe. The first thing we needed to do was drill the clearance hole so the Acme threaded rod won't touch the sides of the rod, and only make contact at the Acme nuts. Since we will be using 3/4" threaded rod, we went with a 7/8" clearance hole. It's a little bigger than we need, but better safe than sorry.

One of three lathes in our shop

Before we drill the hole however, we need to "face" the cut surfaces. Machining has a lot to do with accurate measurements and that means precision is very important. The drop saw doesn't leave a flat or a smooth surface, which means we won't be able to measure off of it until we face it on the lathe. The bar with the triangle shaped end is the tool we use to face the part. We move that tool across the face of the part from the outside towards the middle. Since the tool doesn't move left or right, it makes the surface of the part flat and smooth, which allows us to measure off of it later. 

A facing operation on the lathe

After this, we can drill the holes through the center. In order to do this we start with a center drill, and then move up in drill bit sizes 1/8" at a time. By stepping up in sizes, we reduce the chances of the drill bit wandering off center. It also reduces the amount of metal taken away by the larger drill bits, which keeps the heat down and increases tool life. 

Faced and drilled part

It doesn't look like much, but for all three parts it took us almost 3 hours, which highlights the scariest part of this project. We really can't estimate the time it will take to fabricate any of these parts. This is due to lack of experience (for example, a professional machinist would have a much better idea of the time involved), as well as the fact that we have to work around other project teams in the workshop. Either way I never would have guessed how long these parts have taken so far, and there is still a lot to do. All we can do is be in the shop as often as possible and work steadily so we don't get bit later on.

Designing with Solidworks

A big part of our design work on this project involved Solidworks, a 3D modeling program. The use of the program is taught at VTC during the first year of our program, in Design Communication I and II. A webpage documenting the skills taught during Design Comm. II can be found here.

I've posted our physical models already, but we also built our "final" model on the computer as well. Creating a 3D model allows us to do several important things. First, it allows us to create all the individual parts and assemble them. It gives us the ability to quickly export the models of every part to 2D drawings which we can use when we build the parts. And perhaps most important, it allows us to quickly and easily make changes to both the parts and the drawings. 

Solidworks also has a host of simulation features. These consist of simple analysis, such as the approximate weight of a part or assembly, and much more complex operations like motion, fluid, and thermal properties. These are all helpful tools to aid in designing machines.

The ability to quickly create and edit a model really benefits us, especially since we have so little time to finish the project. Another important aspect of a computer model from a professional perspective is how easily the information can be shared between colleagues. In our case, I can make a change on the computer model, and my teammate will see the changes when he opens the model. And if we need to show our professor the results of a stress test or how we anticipate the model to move, we can send him a drawing with the information or even the whole model. 

So as you can see, having the computer model is incredibly important to quickly and successfully completing our project.

Motor Calculations

Now that we have a preliminary computer model and the testing done, we need to calculate the sizes of the motors. Since we've got the final force, we can work backwards to find the torque needed from each motor. I briefly went through the process already, but we'll go over it in more detail here.

First off, we need to generate 5000 lbs of force at the crush piston. That force then gets transferred along the Acme threaded rod to the Acme threaded nuts that turn the shaft. The nuts then transfer the force to torque from the sprocket turning the nuts. A neat thing about Acme threaded rod is that it acts as a very long, shallow incline slope. From a physics standpoint, it is easier to push something up a shallow slope than it is to lift the same thing straight up. So the 5000 lbs at the nuts get multiplied by .073 (for 3/4-6 Acme rod) to convert to the needed torque. Also, while this magnifies our power output, it also drastically reduces the speed of the machine. 

So 5000 x .073 gives us 365 in-lbs of torque needed to turn the nuts. Now, we are using two sprockets, one small one on the motor, and a larger one turning the nuts. These will be connected by roller chain. The larger sprocket will be 6" diameter, and the smaller one a 2" diameter, which provides a gear ratio of 3:1. This gearing allows us to divide the torque by 3, and reduces our speed even further. So now our required torque gets transferred along the chain and is divided by the ratio, which gives us a required torque at the motor of 121 in-lbs of torque.

We repeated this process for the other two motors, but with smaller initial forces. The results are below. Something of note here is that motor usually has high torque or high RPMs (speed) but not both. So by having a motor with less torque, we can have a faster motor. While speed is not our primary concern the extra RPMs will be nice, especially on the longer 8.5" stroke.

Click to enlarge

Model Version 2!

After we finished our testing, two things were clear. Our machine would have to be much more rugged in design and we would have to compromise on the final crush size.

As far as crush size we decided to go with 1.5" x 1.5" x 1". This brought our actual forces down to about 3200 lbs, which after a 1.5 safety factor brings our target to 5000 lbs. We account for the safety factor because we don't want to design a machine that works so closely to it's requirements. Not only is there the increased chance of failure, the machine would then be working at it's max all the time. In short, having the safety factor will lengthen the life of the machine and it's parts.

As for the more robust design, here is a video that covers the new model:

As you can see in the video we have moved from direct rack and pinion operation to using lead screws. This combined with a gear ratio of 3:1 in the sprockets we plan to use allows us to multiply the forces provided by our motors by a very large amount.

For instance, our final crush which we are designing to be 5000 lbs gets multiplied by .073 to give us the torque needed at lead screw. This is because lead screws operate like a very long and shallow incline, and to lift 1 lb, we only need .073 in-lbs of torque. So that gives us 365 in-lbs at the lead screw. Then our gear ratio of 3:1 divides the torque needed at the motor by 3. This gives us a grand total of 122 in-lbs of torque at the largest motor.

There is a down side to multiplying the torque like this. Our speed or RPM of the motor is drastically reduced. So while powerful, we sacrifice speed. Since our project doesn't need to be fast, however, this is an acceptable sacrifice.

For anyone interested, I will link the work in progress E-Drawing of the computer model we created in Solidworks, a 3D modeling and engineering software. To view the model, you will need to download the Solidworks viewer first,

E-drawings viewer: Solidworks E-drawing Viewer

The assembly E-Drawing:  Crusher Assembly  (mediafire download)

Second Round of Testing

After our first tests yielded such large forces we realized we needed more testing to see how different sizes would affect our results. We did more cans of each size, and crushed the cans to our minimum requirement of 1.5" x 1.5" x 1.5". This gave us the minimum forces that we would have to meet to crush the cans.

Forces are an average of three cans

As expected the force required to crush the can is 1700 lbs. Far more manageable than 8000 lbs. However, this is the minimum. So we did some further testing by taking some cans crushed to 1.5" x 1.5" and tested the forces to crush them to three different sizes on the final crush.

Click to enlarge

After these tests, we decided to go with the final crush size of 1.5" x 1.5" x 1". This brings our final required force to 3200 lbs, and brings us under the minimum crush size. And since more power means more money, this compromise will also reduce the cost of the machine.

While we aren't shooting for 8000 lbs anymore, our initial design is still going to need some reworking to provide more torque. So back to the drawing board.

First Testing Results

Before we could move any further we had to do some crush tests to determine the forces our machine will need to produce. We used a Tinius-Olsen materials testing machine here at VTC. This machine is normally used to determine the forces required to destroy a material in shear, stress, or strain. Here are a few pictures of the machine:

The Tinius-Olsen machine

The middle plate can be set at any height, and the bottom plate is lifted using hydraulics

Our first round of tests was for a target 1" cube. We used various forms that we created to crush the cans similar to the way we would on our crusher. For the first crush we jumped right to the 24oz. cans, since we knew that would provide the most resistance, and therefore require more force to crush. The final crush came in at a whopping 8,000 lbs of force! This was far more than either of us expected, and brought into question whether or not we could feasibly crush down to this size. Either way, we knew we would have to test other sizes as well. Here is our first set of results.

We started with the 24oz can but tested all three sizes for more complete test results

First Prototype Design

After some brainstorming, we came up with a solid first prototype design. As with any prototype, our purpose here was to show fit, form, and function. Here's a video explaining how it would work.

It's important to note that since we have such a short timeline, we had to begin design immediately and change things as our research and testing require. This means that this will by no means be a final design, and spending several hours on a model is impractical because it's likely to change. So using cheap materials and quick building methods are a must.

Having such a model does allow us to have something tangible though, which is helpful in explaining our concepts to our professors and our client. It also gives us a way to test interaction of the machine parts, such as the hopper being able to clear the piston arms. Lastly, it gives us a chance to understand how big our machine will have to be to perform the task.

We'll be posting the changes as they come through.