Prototype Leg Structural Assembly


Yesterday we finished structural assembly of the prototype leg. It looks awesome, so we’ll start with some pictures and then tell you everything that went in to it.

“But doesn’t this look just like your last blog post? What’s changed?” I’m so glad you asked. In the last blog post we had welded together the plates we got back from the water-jet cutters and tied them together with ratchet straps and clamps to get them in to position. The keen observer will note that in the last blog post, the weight of the leg is supported by a crane.

There is much more to these leg pieces than the waterjet plates. Every rotating axis has a shaft going through it and reinforcing bosses that need to be welded in to the structural plates. A reinforcement boss is a machined part that distributes the load from the shaft to the plate over a wider area. There are three axes of rotation on this leg, each of which has three pins that need to rotate (the joint itself and the two sides of the actuator). Each of those pins needs between 2 and 4 bosses machined and welded in to the structure. That’s dozens of machined parts, which are quite time consuming to make.


Another non-obvious but very difficult and important process is bushing alignment. Every rotating shaft on this leg has to go through a set of several bushings. For the shaft to turn, those bushings all have to line up perfectly. When we install the bushings in our structure, they are not well aligned. To line them back up, we need to ream them all. Reaming is the process of taking a special drill bit just slightly larger than the hole and shaving a few thousandths of an inch off the internal surface of the bushing. Hand-reaming a bushing is a slow process that takes multiple passes, adjusting the size of the reamer with each pass. Directionally loading the reamer is another trick we use to force the bushings in to alignment faster. There were 30-ish holes on this leg requiring a multi-pass hand reaming process. Phew!


We’ve also been doing a lot of experimentation with our weld practices. For our heavier plates (1/4 inch to 3/8 inch steel) we were originally using only a TIG process, which is quite slow. We’ve improved our MIG procedure enough that we have almost entirely phased out the TIG. For those unfamiliar with welding – TIG is a precise manual process where the operator controls the torch in one hand and feeds in metal with the other. MIG is a one handed operation where the torch automatically feeds metal in to the weld. TIG is generally more precise, but MIG is much faster and easier on the operator.

Our order of operations will also be totally different for our next leg. Welding the water-jet plates together before putting in bosses made for some very awkward grinding and welding angles, resulting in some lower-quality welds that we can get away with on this prototype but want to avoid when this machine actually has a person on it.

Next up – we need to put all the actuators in place on this leg and hydraulically plumb them. We also need to improve our hydraulic powerplant – we are investigating running our propane engine inside and routing the exhaust out but if the building management is not OK with that we will have to come up with something else. Thanks for supporting open robotics, over and out!


14 thoughts on “Prototype Leg Structural Assembly”

  1. Wow! Looks great!

    If I’m being honest with myself, seeing a human next to such a large mechanical device always makes my inner pessimist a bit uneasy. It’s just so simple to imagine worst case scenarios. I was wondering, what kind of non-human based tests do you have planned for when it all starts to come together? Perhaps this is published somewhere and I missed it. Will the entire device be driveable without an operator present in the “cockpit”?

    Keep up the good work! Can’t wait to see a bunch more of these legs!

    1. Thanks Chris! You’re right to advise caution around this machine. We call it a “robot” and not just a “vehicle” for a reason – all the motion is coordinated by a computer. The computer takes input from a variety of input devices – joysticks, buttons – all of which can be far away. So when we’re testing and debugging this leg and vehicle, don’t worry, all humans will be a safe distance away.

  2. That looks good from here in Kansas! You may want to make a jig for the leg pins as you only have five more legs to go. Jeff

  3. Hey guys – looking great!

    Just a few backseat driver notes from a fellow mechanical engineer (and project supporter).

    1) That’s great you moved to MIG instead of TIG. MIG isn’t just faster & easier – it’s far more consistent because of the simplified controls, producing more reliable welds with less distortion. As a result of that consistency – it’s also much safer once you’ve verified your weld parameters.

    You can further reduce distortion by welding while constraining the piece to something big and flat (like a welding table) and apply your stitches in the same direction on either side of the plates.

    2) You guys are really making it tough on yourselves by welding in those bushing bosses. I expect misalignment will also get worse if you weld them into the plates before assembling the macro part. Would it be difficult to redesign the bosses so that they bolt to the weldment instead? That would eliminate the local distortion (and added misalignment) caused by welding them to the plate. It would also make them replaceable and maybe even upgradable (if you decide you need roller bearings later).

    Admittedly it’s significantly more expensive, but if you wanted it to be really trick you could use spherical roller bearings to eliminate misalignment problems altogether.

    Just some suggestions as I’m sure you guys are busting butt. It looks fantastic & I can’t wait to see that leg crush something while under power!

    1. Thanks for the comment!

      Yeah, we’ve now come to accept that attempting to weld two cylinders on to anything and expecting them to align is… just never going to work. We have a number of potential design changes in mind – our favorite right now is just welding a solid tube in where the two bosses need to go and cutting and grinding it off after welding (which we did for the piston-mounting bosses, which worked well). Our runner up at the moment is weld-on spherical joints (with bushings), which we’re currently using for the foot.

      Sadly, adding bolt-on bearings at every joint that needs them is expensive to the point of being prohibitive.

  4. You guys are incredible! I can’t wait to see this baby in action! When you get a person (safely) in the driver seat, we all are going to see photos of your creation on lots of publications!

  5. What happens when you power the leg and it walks off with the building? O.o

    Seriously though, thanks for the update! It’s great to see progress and robot love.

  6. Oh yeah. great update and thanks for all the action shots..

    In the “Encouraging the column mount” shot, looks like the application of a tool I use frequently – ‘the persuader’

    I really hope none of those largish parts fall on anyone! Will stompy have a ‘friend of robots wristband detector’?

  7. Stompy doesn’t like to be hit with hammer. Stompy doesn’t like that one bit…

    Just thinking about the next version and how cool it would be to find the absolute highest power-to-weight ratio power system to drive this beast. I’m thinking of a helicopter turbine connected directly to a hydraulic pump (and an accessory generator, of course). I noticed that there are some portable power systems used for oil rigs that would be a perfect fit.

    I don’t just want this thing to limp along like a lazy elephant but move at a scary pace. A kind of a pace that would make the average person have to change their shorts after witnessing it. The loud scream from the turbine would only heighten this horror. 😉

  8. Awesome work guys. Great to see solid progress. Please keep the updates coming. Ay idea when we can expect to see the goodies for your sponsors ? Im keen to be wearing my stompy T-shirt around tge sreets of Sydney.

Leave a Reply

Your email address will not be published. Required fields are marked *