Machine Design #09: Selecting a Motor-Driven Transmission – Don't Just Look at Force and Speed
When designing a mechanism driven by a motor, I often do not start from the question "which motor to use", but from a simpler one:
What does this thing need to move?
It sounds ordinary, but in practice this is where it is very easy to choose wrong.
The same horizontal axis, but pulling a light camera unit of a few kg is completely different from pulling a heavy jig unit of a few tens of kg. The same 1-meter stroke, but it is one thing if you only need to bring a product from A to B, and quite another if you need to stop at an exact position for inspection.
So when choosing between a ball screw, a timing belt, rack and pinion, or a chain, I think you should not look at force and speed in isolation. You have to look at the stroke, the accuracy, the eccentricity, the environment, the maintenance, and also what happens to the mechanism when the machine faults.
Many mechanisms look fine on the drawing. But once mounted on a real machine and run for a while, you start to see noise, vibration, belt stretch, tooth backlash, missing guards, or an uneven load that quickly degrades the guide.
Here I write down a practical view of choosing a motor-driven transmission in an automated machine.
Before choosing a mechanism, ask a few questions first
In machine design, if from the very start you think "an actuator will do here" or "just use a ball screw here to be safe", that is a bit dangerous.
Not because those choices are wrong, but because you are choosing components before understanding the problem clearly.
Before choosing a transmission, I usually want to clarify these points:
- How heavy is the load to be moved?
- What is the stroke?
- How long should the full stroke take?
- Is accurate stopping needed?
- Is the position repeated many times?
- Is the load eccentric?
- Is the axis horizontal or vertical?
- Is the environment clean, or is there dust, oil, heat, water?
- Is there a requirement for low noise?
- On power loss or a control fault, is there any danger?
- Later, will maintenance, replacement, and readjustment be easy?
Change just a couple of the conditions here, and the transmission option can change completely.
For example, the same task of pulling an object 1 meter. If the object is light, the environment clean, and it needs to run fast, a timing belt can fit very well. But if the object is heavy, the environment dusty, and high accuracy is not needed, rack and pinion or a chain is more practical.
Machine design is not about choosing the "fanciest" mechanism. Machine design is about choosing the mechanism that best fits the real conditions of the machine.
Ball screw – Very strong on accuracy, but not always the one to use
When it comes to linear motion with position control, the mechanism that comes to mind most easily is:
motor + ball screw
This is a very familiar mechanism in automated machines. The motor creates rotary motion; the ball screw converts it into linear motion.
The advantages of a ball screw are high efficiency, low friction, the ability to produce large thrust, and good position control. So it is often used for mechanisms that need to stop accurately — for example, a pressing axis, a small lift axis, an alignment mechanism, a camera unit, an inspection unit, or the motion axes in an inspection machine.
If chosen right, mounted right, and used with the right load, a ball screw runs very stably.
But the point that is easily misunderstood is: a ball screw does not automatically make the design more solid just by being included.
When does a ball screw fit?
A ball screw usually fits cases such as:
- Relatively good position accuracy is needed
- The stroke is not too long
- The load is not too eccentric
- Large thrust is needed
- The speed is not too high
- Steady motion is needed
- The machine space allows the screw, bearing supports, coupling, and motor to be arranged
In an automated machine, if it is a light pressing mechanism, position alignment, moderate lifting, or an axis that brings an inspection unit to a fixed position, a ball screw is a very strong choice.
But a ball screw also has limits
The first issue is speed.
A ball screw converts rotary motion into linear motion. To run fast, the motor must spin fast. If the screw has a small lead to get large force or good resolution, the motor must spin even faster to reach the same linear speed.
At some point, the motor is no longer suitable, or the screw itself is limited by critical speed, vibration, or resonance.
The second issue is a long stroke.
The longer the stroke, the longer the screw. A long screw sags easily, is hard to align concentrically, is hard to mount accurately, vibrates easily when spun fast, and costs more. Then the mechanism looks simple on CAD but is uncomfortable once actually assembled.
You also have to watch backlash, preload, the accuracy grade, how the two ends are fixed, lubrication, dust protection, and the life of the nut.
In short, a ball screw is very good for accurate motion and large force. But if the stroke is long, the speed high, or the environment poor, it is not necessarily a clean choice.
Rack and pinion – Good over long distances, clear structure, but you accept backlash and noise
Rack and pinion is a mechanism that uses a gear meshing with a rack to create linear motion.
This mechanism is very easy to understand: the gear turns and runs along the rack, or the rack moves depending on the layout.
The biggest advantage of rack and pinion is that it is easy to make a long stroke. If you need to travel 2 meters, 3 meters, or more, rack and pinion is usually more practical than a ball screw. The rack can be extended, the structure is fairly clear, easy to picture, easy to inspect.
For mechanisms such as a gantry, a long slide table, a heavy-load pulling mechanism, or in-shop transport equipment, rack and pinion is a very worthwhile option.
The strengths of rack and pinion
Rack and pinion fits when:
- The stroke is long
- The load is relatively heavy
- Very high accuracy is not required
- A strong structure is needed
- You want a mechanism that is easy to understand and maintain
- The machine space allows the rack to be arranged
Another nice point is that the layout is fairly flexible. You can fix the rack and let the motor move, or fix the motor and let the rack move.
But if the motor moves along the axis, you have to also think about the power wiring, encoder wiring, air tubing, cable carrier, and wire protection. This part is often underestimated on the drawing, but it becomes a real nuisance on the actual machine.
The weaknesses of rack and pinion
The biggest weakness is backlash between tooth and tooth.
If the mechanism runs in only one direction, or stops without needing much accuracy, the backlash can be acceptable. But if it reverses often, or needs a good stopping position, the backlash shows up very clearly.
You also need to watch:
- The mounting accuracy of the rack
- The parallelism between the rack and the rail
- The meshing of the gear
- Lubrication
- Tooth wear
- Noise
- A cover against dust and grease
- A mechanical stopper and a stroke-limit sensor
Rack and pinion runs over long distances, so you must pay special attention to overrun. If a sensor fails or the control is wrong, the mechanism can run past its limit. So do not trust the software limit alone. Have a limit sensor and a clear mechanical stopper.
Rack and pinion is a good mechanism, but a long stroke does not mean you can use it right away. If you are not used to designing it, the installation, lubrication, and stroke safety are easily missed.
Timing belt – Fast, smooth, light, but don't overwork it
The timing belt is a mechanism very often seen in modern automated machines. Many off-the-shelf long-stroke actuators also use a belt inside.
The strengths of a timing belt are that it is light, runs smoothly, is quiet, needs no lubrication like a chain or a rack, and reaches high speed easily. For light or medium loads, a timing belt is a fairly nice choice.
For example, mechanisms such as a camera axis, a light gripping mechanism, a small-load XY axis, a product-sorting mechanism, or a light slide table — a timing belt usually fits very well.
When a timing belt fits very well
A timing belt fits problems such as:
- Light or medium load
- A relatively long stroke
- High speed is needed
- Smooth motion is needed
- You do not want grease lubrication
- Very large thrust is not required
A belt actuator, chosen right, can save a lot of design time. You do not need to design the pulley, tensioner, cover, rail, and sensors yourself too much. The maker has standardized it fairly well.
But even with an off-the-shelf actuator, you still have to understand that inside it is a belt. And a belt still has the limits of a belt.
Watch out for belt stretch, heavy loads, and vertical axes
A timing belt can stretch and vibrate if the stroke is long, or skip teeth if chosen wrong. With a heavy load and large acceleration, the force on the belt is not small.
A common mistake is using a belt for a vertical axis without thinking carefully about the case of power loss or a broken belt. If the load falls, it not only damages the machine but is also dangerous for the operator.
For a vertical axis, check carefully:
- Does the motor have a brake?
- On power loss, does the load fall?
- If the belt breaks, what stops the mechanism?
- Is there a stopper or a drop-prevention mechanism?
- Can a sensor detect an abnormality?
A timing belt is very good if used in the right place. But making it pull too heavy a load, bear large impact, or serve as the safety mechanism for a vertical axis without extra protection is not sound.
Chain – Strong and durable, but noisy, dirty, and needs maintenance
A chain drive is a familiar mechanism in heavy-load transport systems. Compared to a belt, a chain bears load better, is more durable in a worse environment, and is often used for conveyors, pallet-pulling mechanisms, large lift mechanisms, or in-shop transport equipment.
The advantage of a chain is strength. If you only need to pull a heavy load, run in one direction, and do not need high accuracy, a chain is a very practical choice.
But a chain also has its price.
A chain is usually noisier, vibrates more, needs lubrication, has wear, has play, and needs periodic maintenance. If the machine is in a clean environment or near an area that needs little dust and oil, using a chain may not be clean.
When does a chain fit?
A chain fits mechanisms such as:
- Heavy-load transport
- Conveyors
- Product- or pallet-pulling mechanisms
- Environments with dust, oil, or near the outdoors
- Motion that does not need to be very accurate
- A strong mechanism that is easy to replace
In many machines, a chain does not need complex position control. The motor just spins and the chain pulls the product along. So it fits transport-type motion more than accurate positioning.
Designing a chain is not just choosing the right load
When designing a chain, do not just look at the catalog and choose the chain by pulling load. The parts around the chain matter no less.
You need to watch the sprocket, guide, tensioner, cover, lubrication, guide wear, chain slack, and also safety if the chain breaks or comes off.
Many chain mechanisms run at first, but after a while become noisy, vibrate, sag, wear the guide, or spread grease everywhere. If maintenance is not considered from the start, the shop floor will struggle later.
A chain is strong and practical, but it is not the right choice for a clean, smooth, accurate mechanism, or one that wants little maintenance.
Cam – Very nice mechanically, but not flexible
A cam is a mechanism that creates motion following a pre-designed profile. One rotation of the cam can create a certain motion cycle.
Today, in many automated machines, cams are not as common as before, because servos, actuators, and robots have become much more convenient. But that does not mean cams are poor. Used in the right place, a cam is still very strong.
A cam can give fast, stable, well-repeated motion and very neat mechanical synchronization. For a machine producing large volumes with few model changes, a cam can be a very effective option.
The issue is that a cam is harder to design and harder to make, and when you want to change the motion, it is not easy. Once you have designed the cam and the interlocking mechanisms around it, modification is usually quite heavy.
A cam fits stable production, repeated cycles, and high speed. A cam does not fit well a machine that frequently changes model, stroke, or product.
Link mechanism – Nice and compact, but you must understand how force passes through
A link mechanism is the kind that people who love mechanics usually really like. With just a few links and a few pivots, you can create fairly interesting motion: push, pull, rock, lift, fold, open, clamp.
The advantage of a link is that it can create repeated motion mechanically, sometimes very compactly and without a separate actuator for each motion. In some machines, a link helps reduce the number of motors or cylinders.
But a link is not as easy as it looks in simulation.
The force passing through each joint and each link changes with the angle. There are positions where the force is very nice, but also positions where the mechanism almost jams or needs a very large force. You also have to account for radial load, axial load, joint play, bushes, bearings, lubrication, and the rigidity of the link.
Another difficulty is adjustment.
If the link is already interlocked with many other units, adjusting one small position can drag along many things. So when designing a link, think in advance about where it can be adjusted, where it is fixed, where a slot is needed, and where a stopper is needed.
A link is very nice, but if used carelessly, adjusting the machine later becomes very tiring.
Off-the-shelf actuators – Really convenient, but not to be used by feel
Today, using off-the-shelf actuators in machine design is very common. For example, an electric cylinder, a robot cylinder, a linear actuator, a ball-screw actuator, a belt actuator.
This is a reasonable trend. Instead of designing the motor, coupling, screw, rail, bearing supports, cover, and sensors yourself, you can buy a standardized unit. This reduces design time, reduces assembly errors, and is easier to maintain.
But there is one point to state clearly:
An off-the-shelf actuator does not replace mechanical-design thinking.
Inside the actuator is still a screw, a belt, a pulley, a rail, a bearing, or similar mechanisms. If you do not understand its nature, it is very easy to choose wrong.
For example, a screw actuator can bear axial force well, but if the load is too eccentric, the moment on the guide will shorten its life.
A belt actuator can run fast over a long stroke, but is not suitable if the load is heavy, the thrust large, or high rigidity is needed.
An electric cylinder can be very convenient, but if used for a vertical axis you have to check the brake, the falling load, power loss, and the safety mechanism.
An actuator helps you work faster, but that does not mean you can skip the calculation. At minimum you still have to check the load, moment, speed, stroke, duty, accuracy, mounting orientation, and life.
A real example of choosing a mechanism
Suppose you need to move a steel block of about 100 kg over 2 meters in about 3 seconds.
Looking at it simply, you might think: use a bigger motor and it is done. But in practice it is not that simple.
With a 100 kg load, a 2-meter stroke, and 3 seconds, the mechanism has to accelerate, decelerate, bear inertia, and bear relatively large force. If you use a ball screw, the long stroke and speed can make the mechanism unattractive. Here, rack and pinion or a chain can be more practical, depending on the accuracy requirement and the environment.
If the environment is near the outdoors, dusty, the product is steel, and accurate stopping is not needed, a chain or rack and pinion is worth considering.
Conversely, if you only move a 1 kg plastic unit over 1 meter indoors, with a requirement for smooth running and no heavy load, a timing-belt actuator can fit very well. If more accurate stopping is needed, a ball-screw actuator is also a candidate.
The same "move an object from A to B", but with different conditions, calls for a different mechanism.
An easily forgotten part: eccentric load and moment
In an actuator catalog, the allowable-load figure looks easy. But when brought into a real machine, the problem is usually not the simple vertical load or axial load.
The problem is often the moment from an eccentric load.
For example, mount a jig unit on an actuator. The weight may not be large, but the center of gravity sits far off from the center of the guide. When it runs fast or stops suddenly, the moment on the guide rises. If you only look at "how many kg the load is", it is easy to under-select.
This is a fairly real mistake. The machine still runs, but after a while the guide develops play, the actuator makes a strange noise, the accuracy drops, or the life is shorter than expected.
So when choosing a transmission, do not only ask "how many kg". Also ask:
- Where is the center of gravity?
- Does the load stick out far?
- On acceleration, how much does the moment rise?
- Is there an external force acting on it?
- Is the mechanism clamped, pressed, or impacted?
Often, just changing the layout so the center of gravity is closer to the guide makes the mechanism much stronger, without increasing the actuator size.
Don't forget maintenance
A mechanism that is good on the drawing but hard to maintain still causes trouble once it goes into production.
For example, the grease-injection point is too hard to reach. A sensor is blocked, and to adjust it you have to remove a cover. To re-tension the belt, you have to remove half the machine unit. The chain has grease but no drip tray or cover. The rack needs lubrication but sits right in an area that easily collects dust.
These things may not look serious in 3D design. But when the machine runs every day, they become a real problem.
I think that when choosing a transmission, you should also look at who will maintain it later, how they will maintain it, how long it takes, and whether other units have to be removed.
A good design is not just one that runs. A good design is one that runs stably and does not make the person maintaining it later suffer.
No mechanism is always right
If you ask which mechanism is best, the answer is usually: it depends.
A ball screw is good for accuracy and thrust, but not attractive for a very long stroke or very high speed.
A timing belt is fast, smooth, light, but not suited to a very heavy load or a mechanism that needs high rigidity.
Rack and pinion travels far well and bears load fairly, but has backlash, noise, and needs attention to lubrication.
A chain is strong, durable, and withstands a poor environment, but is noisy, dirty, and needs maintenance.
A cam and a link are very nice for repeated motion, but not flexible when the product changes.
An off-the-shelf actuator is very convenient, but you still have to understand the mechanism inside and check the real load.
So when choosing a transmission, do not follow the "this is always right" style. In machine design, the most correct option is often not the fanciest one, but the one that best fits the production conditions.
Conclusion
Choosing a motor-driven transmission is not just choosing a motor, a screw, or an actuator.
It is the process of balancing load, speed, stroke, accuracy, rigidity, life, cost, installation space, operating environment, maintenance, and safety.
A mechanism that runs on CAD is not guaranteed to run stably in the field. A mechanism calculated to have enough force is not guaranteed to be easy to assemble. An actuator whose catalog says it bears the load is not guaranteed to live long if the load is eccentric.
So before choosing a mechanism, return to the initial question:
What am I trying to move, under what conditions, and if the machine faults, what will happen?
Answer that clearly, and choosing between a ball screw, a timing belt, rack and pinion, a chain, a cam, a link, or an off-the-shelf actuator becomes much easier.
In machine design, knowing many kinds of mechanisms is necessary. But more important is knowing how to use the right mechanism in the right place.
A note from MINATA
MINATA shares these notes as reference material from a practical machine-design point of view. If you would like to discuss the choice of a transmission, confirm drawings, compare options, or interpret Japanese–Vietnamese engineering terms during design, feel free to get in touch and we can review it together.
View all MINATA technical articles