Time Lapse Rail Motion Control Design 101.

Time Lapse Rail Motion Control Design 101.

CHRONOSWEB6

Introduction

I am writing this document in an effort to put together a guild to help those who wish to build their own timelapse system. Designing and building your own system can be a very rewarding process, and it can also save quite a bit of money. Most of the parts can be found off the shelf, with a little ingenuity you can build your own brackets and have a system running in a matter of days. However, if you are not careful about your part selection it can also be a frustrating process that can get very expensive, and in the end it may have been cheaper to just buy a turn-key system. My goal is to help you avoid costly mistakes and to give some guidance to help you design your own system.

Design Requirements.

The design requirements will dictate your build. This is the cornerstone of the development process and everything about your build will be reflected by this. I could not possibly come up with all the considerations but these are generally what I consider when designing a system. The main focus in this article is on linear motion systems as this is usually the best place to start in order to get that 3 dimensional depth. These considerations should apply for any other type of motion control as well.
  • Usage goals. Do you care about video speeds? Do you want to focus on timelapse only? Is this something you want to be a Hybrid between a video dolly and a timelapse motion system? I no longer consider these to be the same tool, and would strongly recommend focusing on either timelapse or video. I will get into this more later in chapter 8, chances are if you are actually capable of designing a system that is able to do both and to do both well, then you would not be reading this article in the first place. I dont want to discourage anybody, but i highly recommend reading about Timelapse Vs Video before attempting to integrate the two.
  • Length of movement. How far do you want to go? My personal preference is 1000mm. That offers more than enough distance (usually about 36 inches of travel if you keep the carriage slim) to create a nice parallax effect. My go-to routine is 360 frames over 30 minutes with about 34 inches of travel. This gives about 12 seconds of timelapse video at 30FPS and by keeping the dolly a bit closer to the ground you can pull dolly moves that appear to have much more travel. There are several systems on the market that offer 6 feet of travel, this does give a little bit more flexibility than a shorter rail. Or do you need even longer? if you are looking into 20-30 feet of movement then a rail is not really the best option. If you need to go 300-400 feet then a suspended cable or even a hyperlapse staff would be a better choice, but a staff would have a hard time pulling off 3-4 foot movements.
  • Portability Do you plan to go hiking to remote places with your rail? Do you have an assistant that carries all your gear? are you a single persion operation? Or is this something that will live in a studio enviornment?
  • Length is once again a consideration. If you plan to go hiking to remote spots in the wilderness then you might find longer systems can get snagged and be a pain to manage. If you are building a custom rig for a warehouse that will be used once or you are building a setup for a studio then length should not be a problem.
  • Weight. This is really only a concern if you are going to remote places. Chronos Lite weighs in just over 5lbs with the rail, controller, battery, and built in feet making it ideal to get to remote locations. However at 5lbs its max payload is only rated at 15lbs and may not be suitable for heavier payloads.
  • Power. How do you plan to power the system? Right now most systems can be powered off of a 12v DC source. If you are working in a studio power should not be much of a problem, but if you are in an area prone to blackouts an UPS would be a good idea. If you are working in remote locations with no access to power you are pretty much stuck with batteries. Most timelapse rails are fairly power efficient and can shoot for a full weekend off a single 4800MAh battery.
  • Motion Accuracy. This was a big one for me. I wanted to build a system that was capable of offering full motion control over extremely short distances for macro photography. For a system like this the belt drive was immediatly thrown out in lieu of a lead-screw drive commonly found in CNC for thier accuracy levels. Most belt drive systems offer 1/125th of an inch resolution, which is pretty good, however for motion ramped macro shots i needed something that was capable of much higher resolution. If you do not need this level of accuracy you can save money on the pricey lead-screw and instead go with a belt drive
  • Expandability Linear motion is fine and good, but what about adding a pan and tilt? how would that be done and how would it integrate? This is something you need to keep in mind when designing the control system.
  • Environmental concerns. Another thing to concern is whether or not any weather sealing concerns will need to be addressed. Humidity levels, Rain, Snow, extreme heat or cold. Are the parts you are selecting capable of performing in these extremes? LCD’s tend to freeze in extreme cold, whereas controllers may overheat in extreme heat.
    CHRONOSWEB7
  • Motion Goals What kind of movement do you have in mind? Do you need something capable of taking 1000 images in a single centimeter? are you interested in velocity ramping, do you want to do continuous movement or move-shoot-move? These are all things to consider.

Different configurations for Linear motion

Self Propelled Cart

  • Comments. The first design i tried for a timelapse system was a self propelled cart. The first version was made of wood and used a simple motor to propel the cart. It was a continuous movement system and i built a track for it to run on, it did pretty well on smooth surfaces. The next incarnation was made out of a VEX robotics kit and controlled by an Arduino and pair of stepper motors offering linear and panning movement. It worked well but i was also spending time dealing with post production work cleaning up the movement and stabilizing the footage. This can be a very effective design but is very limited in terrain options, you can start adding track which can become cumbersome.
  • Benefits. 
    • These can get pretty small and pretty portable. With a nice clean surface to ride on they can travel very long distances while snapping pictures.
    • These are pretty easy to design and assemble, there are various roboticskits online that can provide a good chassis and building structure to build your system upon.
  • Drawbacks
    • They can be pretty limited in the terrain they can handle.
    • Vertical movements can be very tricky
    • any dirt or gravel they pass over can introduce jerkiness and bumps in the final video and require extra time smoothing the motion.

Suspension Trolly design

  • Comments.
    • I have not yet tried this design, however it looks like it could be a lot of fun. One of the challanges will be dealing with wind and providing power to the carriage as it travels longer distances. A friend of mine and i have discussed using a pair of rotating gyros about 20 degrees off axis to provide an active stabilization method. I have seen a couple of these designs for sale, they are usually fairly expencive and while they offer very long distance of travel thier limitations due to enviornmental factors makes them very limited in my opinion. A system like this would be a good addition to a timelapse equipment collection, but might prove to be too limiting as a primary time lapse motion control system.
  • Benefits
    • Long Travel, lightweight, these can span very large distances for some real awesome footage.
  • Drawbacks
    • Very susceptible to vibrations and motion introduced by wind.

Timelapse rail

  • Belt drive

    • Comments.
      • This is a very solid design made popular in large part by the Stage Zero dolly by Dynamic Perception. This seems to be a pretty good blend of top speed as well as timelapse capability, however with many designs motor swaps are still required to help achieve video dolly speeds.
    • Benefits
      • Varying track length, 16-20 feet have been done quite often.
      • belts are inexpencive
      • high speeds are possible
    • Drawbacks
      • motion accuracy not overly impressive
      • belts flapping in wind may cause vibrations
      • motor must remain powered during vertical movements
      • backlash may be introduced if belt is not properly tightened
  • Lead-screw

    OLYMPUS DIGITAL CAMERA
    • Comments.
      • As far as I know project Chronos is the first timelapse system that was designed utilizing a lead-screw design. This is basically a single axis off a CNC machine, i opted to go this route due to the extreme accuracy levels of CNC. My goal was to design the ultimate time lapse rail and I did not put any thought into video speeds, I have another article discussing this further and felt that trying to achieve video speeds was an attempt to merge two different tools into a single package. Instead of going that route i decided to focus on timelapse motion only. The result is the most accurate time lapse system ever built, and it is extremely efficient as well.
    • Benefits
      • supports very heavy payloads
      • unparalleled accuracy
      • able to shut motor off between movements on vertical climbs making it very efficient.
      • Easy to achieve zero backlash
      • easy to achieve repeatable motion
    • Drawbacks
      • high speeds extremely difficult to obtain
      • limited to shorter rails.
  • Rack and Gear
    ChronosWeb1

    • Comments.
      • I have only seen one design using the rack and gear system. I really do not see any benefit to this design other than modular tracks, Perhaps somebody will explore this design and find some good reasons for it, but i have difficulty seeing how this design fixes any problems with other designs.
    • Benefits
      • using modular racks you can achieve longer track length
    • Drawbacks
      • Heavy gearing will be required to give this system enough strength to do vertical lifts.
      • gear racks can be expensive
  • Cable/pully

    • Comments.
      • Of all the possible designs for linear motion this one is my least favorite. It does not offer any solution to address problems of any other design. The cable-pully rail systems I have seen usually have a lot of backlash. I suppose better material selections can be done, but in my view this is a poor alternative to a belt design. The one redeeming quality is the simplicity of the build, just attach a motor with a pulley on one end of the track and a cable from the pully to the carriage.
    • Benefits
      • very simple to build.
    • Drawbacks
      • repetition virtually impossible
      • lots of play/backlash in cable systems
      • not very accurate

Hyperlapse Staff

  • Comments.
    • This is one of my favorite types of timelapse. The idea is simple, you have the camera on a monopod, aim it at something in the distance, take a picture, move a few feet, re-aim the camera and take another shot. This can be as simple as a monopod, or a computer controlled leveling and aiming platform on a stick. A properly dont hyperlapse is a very impressive clip, but you will always have to spend extra time in post production smoothing out the clips. Some day i plan to take on the challenge of a self-leveling and aiming gyro/gps enabled head.
  • Benefits
    • Light weight
    • able to move across vast distances, hundreds of feet making very impressive timelapse sequences
    • simple designs can be inexpencive and simple to build.
  • Drawbacks
    • Lots of technique work, can be very difficult to use
    • only good for longer runs

 

Control

CHRONOSWEB5

  • Comments. This is the most important thing to consider period. How far down the rabbit hole do you want to go? There are already several pre-made timelapse controllers on the market that can be used with various motion control designs. Or you can opt to build your own. They can be as simple as a voltage regulator attached to a motor to control speed, to a multi axis and keyframing system that runs on a laptop.
  • Type of Movement
    • Continuous. This is the easiest method. ALl you have to do is get the system moving at a very slow and steady rate. The camera mounts to the rail and can be triggered by the controller, an external intervalometer, or even a built in intervalometer.
      |
    • Shoot Move Shoot. This is a bit more advanced, and this is my favorite type of motion control. This requires the controller to move the camera, stop, trigger the camera, then repeat. This allows extremely long duration timelapses and longer shutter speeds for astro timelapse.
    • Additional motion control options
      • Lead in/Lead outs used to gather frames for a brief time before motion begins or after it ends.
      • Motion Ramping. This is used to gradually increase speed while using Continuous motion, or to incrementally increase distance between frames on Shoot-Move-Shoot sequences to give smooth starts and ends to motion in the final output video. Chronos 2.0 uses Linear Ramping rather than S-curved ramping, i have used both types of ramping extensively and never seen any visual difference between the two forms of ramping. I opted for Linear ramping because it uses less code and is easier to predict.
        • Linear Ramping uses a simpler form of mathematics and results in very predictable and understandable settings. ADD IMAGE
        • S-curve will use gradually increasing and decreasing amounts of motion to offer an s-curved shaped ramp. In theory this should offer a smoother tranition at the beginning and end of the ramp. ADD IMAGE
  • Camera Control
    • Do you plan to control the camera? if so, you need to figure out how to connect the camera to the controller. Most camera work with a 3 lead system, with a Focus, Shutter and Ground pin. By shorting the Focus pin to ground the camera understands to focus the camera. By shorting the Shutter pin the camera knows to release the shutter. This is an active circuit provided by the camera, and it does provide a level of voltage across the lines. Some brands like Nikon require both Focus and Shutter to be shorted out in order to trigger. It is always recommended to keep circuit isolation to prevent the cameras circuitry from making a physical connection to the circuitry in the controller, in order to separate the two it is common to use a relay or optoisolator. Both relays and optoisolators (or optocouplers) do the same thing but in different ways. Optoisolators do not have moving parts which many people favor. A good resource for shutter release and cables can be found at http://www.doc-diy.net/photo/remote_pinout/ADD IN IMAGE OF RELAY
      By using 2 optoisolators you can control both the shutter and focus of a camera. I have never seen a benefit of this and usually combine the Shutter and Focus pins together as you will always be using manual focus with timelapse
    • Microcontrollers This will probably be the brain of your system. Unless you want a very simple design the easiest way to achieve a controller that has some sort of featureset would be to use a microcontroller. These are usually small circuit boards that utilize C or a variant of C programming language. One of the easiest to learn would be the Arduino as there are lots of excellent online tutorials that teach you how to program. If you are looking for something more robust there are also much more powerful linux based systems such as the Raspberri Pi. By learning to code these boards you can develop excellent control of the timelapse rail as well as the camera.

Motor decisions

  • DC motors. These motors are commonly found in all sorts of products. They use voltage to generate movement by creating a magnetic field that rotates the shaft. By varying the amount of voltage you can adjust the speed of the rotation, however these normally run at higher speeds and will almost always require a gearbox to reduce the speed of the output shaft and to increase the torque of the motor. The benefit to DC motors is they do not require a microprocessor or driver chip in order to run.
    chronosweb3 

    • Considerations when using DC motors. DC motors are not the most accurate unless you provide some sort of feedback. They are simple to use and integrate however so if you are not concerned about repeatability they can be very good. They are either ON and moving or OFF, not sure how well these would work for vertical climbs when shooting SMS. They can also be slowed down by weight, so a speed that gives you 1 hour end to end travel in a horizontal position may end up taking 1.5 hours when moving vertically. Thus they are somewhat unpredictable.
  • Stepper motors work differently than DC motors as they do not freely spin when voltage is applied, instead it uses a more complicated magnetic structure that allows you to move steps. Most steppers have a 1.8 degree step angle, meaning that each step will adjust the position of the shaft 1.8 degrees. If you move 200 steps you will move one revolution. By adjusting the amount of time between steps you can increase or decrease speeds. Depending on the motor driver you can also introduce Microstepping where it can cut the step into multiple smaller steps, generally 8, 16 or 32 microsteps per step. The benefit in using a stepper motor is it naturally runs a much lower speed and does not always require gear reduction. You can also position the stepper to whatever angle you want by telling it to move X amount of steps. As you learn more about controlling steppers and coding you can actually count the number of steps taken and be able to recall specific locations for repeatable moves. Stepper motors are more complex and do require special driver chips to coordinate the motion.
    ChronosWeb2 

    • Considerations when using Stepper Motors. They naturally run much slower than DC motors and it might be temped to just put a pulley on the shaft and put a belt around it, which will probably not yield very good results. If you are planning on building a system that can be powered off battery you should look into some gearing to provide a torque increase on the stepper. Also, if a stepper loses power or if the torque exceeds its capability it will spin freely, so using a non-geared setup on a vertical incline risks damage to equipment if it breaks loose. If properly implemented you can achieve much greater levels of accuracy and predictability with Stepper motors over DC motors.
  • Servos are essentially DC motors in tiny gearboxes and use a feedback mechanism to relay thier position back to their controller. I cannot comment too much on Servo motors because I have very limited experience with them, but they are commonly used in robotic applications.

Gearing options

  • Why use gearing? Well the most obvious reason is to slow the speed down to manageable timelapse levels. If you are using a DC motor you will absolutely need some level of gear reduction. You might be able t get away with no gearing on stepper motors but as mentioned when discussing steppers that has its risks. Chronos uses a lead-screw directly coupled to the shaft of the motor which gives a steep level of gear reduction. In the case of steppers gear reduction gives higher levels of motion accuracy.
  • Backlash is something to consider. This is the amount of play in the system with the gearing. If you pick up a geared motor and try to spin the shaft you will notice that there will be a tiny bit that it can move before the gearing prevents movement. generally you want to have as little backlash as possible, but in most setups a little bit of backlash is acceptable and can even be compensated for in the software by moving the motor a few steps in the opposing direction anytime you switch directions.
  • Common types of gear reduction
    • Motors with Gearheads. These are commonly used, especially with DC motors. There will be a gearbox attached to the end of the motor with build in gearing to reduce the speed of the motor. For DC motors this will almost always be the first stage of gear reduction.
    • Pulley size. This is another way to reduce the speed, use a small pulley. If you are using a belt-drive system the larger the pulley the faster the movement, however the larger pulleys result in torque reduction and may decrease the max payload to unacceptable levels.
    • Worm Gears are popular as an excellent way to get some heavy reduction, if using a single lead worm with a 50 tooth worm gear then the result is a 50:1 reduction meaning the motor must spin 50 times for each rotation of the geared shaft. it can be difficult eliminating backlash completely from the system.
    • Lead-Screws or ball screws offer a different type of reduction and can be virtually free of backlash. By turning the lead screw each rotation will move the carriage a certain distance depending on the pitch. On the case of Chronos we normally use a 20tpi screw which means the screw must be rotated 20 times to move the carriage one inch.
    • Harmonic Drives offer an excellent way to reduce speed, by their nature they offer zero backlash, however thier price usually puts them outside the budget of a typical DIY timelapse rail, generally costing hundreds of dollars.
  • Movement Resolution is something to consider when selecting you gearing options, especially when using stepper motors. if you do not plan to do any motion ramping then gear reduction becomes less important.
    • Scenario 1: Stepper motor with 1.8 step angle, using a 1 inch diameter pulley on a belt configuration with no additional gearing.
      • 1.8 step angle = 200 steps per revolution.
      • 1 inch diameter pulley = 3.14 inches per revolution
      • 3.14 / 200 = .0157 inches per step
      • 200 / 3.14 = 69.69 steps per inch.
      • That is not a lot of resolution. at best you can expect maybe 30 moves per inch if you are lucky, and while you may not be interested in doing macro timelapse this will not be enough for smooth motion ramping either.  You can always use a motor driver with microstepping to help, but odds are unless you are doing studio work and have a very robust power source to drive the motor it is going to be too weak to do any sort of lifting movements and will only be able to handle horizontal.
    • Scenario 2 Geared stepper with 5:1 gear reduction using a 1 inch pulley
      • The gearing provides 1,000 steps per rotation.
      • The gearing also provides much more torque and a degree of gear-braking so if it loses its grip on a vertical lift the gear will drop much slower.
      • with a 3.14 circumference on the pully it offers a .00314 inchs resolution
      • 1000/3.14 = 318 steps per inch.
      • This is much better than scenario 1, now we are seeing the resolution that is needed for motion ramping. Plus you could achieve some decent macro resolution when microstepping with a resolution of 5,095 microsteps (x16) per inch.
      • This configuration should provide a little bit of lifting power, probably 7 or 8lbs on vertical movement depending on the strength of the motor.
    • Scenario 3. 1.8 degree stepper motor, using a lead-screw setup instead of a belt drive.
    • Using a 20 thread per inch lead-screw directly coupled to the motor.
      • 20 TPI lead screw means the motor must rotate 20 times to move one inch.
      • 200 steps per rotation puts this at 4,000 steps per inch in non-microstepping mode
      • This setup results in virtually no backlash.
      • In microstepping mode it will cut this out to 64,000 steps in one inch.
      • This gives it the ability to do motion ramping even at macro levels
      • The lead screw design is also extremely efficient, you can use a pretty low powered motor and still do vertical and horizontal movements with substantial weight limits. Chronos lite uses a small bipolar stepper motor and it can lift 25lbs vertically with this setup, where as with a belt drive it would have a hard time listing 2lbs.
      • This setup is also self brakeing, there is little worry that your camera could come crashing to the ground unless you do not properly attach the anti-backlash nut to the carriage.
      • It is very difficult to achieve high speeds with this setup.

 

I will keep updating this document as I have time, I hope this helps. This answers a lot of the DIY questions I have been asked.

The Chronos Project | Time Lapse Motion Control
©2013