Radio Controlled Cable Dolly for Small Format Cameras

Using some relatively simple construction techniques and off the shelf RC parts, you can make a small, fast, reliable cable dolly to carry a small format camera, and get you footage like this:



and this:


At lower speeds, it suffers a little bit of wobbling. Having a second person would eradicate some of this by enabling me to pull the line tighter; the rest, I can design out with the next version. I'm quite proud of having successfully sent it over a cliff! There are many problems I've already solved with various prototypes.

This instructable won't just show you how to make a cable dolly, it'll show you my research, errors and process too. By the end of it, you'll be able to anticipate common problems, and design one to work with whatever hardware you choose. There are many possible upgrades, which I talk about a little in step twelve. For now, I'm sticking with a simple cable dolly, with variable speed forward and reverse control.

After tools, materials and design, steps 5 - 9 will cover enough for you to make an unpowered gravity dolly, then we'll go into motors and radio control. Most prototypes of this dolly were made by hand with aluminium. The DXF used to make the CNC milled version is attached in step 5.

A cable dolly, even a powered one, is quite a simple thing to assemble. Getting all of the little things right, designing and tweaking it to work well is the complex part. You'll probably run into all kinds of problems such as undesirable swinging, not enough wire tension, and insufficient contact between the drive wheel and the line. Don't be discouraged, you can iron these out.

I'm going to show you construction of my first ugly prototype, then take everything I learned to make a neat looking CNCed and improved version. In some steps, I'll be doing more than I really had to. This is either because parts were not available, or it was late, hardware stores had closed and I had to work with what I had to hand.

This instructable is focussed mostly on making the dolly with hand tools, since they're what more people have access to, but I did use some bigger machines to better the results and speed up the process. You can cut out a whole load more of those tools and steps if you have access to a three axis CNC mill.


Tools:

• Sharpie (for marking metal).
• Rubbing alcohol/Nail polish remover (for removing sharpie marks from metal).
• Battery drill.
• Hacksaw.
• Bench vice.
• Speed clamps.
• Metal files.
• Pair of adjustable spanners.
• Molegrips.
• Metric tap and die set (if making or adapting your own axles for pulley wheels).
• Soldering iron.
• Third hand tool.
• Wire snips.
• Wire strippers.
• Calliper (For precise metalwork, and determining your motor's footprint if diagrams aren't readily available)

Optional:

• Drill press/Pillar drill: Much more precise than than drilling by hand.
• Grinding wheel: If you cut down steel bolts, you could just file them, but this with the bolt held in molegrips is much faster and give a better finish. n.b. don't use this for aluminium, it'll clog up the wheel surface).
• De-burring tool: Easy to use, makes your metalwork much nicer.
• Benchtop belt sander: With the right belt, this can neaten up your cuts into aluminium no end.
• Lathe: For making wheels. You might be able to find stock ones you're happy with.

I used all of these things, and they made the job a lot easier.



Materials and Components:

(Not all shown in photos)

• Aluminium strip or polycarbonate sheet (4mm thick seems about right).
• L bracket and/or aluminium strip (to hold tripod head below dolly)
• Pulley wheels with axles.
  • Alternative: Nylon rod + lathe
  • Alternative: Inline skate wheels + lathe (This is what I went for).
• M3, M4, M6, M8 machine screws/bolts, washers, spacers and nuts (depending on the size and types of your wheels and motor fittings).
• 10mm aluminium square rod (Will vary according to the motor/mount you get).
• Threadlock
• Rapstrap (Soft, stretchy zip tie stuff. Not necessarily easy to find, but usually on ebay. Good for attaching stuff to the dolly temporarily).
• Compact tripod ball head (I favour the Joby BH1: Well built, compact, sometimes available for as low as £20, has quick release plates, and secure enough for small format cameras in most circumstances).
• 3/8"- 16 UNC bolt (To fasten tripod head to dolly).
• PVC clothesline (steel core), steel wire rope, or 5mm static line.
• Ratchet strap (For tensioning the line at each end).
• RC transmitter.
• RC receiver.
• Power source (LiPO or NiMH battery).
• Charger for battery.
• Electronic Speed Controller (ESC) with battery elimination circuit.
• Deans Connector if using LiPO and your ESC didn't come with one.
• Motor and gearbox (370 motors and gears are inexpensive and can shift relatively heavy loads).

Optional:

• Nylon bolts, nuts and spacers (For holding the body of the dolly together if using standard pulleys, extra fixings also give rigidity to polycarbonate body if you go that route. They don't necessarily have to be nylon, but nylon fixings are nice to work with by hand).
• Cable ferrules and thimbles (Vital if you're using steel cable).
• Sugru (If you're using a metal drive wheel and want to give it extra grip, though it's not made for this kind of application: higher powers will wear it away).

Regardless of RC parts, the considerations in this step apply powered and unpowered dollies alike.

I thought about various materials for the body, and settled on two as best: Aluminium, and polycarbonate. They're both very light, easy to machine by hand, are easily obtainable, and are also cheap. I'll show how to make an aluminium dolly by hand, then also use a CNC router to make a polycarbonate body.

Beyond materials, the major consideration is stability. Without it, all of your footage will be made rubbish by the camera swinging around. There are a few approaches to making your dolly more stable. One is to run two parallel cables instead of just one, then hang the camera between them. This way, you have to worry a bit less about balancing the dolly. With this method, it may also be practical to modify an existing RC car to run along the lines, but raising, tensioning and securing two lines is significantly more hassle than securing one along a long route. This design by BackyardFX shows how to, but it's bulky and seems like it would work better for shoots in smaller areas.

Most designs I've found, and certainly all the ones chasing mountain bikers, etc., are single line, which reduces rigging time and enables longer cable runs. With a smooth start, gradual acceleration, and proper weighting of the dolly, one line should be fine for the kind of footage I'm aiming to shoot, so I'm going with that sort of design.

Another approach to stabilisation is to use gyroscopes or a flywheel. Some dollys use a weighted wheel for balancing; for instance, you can see one on the rear of this dolly in the first shots of it in motion (Note, this dolly was not made by me, nor this incredible footage shot by me):


I'm not sure how effective this approach is though. It seems the footage was all shot at different times with various modifications to the dolly. Some of it is amazing, but you can clearly see the camera rocking from side to side in other shots. If you wanted to be sure of stabilising one, two powered gyroscopes mounted at 90 degrees to each other would do the job well, but probably have a decidedly adverse effect on battery life.

Another approach is to use more wheels, not in contact with the cable, to create stability. TShed have done just with their professional cable dolly, which uses bike wheels on the sides to balance it:
http://www.tshed.co.uk/1/page.php?17

The results it gives are superb (seen in many BBC nature and anthropology documentaries), but it significantly increases the size. You can see it in operation in this clip. Details on it are scant; it may be that the wheels are geared to spin faster than the one rolling along the line, or it may just be that their larger diameter and greater mass creates a strong stabilising force, even spinning at the same rate as the main central wheel.

There's a lot I could do, but I'm aiming for something very compact and easy to carry, so I'm just going to try weighting and controlling this dolly well.

Body construction will consist of two long, thin, flat pieces running parallel to each other, either side of the cable. The RC parts will bolt and tie onto the frame sides, and the wheels and cable will run between them. It may also need some weights to keep the body balanced on either side of the line.

An important consideration in terms of balance is keeping as much of the weight as possible below the cable. As the cable is what the dolly will pivot around in the event of unwelcome noise, we need to dampen that by lowering the centre of gravity. Extending the camera arm is one way to do this, and the easiest. Your dolly will probably become less compact than you imagined before you get footage as stable as you'd like.

Wheel positioning is also important. The guide wheels (blue in the diagram) sit above the cable, should ideally run freely on bearings, and for stability should be as far toward the top of your dolly as possible (Some dollies have the drive wheel on top and the guides below, such as TShed's). Some designs put the wheels on special hangers above the body for this reason, for instance this excellent commercial prototype:
http://nastynick42.pinkbike.com/album/RC-Zipline-Prototypes/

In my shoddy diagram, the red wheel is the drive/tension wheel, and runs below the cable. Raising it to deflect the cable upward does two things: In the case of a gravity dolly, it will act as a brake. The more tension you give it, the slower your dolly will roll down a given slope. With too much, it won't roll down the line at all. In the case of a motor powered dolly, the red wheel is also your drive wheel. Raising it creates a larger contact area with the line, increasing traction.

A significant upgrade to a gravity dolly would be to cut a slot for the tension wheel axle, allowing you to exert some control over descent speed. Another approach would be an adjustable drag brake. I didn't bother with either, because I knew I was going to make this one motor driven. Speaking of which…

(If you're only aiming to build a gravity powered dolly, this section is irrelevant for you).

When I first thought of this project, I thought it would be good to recycle an old toy for the RC components. Wrong! Looking at the scrap toys I had to hand, I found only slow, small, weak motors. Certainly not powerful enough to drive a dolly along a wire. Also, most toys use infrared comms rather than radio, which is no good for long ranges outdoors, and the few that did use radio had ranges of less than 20 feet.

I thought of a few other approaches:

1. Use an arduino, motor shield, and RF module to control a motor mounted on the cable dolly. Possible, but a lot of work and probably not cheap enough to justify doing that work myself. It has potential advantages in terms of software making the dolly behave intelligently and depend less on a human operator, but good comms range between the transmitter and the dolly are not guaranteed. I also really disliked the idea of working with a laptop on sunny days in the woods.
2. Rip apart a slightly better, but stil cheap, RC toy: I decided against this because it would involve a lot of messing around to adapt an RC car to work with a single cable and, once completed, it might still not work how I wanted it to. Parallel cable make for a much more stable dolly, and might work very well with an adapted RC car, but making such cable runs involves a lot more hassle that single lines, both in setup and hefting gear around. Also, there's little point buying something new to rip apart when:
3. There are off the shelf RC components that you can combine to spec for your project. This gave me the most bang for buck: Guaranteed comms range, lots of work done on hardening against interference, a powerful motor, plug in units that I know will be able to control it, high capacity batteries, and a familiar, standard control method. Far East production and the amount of people playing with RC vehicles has driven component prices right down in recent years, so they're by far the best option. Let's take a look at some.

If you want to look at this subject in far more detail, teamtestbot has posted an excellent guide to RC parts here:
https://www.instructables.com/id/The-New-and-Improved-Brushless-Electric-Scooter-Po/

Electric scooters are a much tougher engineering challenge than something like this half-kilo dolly, but his point about overspeccing parts is a very good one: When specifying your own, better to build extra capacity into it than have something burn out or start a fire.


Motor:
The motor has to be big enough to drive about half a kilo up and down a line. There are many, many motors available, and the source most RC people direct me to is Hobby King (Warning: at time of writing, has autoplaying video on home page). What we're doing more akin to an RC truck than an aircraft. After quite a while spent looking online, my local model shop directed me straight to a cheap motor with a gearbox already attached, telling me it would handle a few kilos with ease.

You can get much higher torque from a brushless outrunner, if you're willing to construct a slightly more complex drivetrain. I opted for a brushed motor: It can cope better with lateral loads, is simpler to mount, and is less susceptible to bits of stuff getting into the motor casing than an outrunner. Since I'll be using this outdoors, that counts for something over a little more torque.


Transmitter:
I'm using a four channel Futaba because it's what I had available. With no budget restrictions, I'd probably get a Spektrum DX7 or DX6i, as the receivers are smaller, and extra channels mean they're good for other, much more complex RC projects too. For now though, this dolly is only going to have forward and back speed controls, so a four channel transmitter is more than enough.


Receiver:
This is matched to the transmitter, and has channel outputs for both the main motor (throttle) and servo motors. Servos are usually for control surfaces on model planes or steering on RC cars and boats, but if we wanted to get really fancy with this dolly, we could get some servos and use them for panning and tilting the camera. Let's keep it simple for now.


Electronic Speed Controller:
I recommend a marine ESC, as it will probably have a bigger heat sink. Also, what's not to like about waterproof components?

A couple of other considerations for your ESC:

• Make sure it's matched to your motor type: brushed or brushless.
• Verify it can do forward and reverse control of your motor. Not all can.
• Get one with a battery elimination circuit (BEC).
• Once you have it, calibrate it to make sure you're geting the full range of power.

The BEC in particular keeps your electrics simpler and lighter, by enabling you to power them all from the same battery. Without a BEC, you need one battery for your motor and one for your receiver (and servos, if you're including any of those). Buying an ESC with a BEC is a little more expensive, but it greatly simplifies your build as well as halving the number of batteries you need to charge, maintain and carry. Well worth it.

Also, you may need to replace the connector on the ESC, depending on it and your batteries. Tamiya connectors are said to melt when used with LiPO batteries, but Dean's connectors have a good reputation among RC enthusiasts, and putting one on is a quick and easy soldering job (detailed in step 9).


LiPO batteries and chargers:

Get the right voltage battery for your motor. I'm using a 7.4v, 2 cell, 2100 MAh LiPO.

DANGER: LiPOs are not to be toyed with. They're soft, and shouldn't be dented, pierced, or subjected to impacts. If mistreated, the resulting explosions and fires can be spectacular. Ideally, for high speed applications you should use a hard-cased LiPO battery rather than a standard soft one.

Never leave LiPOs to charge unattended. To be extra-safe, use a fireproof charging bag, which should be available anywhere that also sells the batteries. Get a balancing charger too, and make sure any LiPOs you get have balancing plugs (My charger: EK2-0851. It's small, lightweight, and inexpensive). While balancing and a fireproof bag do make charging safer, they don't eliminate the risk of fire. It just means that if something does go wrong, you'll have a much better chance of getting it away from anything flammable.

Why LiPOs? Simply because they're about the cheapest yet most energy dense battery out there. This is the reason they're used for RC aircraft. It may be that you could use something safer like NiMH, and your local model shop might advise you so. A part of the reason I went LiPO was to reduce weight (which will count for everything from performance of the dolly and strain on the motor to how comfortable days spent filming in the woods and carrying gear are), and another factor is that I want to start building up gear that will be useful for other projects. A different battery option may be the best for you.

I tried a few different approaches during this project, some of which failed at various stages. These photos and videos explain.

Failure one: Expecting to be able to salvage something from another gadget that would work. I wasted several hours on this, and by then it was obvious all the garbage I had at my disposal was too underpowered. It was far better to go to a model shop and get equipment I knew was up to spec before I started.

Failure two: Attempting to use unmodified inline skate wheels, which put far too much weight up top and made the dolly too unstable. It swang far too easily to get good footage. This approach might work with thicker cables and heavier cameras, or perhaps just a very heavy camera hanger, but there's no way it can be suited to the small format cameras I'm aiming to use without drastic over-engineering.

(Partial) Failure three: Metal pulley wheels on metal axles. Cleaned up and lubricated, they ran, but nowhere near as well as I wanted.

I fixed both of these issues by hacking apart skate wheels then turning them down on a lathe, making a set of small, single bearing wheels with grooves in. You can see that done in step 6.

Failure four: Camera arm too short, resulting in lots of lateral wobble. Having a longer, heavier arm helps to compensate for this by lowering the centre of gravity.

Failure five: Paracord. A bare metal pulley wheel for drive ran fine on PVC clothesline, but slipped on paracord, burning it in the process. Attempted fix with heatshrink (pictured), but it was too uneven. Attempted another fix with Sugru, but the paracord wore through it in seconds when the motor was revved right up.

Failure six: A loosely hanging camera arm. I thought this might work and be self-levelling, but instead, it introduced line noise whilst failing to level. I found a rigid camera mount was much better.

By far the most difficult aspect of this build was determining the right wheels for it to run on. Not many applications call for recessed, gripping pulley wheels on bearings, so they're not the kind of thing you can pick up easily. While researching cable dollies, I found the best performing ones tended to be either large and made from bicycle wheels, or be small but have custom turned wheels. The latter is what I ended up doing myself. You could make a dolly run on stock hardware, but it might not be easy and it won't perform quite as well. For descending cable runs and gravity dollies, pulley wheels seem to be fine.

Time to build! I've added a DXF file for my final design to this step, which you can either print out and stick on your metal to fabricate identical sides by hand, or use with a CNC mill if you have one. The curvy shape doesn't matter except to save a little weight; the same hole pattern in a couple of normal strips of material will do the job.

If improvising your own design of gravity dolly, you only need to make four holes in each side: One pair for each of the three wheels, and one for the camera arm to mount to.

By hand:
Cut a strip of 4mm aluminium into two, then mark all of the holes to make on one side. Clamp both parts together, then drill it out and deburr the holes. Beware if you're more used to wood: For metal, work your way up from smaller drills to larger ones in increments of a few millimeters at most.

I used a belt sander to neaten up the ends of my cuts, and a deburring tool to neaten up the holes. These things make all the difference to your metalwork, and also mean you won't end up with a million tiny cuts on your hands just from handling it. A deburring tool is a cheap thing; here's a quick video showing use of one:


The tool has a curved cutting edge. Starting with the part of the edge parallel to the side of the hole, you simply work it around and gradually push it further in, removing any burr and creating a nice chamfered edge if you want to push it in that far.

In the photos here and subsequent steps, you'll see a couple of aluminium plates slowly starting to look like swiss cheese. This is because of all my experiments and failed attempts. I swapped wheels out a lot, then finally ended up tweaking the length and position of things before committing it to CAD and making the final version.

By CNC:
Not many people have access to a CNC mill (I only do because we have one on loan to Nottingham Hackspace). There's a DXF of my final design attached to this step. It's designed to only require 2.5D cutting, and a maximum bit diameter of 3mm. It will fit in the bed of a 6040 hobbyist CNC mill. Set up your machine and material, make some G-Code, and mill away, you lucky so-and-so.

I was hoping that I'd be able to get some standard wheels to run this with, but that turned out to be very difficult. I happened to have some spare skate wheels, so sawed the excess off, then turned them down on a lathe, putting grooves in them for the cable then parting them. Each wheel has one bearing in. Much better than a standard pulley!

To quickly test it rolls, assemble the dolly with just the top wheels. Use an M8 nut and bolt, with a small washer placed either side of the wheel to clamp onto the inner part of the bearing and make sure the body doesn't foul the wheel or outer bearing. Stretch out a small piece of line and test to make sure it will work.

If there was one thing I'd do differently for this project, it would be to buy some nylon round bar and make my own wheels from scratch. It wouldn't have been any more work than turning down these inline skate wheels, I could have made them slightly bigger, and I could put a much better profiled groove in them. Next time. For now, this is more about what you can do with fairly standard hardware. While I got pretty fancy with those skate wheels, well mounted and lubricated pulleys would do the job.

At this stage, you have to make a compromise between compactness and stability. When the dolly is moving, it will be able to swing from side to side on the line. The closer the camera is to the top of the line, the more obvious that will be in your footage.

With a longer hanger, the centre of gravity will be lowered and the motion will become both smaller and less obvious on film. Make it too low though, and you might struggle to get exactly the shot you want, or get the dolly through gaps and over obstacles.

I tried a short hanger first, found it too unstable, then added about a foot of thick aluminium strip, creating a marked improvement in performance. In conjunction with the tripod head, this also had the added benefit of making the camera position far more adjustable.

For this stage, I'd advise a couple of ratchet straps, and one of the following:

Steel core clothesline (Available just about anywhere in lengths from 10 - 100m. You might struggle to find more than 20m lengths locally, but you can buy longer ones online).
4mm static line (Available from any shop that stocks climbing equipment).
Steel wire rope and appropriate fittings.

Some of the examples here use paracord to show techniques applicable to static line, but I wouldn't recommend paracord for running dollies on at all. It's difficult to knot, quite stretchy, and whatever you use for a drive wheel on a motorised dolly, it will probably slip and create a lot of heat on it. I was hoping it would be a cheap way to do long cable runs, but alas, no.

PVC coated steel core clothes line is strong, fairly cheap, and can maintain a lot of friction with your drive wheel if you're motorising your dolly. It's a pretty good, readily available compromise. It's also available in various colours if you're concerned about safety while filming (For instance, a red clothesline would the easiest to spot for snowboarders on piste or mountain bikers in the woods, but without extensive safety procedures, you almost certainly shouldn't be running lines where anyone might collide with them).

Of course, if you can budget for it, you can go for proper high tension steel wire rope and specific fittings manufactured for it. Research this well and use it carefully if you're going to: Steel cables are *incredibly* dangerous if they break or come free under tension.

One of the biggest things that can make this step easier is getting help from one or a couple of people, which lets you work faster as well as tension the line more. In the photos, there's a method for tensioning cord/rope with a secondary line and ratchet straps. Alone, doing this is awkward at best, and you'll get a much slacker line than you would with help.

With steel cable and steel core clothes line, you can't just tie a loop in the end; you need to terminate it properly. I made some fittings for it out of bits I had to hand, but since the fittings are wider than the dolly body, setting it up requires removing the fittings from one end, feeding the line through, then replacing the cable termination (still less hassle than dismantling the dolly and reassembling it around the line). Steel cable terminations are also commercially available, and a must if you're buying steel wire rope.

With clothesline, over the course of a few hours you'll find that it sags and stretches. Your cable terminations will also dent, and may sever the casing. PVC+steel clothesline is very cheap, but I certainly wouldn't regard it as very reusable. At least, it becomes shorter each time I use it, with the crimped parts trimmed off after use.

I ran a 20m line about four times before it started to show signs of failing. That may seem like a short lifespan, but it did only cost £2 though, and it can be acquired in lengths of 50 - 100 metres too. Longer lines are a bit more expensive per metre, and may fail faster under their own greater weight. One advantage of clothesline is that when it does fail, it tends to just fall or sag rather than flying free and whipping into nearby things. It can't take as much tension as proper wire rope, but that makes it safer in some ways, and for the weight of this type of dolly and a small camera, I've found it's absolutely fine.

Another advantage is that it will almost always fail at the ends, shortening your line slightly rather than cutting it in two. This due to the properties of a line under load: It's under much greater strain at the ends than the middle. Think of it as a series of weights strung together, and that should start to make sense. The line sags in the middle to some degree no matter how much it's tensioned, so the parts further toward the ends are supporting a greater weight than any piece of line near the centre.

Proper cable setup makes a big difference to your footage. For a gravity dolly, you need to make sure the downward cable termination sits below the lowest sag point on the line as the dolly hits it. That way, it will run all the way down, and probably decelerate as it approaches the end unless your cable run is steep.

For level spans with an RC dolly, you'll find that as a result of sag, the speed changes at the halfway point because the dolly switches from descending to climbing, putting the motor under a greater load. Putting just a little bit of an incline on the cable to compensate for the sag can put a closer-to-continuous load on the motor, making it much easier to control smoothly.

If all you want is a gravity dolly, you can stop here. I'm going to push on and make this one radio controlled though.

UPDATE: Static line is much better than washing line in terms of traction, and 4mm line is available for about £0.7 per metre. For spans of more than 20m it's more economical than steel core clothesline, and it comes in lengths up to 100m (Clothesline, the maximum is around 30m). It's easier to work with because you can use knots rather than hardware terminations, though it is quite stretchy. My experience so far is that after tensioning it by myself, with the dolly hanging at the centre it will sag about one metre per thirty metres of span. With a friend, I'd be able to get more tension into the line, because they could stretch it while I take care of terminating it. In terms of loops, I've tried a few more and quite like the perfection loop.

This will be a quick test of the RC gear to make sure it's all working before we attach it to the dolly, but before that, I need to make a small modification. Many RC power controllers come with Tamiya Connectors, which are reputed to melt under high loads. Best to replace it with a Deans connector, as shown in the photos.

Apologies if the little soldering how-to is teaching you to suck eggs. There's also a diagram of the RC parts and how they connect to each other, which will be deeply patronising if you've ever dealt with such before.

Once that's done and you're hooking the components up, changing the connection between the ESC and the receiver lets you select which control will adjust the power to the motor. By switching the connection between 1 - 4 on these Futaba bits, I can choose left-right or up-down on either stick. I chose one which works with up-down on the right stick, because that one has a spring return and gives me better feeling analogue control of the motor speed. The left stick on this transmitter has steps and doesn't return to the neutral position automatically. Under some circumstances, that might be useful for setting a constant speed.

I'd advise always switching off your receiver before your transmitter. It may not be the case with every brand, but with my Futaba components, even after calibrating the ESC, I quickly found that switching things off the other way round creates a surge of the motor. Not what I'll want when the dolly is parked at one end of the cable.

Mounting motors can be tricky. The one I had is designed to mount on a piece of 10mm square bar, as well as having a couple of holes in the front of the motor. It required a bit of modification by rethreading the axle, plus the fabrication and some spacers and brackets, as shown. The result was a very solid three-point motor mount, and a properly aligned drive wheel.

The pictures show my method from start to finish, you may need to adapt depending on the size and type of motor you get. The one I got is a brushed 370 motor. There are a lot of different gearbox housings for them, some of which are simpler to mount than the one shown here.

If you have to retap a shaft to make your drive wheel align properly, you may also need a second person to help; I did.

The images of this step show several different configurations for the RC components, which I found best to mount on the camera arm. They also show the final, CNCed body I made from polycarbonate.

A few things are worth considering here.

Temporary fixings: Test some stuff out with non-permanent fixings. I used stuff called rapstrap, which is like a rubbery reusable zip tie. Especially good for tying down the LiPO without risk of denting it, as long as you're careful and don't overtighten (nylon zip ties should definitely not be used for securing LiPOs).

Weighting: If possible, the RC components should balance the dolly. In my final configuration, I opted to put the battery and power controller on the camera hanger to help a little with stability. Putting them there also made them more centralised.

Wiring: At all stages, but *especially* when doing initial tests with parts strapped rather than properly mounted, you should make sure that none of your wiring will foul on any other moving parts.

Antenna: Just in case of the body interfering with radio reception, I ran the antenna wire from the receiver around both sides of the body and camera hanger.

Using the methods from step 9, tension a line, power it up and try it out!

Here are my first tests, done at various stages of prototyping. The camera arm was too short in the first set, and in the second I found out paracord was a rubbish kind of line to run powered things on:

When it comes to filming with small format cameras, a very large amount of the quality is in the setup. Many things come from practice and incremental improvement. I'll share what I know and that seems relevant here.



Lenses:
If using small format cameras such as phones or pocket camcorders, lenses are actually one of the easier and cheaper parts of this project. These devices usually have a very narrow field of view, designed for capturing social situations. You may find yours works for what you want to do, but if you want wider angles, extra optics are required. Here are some small, cheap, readily available lenses that are designed for phones and attach via magnetic rings:

2x Telephoto: Awful lens. Rattles, has blurry edges, and tends to vignette.

0.67x Wide: The smallest of the lenses. Very cheap, and gives a good, non-fisheye wide angle for phones (tried with HTC One X and iPhone 4). With Flip cameras though, it makes too small a difference to really be noticeable. You can unscrew the front to get a tiny magnifier/macro lens.

0.5x Wide: This one can be harder to find than the others, and is a little more expensive than most, but I think it's the nicest of them all. It's the only one with a concave front element, and it strikes a nice balance between extra field of view and minimal distortion. Not quite right for POV action footage, but great in lots of other situations and would be fine on a dolly.

FE12: Also listed as a 180 degree fish eye, this will do just that on most phones. On a Flip, this will give you around 150 - 160 degrees of viewing angle. Not bad.

FE18: The most expensive of them, but that's still only about £15/US$25. Also listed as a 185 degree fisheye. Converts things like flip cameras to around a 170 degree viewing angle, which makes it ideal for POV sports footage, but grotesquely distorted for most other stuff. Tends to cause vignetting.

There are many more such lenses available, but the latter four of those are the best I've found in terms of budget and performance. Camera geometry will mean you get varying results with all of these lenses when coupled with different devices. Usually, small camcorders and phones have a glass dust cover, with the actual tiny lens and CCD set back from it. The greater that distance is, the less of the viewing angle of your add on lens will be utilised by the camera. Flip cameras are about the worst for this with 6mm between the two. Phones usually have the lens and dust cover much closer to each other.


Shooting With Extra Lenses:
Wide angle lenses make chase shots look great. Filmed with normal focal lengths and viewing angles, the same shots look decidedly pedestrian, because the edges of the frame sit further along the path and exclude nearby context. To an extent, the wider viewing angle simulates peripheral vision for the viewer: Things leave the sides of the shot more or less as they're passing the camera, creating a greater feeling of speed and making the viewer feel not only as if the camera is moving faster, but also more like they're there.

Wider angles are also easier to frame things with, and more forgiving of small camera movements. Camera shake is less noticeable because it's effectively scaled down, and in the case of action sports, when the subject is moving around a lot, a wider angle helps to keep them in frame more consistently.

(Note: The test videos below were shot with a Flip camera, and the corresponding reduced viewing angles you'd expect compared to phones).

Here's a comparison of chase shots with various lenses:


For sideways pointing tracking shots, the compression of wide angles may not be the prettiest effect. Here's a pan with a 0.5x wide lens, which I quite like:


Here's the same pan, but with FE12 and FE18 fisheyes:


The barrel distortion of a wider angle forced into a rectangular frame can look distracting and ugly if used badly. Also, particularly when shooting action from an external perspective, by compressing more of the scene into the frame a wide lens can make things look more remote to the viewer, and hence a lot less dramatic.

Whatever lens you choose, you'll be making some compromise between ease of framing and noise compensation on one hand, and distortion on the other.



Tips on Filming With a Cable Dolly:

Using a cable dolly to film is potentially dangerous. You should be careful, and anyone you're with should also be aware of the risks.

• When picking a line, try out a few shots by hand roughly from where you think the camera will be. Better to make sure of your shot in advance, rather than run cable and find out your footage is all taken too close to the subject, etc.
• Account for cable sag. If you're not an engineer, just like I'm not, experience will tell you what kind of sagging you're likely to get over a given span. Raise your anchor points accordingly.
• If you're working with performers of some kind, i.e. downhill bikers, BMXers, etc., walk the line with them and make sure they understand where the cable is. Make sure they're happy with the line and it doesn't pose undue risks to them (i.e., will they get it in the neck when they go over that double? Is there a chance of the camera hitting them?).
• Likewise, wherever you are, run your lines such that they aren't a risk to other people using the trails/park/piste/etc.
• Never tension damaged lines, regardless of type. Even something like paracord could hurt you if it breaks under tension.
• Unless you can run cables safely high up off the ground (and are competent plus equipped to work at height if you need to), never cross paths, trails or roads with your cable runs.
• Don't run cables where you're likely to annoy people or otherwise get negative attention. For instance, it's doubtful you could run cables at a ski resort without getting permission from the company that maintains the pistes.
• If your lines have to be low, only run an RC dolly within sight of the operator. If you let it run out of sight, for all you know it's beheading a dog walker or something.
• Don't attach cable to weak anchors. That gazebo pole, the sapling, and the rusty iron fence? Almost certainly not.
• Don't hurt trees: If your straps aren't stretchy, use foam padding around the back of your cable and straps to prevent them eating away at tree bark. Even when not necessary, it gives you a greater appearance to onlookers of acting responsibility.
• Wait for your cable to settle before starting a run. Starting too soon will just result in a lot of bouncing and swinging.
• If you have to move the dolly by hand, poke and push it with a stick rather than grabbing it. Grabbing will create more line noise, then subsequently more thumb twiddling and less footage.
• With an RC dolly, don't let it hit the ends of your cable run. It will likely derail the dolly, which is awkward at height, and also results in more time spent waiting for cables to settle.
• A hook on a pole will come in handy if the dolly goes beyond arms reach.

I absolutely know a better dolly than mine is possible. Body design, weighting, wheels, line materials and tension could be tweaked a lot more to optimise. There are also many obvious upgrades, but this is largely bike-shedding: wiring recesses, hard-cased LiPO for extra safety, pan and tilt servos, gyro-stabilisation, RF cameras to give a first person view while filming, onboard semi-automated control, sprung buffers, limit switches, etc. etc. etc. As it is, the major work on the dolly is done, and it does most of what I want it to.

I'll update if I learn anything major, and will also update this with more footage as I go out and experiment with it.

Now, go out and get some excellent shots!