While gimbals with great performance are available for the GoPro, it’s another story if you want to stabilise a bigger camera like a DSLR.  DJI makes the lovely Zenmuse Z15 family of products if you want to spend a fair bit of cash.  There are kit-form carbon-fibre gimbals which aren’t as nice as a Zenmuse but they are a bit cheaper…

There’s an alternative: buy some iPower 5108 120T motors, an Alexmos/Basecam board and some aluminium tube and do a maker project.   Aluminium has some great properties for a gimbal: cheap, light and strong, and it’s easy to work.  But it’s notoriously tricky to weld.  The first thing though, is to find the centre-of-gravity of the camera and make some rough sketches with dimensions that will allow the gimbal to balance neutrally with the camera installed.

With a rough sketch of the design it’s pretty easy to bend tubing into a simple frame using a miniature pipe bender.

Aluminium is a bit stiffer than copper tube, and it can slip through the pipe bender and kink, but using a small clamp to hold the tube in place allows easy and reliable bending.  It’s helpful to do a test bend or do some simple math for the circumference of a circle so it’s possible to plan the places to bend and the relation between the straight tube length and the dimensions of the finished part.

Welding or brazing aluminium is notoriously tricky but it can be done just using a MAPP gas torch and the right solder and flux materials.   There are some notes on brazing at the bottom of this page.

It’s possible to braze without using a jig or fixture so long as everything is held in place by some means.  On this project I made a couple of different prototypes for the pitch axis, eventually resulting in the below.

 The pitch axis has some extruded parts that create a sliding joint that provides for a way of getting the centre-of-gravity of the camera exactly aligned with the centre of rotation on the vertical axis.  This locks off with some cap screws.  The camera can mount on two rails that allow the alignment of the COG to align with the centre of rotation on the horizontal axis.  The design offsets the sliding joint extrusions from the centre of rotation so it can be easily adjusted once fully assembled. The extrusions are simply some modified DIY store hardware.

One of the tricky aspects of making the roll axis is to mount the roll axis motor in such a way as to avoid a huge ‘mechanical moment’ acting on the motor bearings.  I modified the iPower motor by driving out the 5mm motor shaft and pushing in a much longer shaft that allowed me to fit a frame right around the motor/roll axis assembly.  This meant purchasing a 5mm dowel pin, sanding it down slightly to 4.97mm dia to be a moderate interference fit in the motor outrunner, cutting a groove into it for the circlip, and putting it all back together.  It was pretty easy using only basic tools.

The modded/lengthened shaft of the roll axis motor is mounted in a bearing that’s visible in the above photo.

I mounted the Alexmos/Basecam ‘Simple BGC’ controller using some 16mm boom mount clamps that cost almost nothing via a Chinese website, and held the wiring in place with zip-ties.  Job done, and time to fit it to the airframe.

Tuning the SimpleBGC controller PIDs is similar to the explanation here:

https://www.youtube.com/watch?v=8_fAoGj5MWg

Here below is a video of this stabilising the camera:

This T960 was originally flying on 2x 4S LiPo, with all-up-weight somewhere around 4.5 kg.  With the Canon and gimbal fitted, 6S LiPo packs are needed.  With two 6S 5000 mAh LiPos and the 3S auxiliary battery for the gimbal, the AUW is 6.25 kg, and the prop size needs to come down to 15″ to avoid cooking the motors.

Aluminium Brazing – Notes

It can be tricky to join aluminium by welding or brazing.  Welding this type of small gimbal structure is really only possible with TIG (GTAW) using ac welding current.  Brazing is possible, or even relatively easy, but there some details that need to be right.  It’s also going to need a bit of patience and practice on scrap parts to get a workable approach.

The first thing is the metallurgy of the aluminium parts.  This is what welding engineers would call the ‘parent material’. Some aluminium alloys are classed as unweldable, and this can be because the heat input causes the metallurgical structure of the alloy to change in ways that fundamentally alter the mechanical properties.

There’s some extremely useful info here:

http://www.lincolnelectric.com/en-gb/support/welding-how-to/Pages/aluminum-design-mistakes-detail.aspx and also here:

http://www.lincolnelectric.com/en-us/support/welding-solutions/Pages/aluminum-faqs-detail.aspx

As a rough rule-of-thumb, any of the high-strength aircraft alloys are unweldable.  Any alloy that has been subjected to careful heat treatment during its manufacture is likely to be unweldable.  Generally it’s better to pick an alloy that isn’t a high strength alloy because it will be extremely low strength after heating it.  Moderate strength alloy may retain its moderate strength.  Some alloys ‘self heal’ after welding and some can be heat treated after welding to restore their properties.  But, for a hobby project you are never going to have spec sheets on the material, and you are not going to have access to carefully prepared welding procedures, so just buy some different types of alloy and do some tests with them to see what works.  Or use screwed joints where you need to.

The second thing is the type of solder and flux you use.  There are basically three main types.  The first is the ‘friction brazing’ solder which is fluxless.  This is hideous to use for brazing, really it’s a filler material that can be used to fill holes.  Then there are both low temperature and high temperature brazing materials that work with specialised fluxes.

Aluminium has an oxide coating that has a much higher melting point than the alloy itself.  The job of the flux is to chemically remove the oxide, because if you try to melt it the underlying aluminium will melt first… messy.

Personally I had quite a bit of success with the following:

High Temperature:

• Fontargen A 407 L rods with F 400 MD flux

Low Temperature:

• Fontargen AF 631 NH which contains a flux core

I somewhat prefer the low temperature solder because there’s a wider window between the melting point of the solder and the melting point of the parent material, which gives a helpful margin when you don’t have the skills of a professional welder.

The third thing is to break the oxide coating near the weld using a file, and to absolutely eliminate any anodizing from the weld area since chemical fluxes can’t remove anodizing.  Although chemical fluxes will remove oxide to an extent, excessive oxide doesn’t help and, being heavy, can sink into the molten metal.

The fourth thing is to get the parts jigged or held together stably.  It’s all going to go wrong if you get the parts up to 380 degrees C or 650 degrees C, and then they start to move around.

Then controlling the heat input is quite vital.  Both parts at the joint need to reach the same temperature.  Aluminium is an amazingly good heat conductor, and virtually the entire structure will get to the same temperature, so it needs a lot more heat input on the bigger part.  But too much and it will melt. Once enough heat has been put in to melt the first bit of solder, the rest of the joint will likely be already near the ideal temperature so further heat input needs to be reduced.

The rest is pretty much the same as any brazing or soldering.  Feed in just enough solder to wet the joint.  The solders I mentioned are extremely good at wetting the joint and will flow very quickly.

Practice is needed.  Don’t mess up parts you have carefully crafted by using them as test pieces for your first brazing attempts.