Linear Motion Drivers

Linear Motion Drivers

In this post we will be going over the different methods to drive the axes of a CNC machine.

Threaded Rods

1/4 inch threaded rod with nut

Close up of 1/4 inch threaded rod

Threaded rods are a good choice for lead screws if you require minimal precision and are planning to cut softer material like foam or wood. They are the cheapest common option available and can be found at a local hardware store like Home Depot or Lowes. The pricing for this item is approximately $20 CAD (high estimate) for a 4ft, 1/4 inch diameter rod (if my memory serves me correctly). Threaded rods are the cheapest option mentioned in this post.

If you are aiming for precision or desire to cut harder materials like aluminium and possibly steel, you do not want to cheap out on the lead screws and go with threaded rods. The main reason is the tolerances for the thread and nut pitch is low. This will cause an axis to move at uneven speeds, due to friction between the threads, which will have a negative affect the finish (smoothness) of the surface. Also the motors will, in general, need to apply more torque since large amounts of friction will be created by the mismatch in thread pitch. If the motor drivers do not have a current protection circuit, you could very well blow your motors if the friction between threads is greater than the peak torque generated by your motor.

Backlash is one other issue that arises with mismatched pitches. Backlash occurs when the pitch of the screw is larger than the pitch of the nut (when it is reversed, the nut will be unable to move along the screw). However, in a threaded rod, the pitch is not consistent across the entire rod so the amount of backlash will change from position to position which makes it impossible for any CNC controller on the market to compensate. It would be near impossible to program your own controller since you would need to know the backlash at every point along the screw. Backlash will cause the dimensions of the part being made to be thrown off. Granted, if you don't care too much about accuracy, then this is of little concern. The gap between the two threads and the nut thread can also cause chatter in the machine because the nut threads are free to bounce from one screw thread to the other without resistance. While the spindle is spinning, it could cause the nut to oscillate back and forth within the gap between the screw threads and cause this phenomenon known as chatter.

If threaded rods are still the preference, we advise that you test a nut with a rod before you purchase them to see that the movement is fairly smooth. To connect the rod with your motor shaft, you can use shaft couplers off ebay or attempt to make your own by buying a longer nut and making set a pair of set screw holes on each end. If the motor shaft does not fit in the other end of the nut, simply drill half way into the hole using a drill bit the size of the motor shaft. This home made shaft coupler is cheaper than buying a shaft coupler off ebay most likely.

Here is a CNC machine that was made with threaded rods.

Summary
pros:
    - cheap!!!
cons:
    - largest amount of backlash out of all the lead screws I will mention
    - rotation of nut along thread is rough and uneven due to a lack of precision
    - low manufacturing tolerances

Timing Belts

Timing belt with pulley

Timing belts are a fairly good options for those builders on a budget. A fair amount of diy CNC machines and lower end machines, like ShapeOko, use timing belts since they are relatively inexpensive. I bought a 720mm belt for less than 15CAD when making a moving target. I inherited the pulleys from a friend but doing some research on Ebay, they are about 10CAD each. This allows for approximately 350mm of travel on a single axis for 35CAD. Like the pulleys, timing belts can be found on Ebay and are also sometimes sold in automotive part dealers.

Timing belts are suitable for machines that require quick and smooth movements and cut softer materials like wood. However, there can be backlash issues as well since a belt, made of rubber, is able to stretch to some degree. Generally a fresh belt will hold its length fairly well; after all, they are made for automobiles and other machinery. Tensioning the belts appropriately can be difficult in order to squeeze as much precision out of the belt without having it stretch too much or be too tight to so that the torque required to turn one of the pulleys is too much for the motor. Belts will need to be replaced overtime, more so than screws, and the harder they are run, the faster they will wear out and stretch. In conclusion, I would recommend belts if you are making a smaller machine with the intention to cut wood and other softer materials.

Summary

pros:
    - relatively inexpensive
    - able to drive an axis at higher speeds

cons:
    - belts will stretch over time
    - difficult to tension properly
    - backlash issues increase as load on the router increases due to stretching
    - belts wear and tear faster than screws

Acme Threaded Rods

Acme Threaded Rod

Acme Threaded Rod Nut

Acme threaded rods are an excellent choice for those who want precision but are still constrained by a budget. These types of rods use a trapezoidal shape for the threads which allow transmission of higher torque loads and greater precision in comparison to the options mentioned above. These components are often used in lathes because they have a self locking property.

The threads, being trapezoidal, have a large amount of contact surface area, which is prone to high friction under certain forces. When the nut, threaded onto the screw, has minimal load (i.e. just itself), the force of friction is minimal. The small force of friction only occurs because the nut is moving axially. In order for the nut to move, the threads of the rod must push against the threads of the nut creating friction. Therefore, we can conclude that axial force give acme threaded rods their self locking property.

However, not all threads have a self locking property. The main factors that contribute to the self locking property are the materials of the thread on both rod and nut, and the angle of each thread, which is associated with pitch. The reason material and angle are the main factors is due to the friction equation:

 Friction Force = (Coefficient of Friction) * (Normal Force). 

The material(s) of the threads will influence the coefficient of friction, which is determined through lookup tables (these tables will show pairs of materials and their respective coefficients, static and kinetic), while the angle of the thread will affect how much of an axial force gets translated into a normal force.

A rough diagram below was drawn for a bit of a visual. In each sub-diagram, the rod is on the left hand side while the nut is on the right. The threads are assumed to be right handed. When a force is applied on the nut in the y direction, it is transferred to the rod's threads and broken up into two components, Fthread and Fnormal. In order to accurately discuss the directions of the force, you will have to imagine the force vectors in 3D as we describe them.
First of all, with a right handed thread, the nut will spin the direction noted by the rotation arrow in order to move in the positive y direction. The interlocking threads prevent the nut from just sliding along the thread and as a result, the force is split up at the contacting threads. Fthread is the component of Force that is parallel to the thread's surface and in these pictures, the z component of the Fthread vector is pointing into the screen. Fthread encourages the thread of the nut to slide along the thread of the rod, therefore causing a rotation and positive y displacement. Fnormal is the force that is perpendicular to the thread's surface and has a z component that is sticking out of the screen. The Fnormal force causes friction which acts in the opposite direction of Fthread, therefore opposing thread sliding. It is to be noted that the force of friction can never cause the thread to slide in the opposite direction of Fthread.

We hope that you understand the forces and will compare the two scenarios. It can be seen in the diagram below that a constant force and decreasing pitch (and thread angle) will cause Fthread to increase and Fnormal to increase. In laymen terms, the force that causes rotation decreases AND the force of friction increases.

Now looking at these fundamentals at a point on the thread, we can say that at a 45 degree thread angle and a coefficient of friction of 1, the Fthread force and Fnormal force balance and therefore no motion will occur. However, most amce threaded rods will have a thread angle considerably smaller than 45 degrees so that the threads aren't spread out as much and giving more rigidity to the system.

Another way to picture this is think of a really long swirly slide that represents the threaded rod and a long spring shaped object resting on the slide like the threads of a nut does with the rod. When the spiral shaped object sits on the slide, the force of gravity will act on it. The angle of the slide (or steepness) and the materials of the slide and object.

Generally, acme threaded rods use some sort of plastic-like nut. One of the most common materials is delrin and can be purchased on ebay once again. Search "delrin rod" and you will find a variety of diameters and lengths. We took a look on ebay today (March 11/2014) and found a 40mm diameter X 4.25 inch length rod for $4.50 USD plus $10 shipping. We suggest you message the seller to see if you can get a discount if you buy more than one. Some sellers advertise free shipping if you order multiple products from them. Once you have your delrin, follow this link to find out how to make the nut. In case the thread goes missing, here is a summary of the instructions:

1. Make a hole through the centre of the rod with the diameter of your acme threaded rod minus the depth of one thread.

2. Cut the rod down the centre so that the hole is cut in half.

3. Prepare the lead acme threaded rod by polishing the surface and in between the threads. Wipe on a thin film of mineral oil.

4. Place the two halves of delrin around the acme threaded rod and square up the delrin so that when the delrin is heated, the two halves will join back together and create a cylinder.

5. Place the two halve delrin rods (which wrap around the acme threaded rod) in a vise with the gaps facing upwards and downwards. The vise will push the two halves of delrin to deform around the nut and eventually meet up and melt together.

6. Take a heat gut and aim at a part of the acme threaded rod that is not covered by the nut. Do not aim it at the delrin itself because it could cause the surface that is exposed to heat to melt at a higher rate. Allow the heat to travel through the acme threaded rod and melt the delrin from the inside so that it is moulded to fit in between the threads. Slowly tighten the vise in order to encourage the two halves to come together and eventually melt together. This process is said to take 10-15 minutes.

7. Chuck the nut in a lathe (put the nut in the part that grips a round object) and use a hand vise to grip. Turn the screw with the hand vise and do this slowly since there will be a lot of heat generated from the friction between the nut and screw. Use some oil as a lubricate.

The backlash on acme threaded rods is minimal in comparison to the options mentioned above and is further reduce by making this custom nut since it was moulded to thread of the rod. There is still space between the nut's thread and rod's thread which will create some backlash; if there wasn't any space, the nut would not be able to move along the thread. A caution should be mentioned that as the nut is pushed along the screw more and more under a load, the nut could wear and backlash will increase. We have not looked too much into ways to further minimize backlash but we suspect it would involve two separate nuts that are locked in rotation with a heavy spring in between.

One more thing to note about acme threaded rods is that they can have multiple starts (see the wikipedia page under the subheading "Lead, pitch, and start"). The number of starts refers to the number of different threads that spiral along the rod. In the picture below, it shows a threaded rod with 5 starts. As the number of starts increases, with pitch and thread width remaining constant, the lead increases (the space between a single thread's threads) since with extra threads, each thread is forced to be stretched out axially. The advantage of multiple starts is to increase speed but at the same time, precision is sacrificed. Speed can be increased because each thread alone is spaced out more which causes the nut to move further in one rotation. Multiple number of starts helps to stabilize the nut as it moves along the screw. Also note that as the number of starts increase, so does the cost.

There was a post on CNCzone about cheap acme threaded rods from an online store by the name of Enco. We have personally not tested them but it is worth taking a look at if you would like

Summary

pros:
    - very robust
    - minimal backlash

cons:
    - most expensive (in general)
    - lots of friction

One last note, and it should be common sense, is that for CNC machines, lubrication becomes crucial, especially for single start rods, so that your motors will be able to drive the screw. If you have unlimited torque in your motor, then you can ignore this note :)

Ball Screws
Ball screws improve upon the basic concept behind threaded rods. Instead of direct bearing between the screw and nut surfaces, ball screws contain ball bearings which fit in between the two surfaces. This makes the transfer of motion far more efficient (aka less friction), since there is now rolling friction as oppose to two surfaces sliding across one another (which is easier: rolling a cart or sliding a box across a floor?). This is significant when you consider that acme threaded rods are generally about 30% efficient, while ball screws can be up to 95% efficient.

With less friction in between the nut and the screw, the ball screw, in contrast with the acme threaded rod, requires more force to stop the nut and any attached load when they are in motion since the system has less friction. As a result, the motor size will increase if the desired maximum speed increases. Also, if a ball screw is used in a vertical application, you MUST keep in mind that once power is removed from the motor, a load on a vertical ball screw assembly (like gravity acting on a spindle) will cause the ball nut to spiral down. This is considered back drive. One way to avoid this is to increase the threads per inch on your ball screw (maybe 10 tpi) or add in a locking mechanism for when you shut the machine off (this is your safest bet).

The main drawback of ball screws is their price. However, the added quality could be worth the investment depending on the application. Since the nut and screw aren't grinding against each other, the life span of ball screws is usually longer and therefore they require less replacement.

Summary

pros:
    - very good precision (very little backlash)
    - longer life time
    - minimal friction

cons:
    - cost
    - require more braking force due to less friction
    - require a locking system in vertical applications
    - require a lubricant

CONCLUSION

The recommendations we have for each linear motion driver will be listed below:

Threaded rod: Only use if you are on a budget!!!!!

Timing Belt and Pulley: Use if you are intending to limit the machine to wood.

Acme Threaded Rod: Can be used for any machine. Select higher grades if you want to cut harder materials with higher precision. Be prepared to use bigger motors in order to overcome the friction forces

Ball Screws: Preferable if budget is not an issue. Ball screws will give you the most efficiency (reduce the energy spent to move the CNC machine around).

These are general remarks on each type of linear motion driver. Another large factor that will come into play is the type of material each one is made of and how much force the component can withstand.

Basic Design - Frame

Common Types of CNCs

DIY CNC mills usually come in two styles: moving gantry, stationary base or moving base plate and stationary gantry.

We decided to go with a moving gantry and stationary base due to the overall size of the machine we wanted to build. The minimum cutting area we hope to achieve is at least a 2ft X 2ft X 4in with overall dimensions of 2.5ft X 2.5ft X 1.5ft. The reasons for a moving gantry style is that it shields the linear components in the base from debris, has a smaller overall size to cutting area ratio (very big factor), and looks cooler (in our opinion)! However this design is trickier to build and tolerances of the parts need to be held tight in order to get a rigid design (how well the machine can hold together when forces are applied on different sides which occur when cutting). 

Generally there is a saying throughout the hobby CNC community that whatever you make your CNC out of is the hardest material you will be able to accurately cut. For this reason, we chose aluminium since this will probably be the hardest material we will ever need to machine for home projects or the fourth year design project at the University of Waterloo. To increase the precision of our machine, we plan to use the lathe and vertical mill at the university machine shop in order to maintain tolerances for each part.

Throughout our research, we found that the moving base is more popular among smaller machines since it provides a more rigid setup. The gantry, being stationary, can be build to withstand torsional forces around the z axis if the router is off centre from the y axis and cutting in the x direction. The way to achieve this is to widen the supports for the gantry in the x direction to prevent the gantry overhanging beam from twisting. The biggest downside to this design is that your x axis length will equal your base plate length plus the amount of travel. For example, if you want a 3ft X 3ft baseplate with 3ft of travel, the dimensions of the stationary gantry machine will almost be double that of a moving gantry.

After watching many different diy CNC machines with a moving gantry cut aluminium and even steel , this helped affirm our decision on this style of CNC. At first we were worried about how rigid our machine would be. However, seeing many other peoples success with a moving gantry and milling aluminium with a fair amount of accuracy, we set out to make design plans with this architecture. One reason we are not too worried about the twisting around the Z-axis is because we will not be pushing our machine at ridiculous speeds (1000 inches per minute [IPM]) that will cause large forces opposing the router.

Here is the website that we used when learning about the two different designs. As you design your own on on autodesk Inventor or some other CAD software, you will be able to see the advantages and disadvantages more clearly.

Frame Material

Now a lot of you are probably wondering how this machine we are making is for a slim wallet if it is built out of aluminium. If you have never made an inquiry about aluminium, a 2ft X 2ft X 1/2in piece of aluminium is at least $120 CAD which will be enough to make the gantry side plates with little to spare for the rest of the machine. Well we got lucky and obtained a lot of aluminium t slot extrusions for free from a company that was clearing out old materials. The aluminium t slot extrusions are most likely made by 80/20 Inc since they have the same dimensions as their "15 Series" extrusions. Also we obtained some aluminium plates (10mm thick) from a friend at the University of Waterloo who occasionally receives donations of used components and materials. All these salvaged pieces of aluminium will allow us to make most of the framework. If you are looking for t slot extrusions, we recommend going onto ebay and searching "aluminium extrusions" and buying lots of up to 10 pieces since it is cheaper than buying from 80/20. You can get all the extrusions you would need for a machine to match ours for around $300 or a little less.

Another option for aluminium extrusions is ordering from Misumi. They have fairly cheap extrusions ($13.60 for a 610mm length slot with a profile dimension of 30mm X 60mm) but shipping could run your costs significantly higher. Our guess for shipping cost would be about $10 USD at least per extrusion. You can check the approximate cost of shipping by entering your orders info into the UPS shipping calculator with the weight Misumi gives you for your order when you request a quote and approximate dimensions.

Aluminium is quite expensive and if we hadn't gotten materials for free, we would have settled with a machine made of MDF. A 2ft X 2ft X 3/4 inch piece at Home Depot sells for around $22 CAD which is doable. Extend one of the 2ft dimensions to 4ft (2ft X 4ft X 3/4) and its about $35. Buying two pieces of 2ft X 4ft X 3/4in should be enough to complete most of the frame of a machine. Another type of aluminium framing that can be considered is square tubes, a 2in X 2in square (1/8in thickness) and 6ft length can be bought for around $55 on McMaster-Carr. We have yet to look up the costs on Misumi but they might have it slightly cheaper. Currently they have a promotional $150 off your FIRST order. This ends March 31, 2014.

If you are not in a hurry to get your materials, look around and try to go to some local scrap yards or search on sites like kijiji for people selling metal or giving away free wood. Emailing manufacturing companies near by could help out a bit on the budget. One recycling company we stumbled upon is foxy recycle located in Ottawa but they do shipping as well. These recycling businesses might have some aluminium plates on hand that have a few holes but still reusable for making some parts. Also if you go to home depot, they sometimes have some wood with defects like a piece of MDF with a crack down the center. They sell these pieces at a steep discounted price.

If you plan to make your machine without access to a machine shop and need a way to make straight cuts, we have CADed a jig in this video (click here). The materials you will need are some MDF, linear rails (which you will need for your CNC machine anyways), a hand drill (for making holes unless you want to use your router), and hex sockets/screws or whatever fasteners you can find.

Estimates of Different Frame material costs:

Aluminium - be prepared to spend at least $400 for an entire 2.5ft X 2.5ft X 1.5ft frame if you order plates that are 1/2 inch thick

MDF - A 2.5ft X 2.5ft X 1.5ft frame will cost ~$80 if you purchase the 2ft X 4ft X 3/4in boards and some other small blocks

Summary of my basic design:

Material:              Aluminium 

Style:                   Moving gantry, stationary base

Cutting area:        x-axis: 2ft (approx 609mm)

                           y-axis: 2ft (approx 609mm)

                           z-axis: 4in (approx 102mm)

Budget:               $1000

Introduction to Slim CNC milS

Thank you for stumbling upon, searching, or bookmarking this blog. The purpose of this blog is to documents the design and build process of a personal CNC machine. Throughout all the posts, the intent is to give as much information as possible to help others who aspire to build their own but have many questions or don't know how to start. There are currently many DIY websites and blogs on CNC machines but this blog is written to give information overload so that those who have very little experience can follow. This is written from two university students' point of view with a starting knowledge that was minimal in terms of CADing, machining, electronics, and testing.

The rest of this post is a bit of an intro as to how each of us got into engineering. If you are trying to pack as much information into your brain because you are a student on work term with only four months to build the machine, a work term report, and that PD20 course (which is totally useless in my opinion) then skip the rest of this post.


Hello,

My name is Daniel Choi and I am an engineering student at the University of Waterloo. When I was young and naive, I said, "I want to be an engineer!" My dad is an engineering and I thought it was a cool job because he did it. Little did I know that was what I would end up studying one day.

When looking back, I can see that a few key things contributed to the birth of an engineering career. A lot of my childhood consisted of building lego vehicles with my brother and adding as many clear coloured circle pieces to the guns of our ships to make them "stronger". Also I enjoyed challenges to various problems like how to build some contraption around an egg to protect it when dropped from the second floor window.

Side note: If you are a parent reading this, I suggest that if you think your child has the slightest interest in engineering, start young! Get a cheap 3D printer and allow your kids to make designs on the computer. Buy them things to build or throw project ideas at them to try. Don't let them sit on the computer all day on facebook or become couch potatoes.

However, it wasn't until I enrolled in a computer engineering course in grade 10 at Sir John A. Macdonald SS (Waterloo) that I found my interests in engineering become more than words. This course taught basic electronics and circuits and allowed students to have hands on experience with manufacturing PCBs from a copper board to flashing fancy LEDs. Some of the projects included a line following robot and an LED chaser. Schematics were provided for both these projects but the board layout was left up to the individual. We used traxmaker to layout the components and etched our personal layout with chemicals that are probably not good for bare hands. Then we were given the privilege (or nightmare) to drill all the holes for through hole components with a fragile $1 drill bits (good thing I never broke one and still have them in my pencil case) and trusted with hot soldering irons to mount all the individual components. For the line following robot, we made the PCB and programmed a MCU to have the robot follow a windy oval loop made of electrical tape. Phototransistors and a super bright LED were used to detect when the robot was deviating from the main path.

After having a blast in grade 10 without any burns or cuts, I decided to continue on with the grade 11 course taught by the same awesome teacher. The main project for this term was to build a sumobot that would face 1 on 1 in a ring with another sumobot and try to push the competition out or immobilize them if that was even possible... All the parts were built and and assembled by my partner and I. In the class competition, our robot came out on top with a tough best out of five finals.

It came down to the last match where it was tied 2-2 and all the eyes in the class were glued on two unprofessionally made robots. When the ON switches were flipped and the robots started spinning in search for their target, the class erupted into cheering. After a couple dances around the ring and a few light nudges exchanged, there was a window of opportunity as our robot stopped, locked onto the opponent, charged forward (at 5cm/s more or less) and made full contact with the opponents side. Everything went in slow motion (partly because the opponents robot weighed a ton and was hard to get under) as all of our efforts for the win (an 100% on the project) could come down to this final collision. All of our upgrades to the robot were being put to the test; the enhancement to the plough with the two curved strips of metal (seen below), countless hours of filing done by my partner on the tip of the plough, and a hot glued hunk of aluminium under the plough for a tighter fit against the ground. The final moments could be described as a battle between a spatula and a stubborn fried egg with the spatula slowly peeling the egg off the frying pan, onto the surface and finally flipping the egg over. In this case, the opponent's sumobot was flipped on its side and the crowd went wild!

This is the only shot of the two sumobots battling with mine on the left (the true outcome is not reflected in this picture!)

In grade 12, we were given the opportunity to make a firefighter robot from scratch. Multiple sensors were used like infrared to detect a candle flame and wall distances and phototransistors/super bright LEDs to detect a line. The coding was the main challenge in this project in order to have the most efficient robot (and get 100%). In the end, my fastest time around the course (and fastest in the class) was about 27 seconds. This was achieved by increasing the voltage to the motors but required the response to sensors to be just on time and a few tweeks to the code to compensate for the higher speed.

This video is about 1 minute long with motor voltages at a nominal 9 volts...I think...it was too long ago (if not 9, then 5 volts)

In my grade 12 year, I was accepted into Mechatronics engineering at the University of Waterloo and currently I am in my third year. So far I have had 3 work terms and am in the middle of my fourth but realize that I have very little exposure to the practical side of engineering (programming, designing circuits, machining, etc). Last work term (summer of 2013) I made a moving target using an arduino, a servo motor, timing belt and pulls, a drawslide, a 10-turn potentiometer, and accelerometer. This was a fun little project but was not very impressive because the target moves at around 5cm/s and is easy to shoot from 15 meters away. I plan to upgrade the motor to a more powerful stepper motor one day when I get around to it. 

In search of a "bigger" project for this work term, I stumbled upon a CNC machines one day during the Fall 2014 term of school. I got very excited since I would be able to develop my practical skills in all the areas listed above and I would be making something useful. At this point I share the idea with Daniel and got him on board. We hope to have most of the mechanical design completed by the end of April 2014 and maybe some electronics hooked up as well. Stay tuned!

Hi,

I'm Daniel Mok, I'm also a mechatronics engineering student at the University of Waterloo and another person working on this project.

Why did I choose engineering? This is a difficult question to give a straight answer to, but it mainly started from an early interest in math and science. Drawing was also a big interest of mine, and among my doodles there were a fair amount of amateurish designs for fantastical machines that I hoped to one day bring to life. As Daniel Choi has mentioned, having parental support is a huge help: much of my early exposure to engineering came from books, magazines, some summer camps, and of course Lego.

It was not until high school that I started to seriously consider engineering as a career, during which I was strongly influenced by my technology and computer science courses. During this time I was able to work on several projects which continued to build up my interest in engineering, such as building a sumobot, learning woodworking, creating designs with AutoCAD, and developing a simple Android game. My manufacturing course was also where I first encountered CNC machines, where I used one to carve out a design into some wood.

I am now in my third year of engineering and I really enjoy my program. Last summer I was thinking about potential engineering projects and thought about making a CNC machine. One term later I was asked by Daniel Choi if I wanted to help him with creating one, so here I am. :)

Closing Remarks

One last note, you might be wondering why the name of this website was called slim CNC mils. We wanted a name that was unique (not boring like diy CNC tutorial). If you haven't already noticed, you can flip "slim CNC mils" around and it spells itself. Also, we are trying to make the CNC machine as cheap as possible yet with the best components we can get for our dollar. So if you are also on a tight budget, there will be some tips on how to save money and also ballpark prices for components mentioned.

Thank You

We would like to thank Chris McClellan for providing a lot of the aluminium used to machine our smaller mounting and connecting pieces.