Compact gearbox design, with two linear shifters, fixed input and output shafts, and large span of gear ratios provided.

This particular idea started with the Truck Trial vehicles in mind. As the trucks used in the Polish Truck Trial are being constantly developed, more and more of them became equipped with simple gearboxes. Such vehicles have the obvious advantage of being able to adapt their drivetrain’s gear ratio to the requirements of the track, sacrificing speed when more torque is needed and gaining extra speed when less torque is needed. Developing a gearbox that would last through a Truck Trial race is quite a challenge, which is why gearboxes only became popular recently.

Aside from highly experimental or lightweight vehicles, all gearboxes currently used in our races are linear gearboxes. There are two reasons for this: firstly, a linear gearbox usually has more compact and thus more robust structure, and secondly the transmission driving rings used in more sophisticated Lego gearboxes are much less torque-resistant than the gear wheels are. Since linear gearboxes only use gear wheels, no transmission driving rings at all, this problem is eliminated. Additionally, with very simple, usually 2-speed linear gearboxes the shifting mechanism is typically very simple too, making the general complexity level of such a gearbox fairly low.

As Truck Trial races involve a significant torque to be handled, the gearboxes are made as simple and robust as possible, rarely using more than 2 speeds. We stick to the rule than any mechanism can be complex or robust, but not both – and it works very well for the Truck Trial constructions. Even a gearbox with just two gears is quite an advantage; a vehicle that can gain extra torque to drive uphill and extra speed to drive downhill will always beat the one that has no gearbox and drives at constant torque and speed. Since the amount of torque required to overcome any given obstacle is more or less the same for every truck, the gearbox-equipped trucks win by simply jumping to high gear and driving faster whenever possible.

There are two basic problems involved when building a torque-resistant linear gearbox: the choice of the gear wheels that will engage and disengage, and the length of the axles that will transfer the drive. As for the first problem, the double bevel gears are a very common choice because the shape of their teeth makes them engage and disengage easily, and also because they are strong (the smallest double bevel gear wheel, the 12-teeth one, is many times stronger that the smallest regular gear wheel, the 8-teeth one). As for the latter problem, the solution is simple: the axles have to be as short as possible and reinforced in as many places as possible to prevent them from bending under torque. Additional protection from the torque for the gearboxes is achieved by using gear reduction between the gearbox and the wheels. Thus a gearbox is placed directly on the drive motor’s output shaft, and it only handles motor’s nominal torque, not the torque multiplied by the gear reduction.

With this said, I started to work on a gearbox for my next Truck Trial vehicle. I wanted it to have fixed input and output shaft, with the only moving elements located somewhere between, to eliminate possible weak points. It meant that I had to use separate shifter axles. Moreover, to keep the axles short I had to keep the moving by a single stud only, so a single shifter axle could only provide two gears. That’s why I used two shifter axles, looking for a gearbox that would provide at least three gears.

The resulting gearbox was supposed to be mounted vertically, but is shown laid on a side in the pictures and video in order to show its internals better.  First video shows bare gearbox and its construction, while the second one shows a version with remote-control already integrated into it, and reinforced adequately. The gearbox provides a wide span of gear ratios, from nearly 8:1 reduction to nearly 1:3 overdrive. It is reasonably small and can shift from any gear to any other gear. Unfortunately it needs two separate motors for remote control and has a rather unusual switching pattern. As shown in the videos, the shifting time is reduced to as short as possible in order to make the gears engage and disengage seamlessly every time.

Eventually, I kept this gearbox for practical testing in some more ordinary vehicle and choose a much simpler 2-speed one for my Trial Truck. I do, however, find it very useful and convenient to control remotely, and I will certainly look for more than just one opportunity to use it in my future models, most likely some at least medium-sized trucks. If it proves successful, it may very well be installed on a Truck Trial vehicle, as intended from the beginning.

Photos:

Videos:

Click here to view the embedded video.

Click here to view the embedded video.

Share/Bookmark

Model of a Scania R500 truck modified as a racing version. Features full suspension, powerful propulsion system with two independently powered RC motors, lights, custom stickers, cabin with full interior and opened doors.

Datasheet:

Completion date: 05/07/2010
Power: electric (Power Functions)
Dimensions: length 48 studs / width 18 studs (not including the side mirrors) / height 27 studs
Weight: 1.86 kg
Suspension: front – independent / rear – dragged axle
Propulsion: 2 x RC buggy motors powered from separate batteries through separate IR receivers, geared down 1.66:1
Top speed: 7.92 kmph
Motors: 2 x RC, 1 x PF Medium

I’m really fond of modern Scania trucks. I have built a model of a Scania dumper truck in 2009, which – while being technically simple – has gathered a significant acclaim because of the quality of its aesthetics and because of the number of details incorporated. I was not fully happy with it, however, because I found the cabin’s look inaccurate and simplified. This was the primary reason why I started to toy with an idea of building a much more accurate Scania model in a racing truck version. Right from the very beginning I wanted it to be red with some chrome and metallic silver elements, which is why I was not convinced of realizing this model, as it had to be expensive to build. Over the time, however, I have gathered a number of chrome and metallic silver elements, and I have determined that it would take less red parts that I initially thought, as Scanias usually have their lower part of the cabin plain black.

Building this model became eventually achievable when I designed the complete exterior of its cabin using the MLCad software and when I bought four metallic silver wheels (to go along with the two I already had). With MLCad model serving as a very accurate shipping list, it became only a matter of time before the red racing Scania would be built.

I have built a number of racing trucks before, which provided a significant building experience. Based on it, I didn’t want to focus just on the aesthetic side of the model – I wanted it to perform well too. The key to surpass the performance of any of my earlier racing truck was the use of RC buggy motors for propulsion, as they offer somewhat more mechanical power than PF XL motors do, and they are also more suitable for fast vehicles, with less gearing or no gearing whatsoever needed. To use two RC motors at the same time while keeping the model’s weight reasonable I had to use the 8878 battery as a power supply. Since it was downright impossible to power two RC motors to a single battery because of their extremely high power consumption, I had to no other choice but to get a second 8878 battery. Thus, the model needed some serious investments before I even started to build it.

The model’s chassis was initially built with both RC motors using no gearing whatsoever – both motors were simply connected directly to the rear axle. Thanks to their convenient shape, it was possible to use them as the main bearing structure for the whole rear suspension which dramatically increased the internal space available inside the chassis. Thanks to it, chassis housed not only two 8878 batteries and two IR receivers, but also PF Medium motor for the steering, with all these elements located between the front and rear axle.  The use of two IR receivers was necessary – since the drive motors were independently powered, they couldn’t be connected electrically and the only connection between them was rigid rear axle. With the initial rear axle set-up it was possible to use a differential between the motors but I dropped this idea for two reasons: firstly, a rigid axle was more efficient (no elements prone to slipping) and much less prone to a mechanical failure, and secondly the rear wheels had only limited traction while turning, since majority of the model’s weight was located above the front axle.

The independent front suspension was based strictly on my previous experiences with the models of racing trucks and used a blend of suspension components from the Lego sets 8448 and 8297. To handle the weight of the cabin, the front suspension was made very stiff, with two hardest shock absorbers set almost vertically above each front wheel. I have also used thin wedge belt wheels between the actual wheels and the chassis to prevent the expensive metallic silver wheels from rubbing away against the chassis – as it already occurred earlier, with slower and lighter racing truck models. The steering system had to be adjusted to the speed and the weight of the model, which is why it used a relatively insignificant gear reduction and no safety clutch whatsoever.

As for the aesthetic side, I came to think that the main disadvantage of my Scania dumper model was the fact that its distinctive front grill was completely flat, with air intakes simply marked with bricks of a different colour. Therefore I have put a lot of effort into inventing a system of recessed air intakes, which was eventually achieved with the use of 2×2 tiles. With plenty of internal space available in the cabin, this system could be used with a relative ease, providing an excellent aesthetic effect, as anticipated. Other than that, the cabin houses not only its own interior, but also a PF switch which controls all the lights (the lights are coupled with the drive motors and go brighter as the truck speeds up, so the switch is used primarily to turn this feature off when not needed, and thus to save energy which is so highly demanded by the RC motors). With 10 PF LEDs installed in the cabin, the switch takes a lot of space to contain 10 PF plugs with their respective wires. I was able to fit it in efficiently by mounting it on its side, so that the plugs are kept horizontally below where passenger’s seat would be located normally.

An important thing to know is that the real racing trucks are usually subjected to very radical visual modifications which make them look pretty much unlike any stock vehicle. Being a huge fan of the original Scania design, I wanted to avoid this. Therefore my approach was to use a stock Scania R500 as the model’s point of reference, with a few typically racing elements added so as to minimize their impact on the model’s look. Eventually, it ended up with five such elements: the four extra headlights (which are not really used for racing, but I viewed it as an opportunity to experiment with the PF LEDs a bit), the racing air intake on top of the roof, the engine moved back from under the cabin, the large spoiler instead of a fifth wheel, and a custom rear bumper. Additionally, the cabin’s interior has been arranged in a ‘racing’ fashion, with a mock-up of two blue N2O bottles replacing the passenger’s seat and with a distinctive safety net in the driver’s window. The net itself is not a Lego element – I made it as a sticker, by printing a black square with a dark grey grid on it, carefully cutting it to net-like shape with a modelling scalpel, and then painting its other side black with a felt-tip pen (I could also print the black square on both sides of the same sheet, but since large portion of it was meant to be cut out, it felt like a waste of ink). Eventually, the paper net was fixed to the door with a thin stripes of a Scotch tape on its edges.

The other, more traditional stickers, have also been intended to reflect the racing ‘nature’ of the model, with the most obvious example being a note in front of the driver’s window that says “Tävlingsfordon”, which is Swedish for “Racing vehicle” (all stickers use Swedish as a nod to Scania’s country of origin). The sticker above the windshield says “Sagodjur”, which can be translated as “Griffon” (the animal used as Scania’s brand mark) and which was intended to look like driver’s nickname, while both front and rear registration plates feature a miniature Scania’s griffon head and a number 5AR13L, which is an obvious reference to my own nickname.

As the model was getting close to its final weight, the propulsion system started to show signs of weakness. To solve this, I have moved the rear axle back a little while keeping the motors intact, and  added a 1.66:1 gear reduction between the motors and the axle, with gear wheels doubled to handle the significant torque. It has greatly improved the overall performance of the model by making it faster and decreasing its accelerating and braking times, but it also made the chassis seem a bit longish and it made the rear suspension very soft. The manoeuvrability of the model, on the other hand, was not noticeably affected by this change. I have ran a number of tests with Lego speedometer mounted on the truck and the top speed it recorded was 7.92 kmph. This is more than any of my earlier racing trucks could achieve, but I feel like it could be surpassed by this very model. The model clearly has plenty of power in the drivetrain and appears to best build up its speed in long runs. It led me to believe that it could break 8 kmph given a long section of a perfectly smooth floor (unlike the floor I had to run tests on). Anyway, 7.92 kmph is a fairly good result, as it requires model’s wheels to rotate at over 600 RPM (that is 10 rotations per second) and – given the model’s 1:17 scale – corresponds with the speed of 135 mph in the real life.

The truck was highly acclaimed for its aesthetics and performance, both emphasized by the video which took me 4 hours straight to produce. It was my second construction to use a slow motion footage (shot with Casio Exilim FS-10), and my first construction to be accompanied by 1200×800 px photos rather than usual 800×600 px, as I really liked the look of the model and the results of its photo-session.

The model will be kept intact at least until a public event that takes place on July 10th, giving my fellow LUGPol members the chance to see, and perhaps even to drive it.

Photos:

Work in progress photos:

Video:

Click here to view the embedded video.

Media reference:

Legocn.net (Chinese only)

Share/Bookmark

Simple table soccer game for two players, computerized and controlled with the WeDo kit.

Datasheet:

Completion date: 28/06/2010
Power: electric (Power Functions)
Dimensions: length 54 studs / width 36 studs (not including axles) / height 45 studs
Weight: 1.88 kg
Motors: 2 x PF Medium

A spontaneous build which started as a joke with my co-workers. It’s a simple realization of the classic table soccer game adjusted for Lego minifigures and automated with WeDo components. Since I have completely no interest in soccer whatsoever, I didn’t pay real attention to keeping the number of players or size of the pitch accurate. My goal (sic!) was rather to demonstrate an innovative use of the WeDo components that to render a realistic model of a table soccer game.

Counting the game score and returning balls to a box are both fully automated. The actual game remains, obviously, fully manual, with eight players on two axles for each team, and a single goalkeeper on a separate axle. The goalkeeper’s sideways movement is limited so that the gate is not accidentally left empty. The players’ axles use different colors for improved distinction. The game is played on a pitch paved with large tiles and surrounded with solid walls build of classic bricks. Both gates have WeDo motion sensors inside, arranged in such a way that incoming ball must pass in front of the sensor. The ball then falls into a simple system of hanging rails, with one rail for each gate and a single central rail that receives all balls and directs them to a box.

The motion sensors trigger a simple algorithm: there is a sound played, an announcement displayed on computer’s screen, and then a second sound follows. Finally, there is a sliding beam above the respective gate which is moved by a given distance.  Both gates have such beams above them, each with a sticker representing numbers from 0 to 10. The beam is concealed behind a wall with a 2×1 stud window, and the desired number is shown through the window. Every time the algorithm is triggered, the beam is moved by a single number, on an obvious assumption that only one goal can be scored at a time. Both beams are controlled by PF Medium motors with a substantial gear reduction – it is possible to control the motor’s speed precisely with the WeDo software and thus eliminate the need for a gear reduction, but it was easier to keep the beam’s movements accurate with a high gear reduction. Additionally, there is a clutch gear at each motor – thanks to this, it is possible to reset the beams’ positions with a single key press, regardless of their current position. Upon pressing R, there is an announcement displayed, both beams are  returned to the ‘zero’ position, and once the procedure is completed another announcement follows.

The construction has seen little of actual gameplay, as it was completed during a holiday weekend and it was difficult to find another player. It was, however, successful in using the WeDo elements exactly how I intended to. Additionally, it was my first opportunity to shoot a video using my new camera with a high-speed filming mode. Therefore a large portion of the final video is composed of a footage shot at 210 FPS.

Photos:

Video:

Click here to view the embedded video.

Media reference:

Ekstra Bladet (Danish only), Engaget.com, Gadgets Now, Gizmologia (Spanish only), Hobby Media (Italian only), Hogargeek (Spanish only), Makezine, Omgunmen (German only), Sema Geek (French only), Technabob, Technoblogia (Polish only)

Share/Bookmark

Simple arm with 4 motors, controllable and programmable with the LEGO Education WeDo software.

Datasheet:

Completion date: 19/06/2010
Power: electric (Power Functions) / pneumatic (fed from internal compressor)
Dimensions: height 58 studs / full length of the arm: 113 studs
Weight: 1.53 kg
Motors: 4 x PF Medium

I had access to the LEGO Education WeDo kit for some time, but it was only after I bought second USB hub that I was able to come up with some ideas on how to make use of it. The kit is in a way similar to the LEGO Mindstorms, except it employs the PF system’s components and it needs a computer to be run on. Therefore it is not as mobile as the LEGO Mindstorms with its separate CPU brick, but it’s much more flexible as it needs a computer and PF components that most of the builders already have.

One of the limitations of the WeDo system’s architecture is having just two independent outlets on each USB hub. While it is consistent with how the PF IR receivers work, it also means that you can only control 2 motors/sensors with the basic kit, and up to 6 motors/sensors maximum as the software can control up to 3 USB hubs at once. This is still, however, better than 3 motors that can be plugged to a single NXT brick.

The first version of the arm was built in a couple of hours to be displayed on the same public event as my Caterpillar excavator was. It was controlled by 4 PF switches, to allow viewers to control it manually. It was, however, built with the WeDo software in mind from the very beginning (hence just 4 motors) and after the event I fixed some minor shortcomings and plugged it to my notebook.

It was easy and convenient to control all the motors using a simple keyboard layout as presented below. Basically, there were two keys for each motor, for running it in one direction or another, and there was a single universal stop key. A separate key was responsible for executing complex programs, and I have prepared one such program for the video (it was a time-consuming task due to the need to manually pre-adjust range of each movement).

Another shortcoming of the WeDo software is that it controls motors using a timeout function and not a rotation counter like the Mindstorms do. The advantage of this solution is the ability to use any motor – the PF extension wires make it possible to plug any 9V motor into the PF system and thus to make it WeDo-controllable. The disadvantage is that motors are friction-sensitive, which means that a motor that is run forward and backward for exactly the same time rarely returns to the exact starting position because of the differences in friction it encounters. With the executable programs, I solved this issue by pre-adjusting every movement manually.

The arm itself consists of a simple grabber with 4 independent fingers, each with a rubber tip for improved grip. The fingers are controlled by an autovalve and thus are capable of adjusting to the shape of the grabbed object. The grabber has a wrist that can be rotated through 360 degrees and uses a safety clutch. The arm is elevated with two linear actuators, with the arm’s level being maintained throughout the elevation process. Two battery boxes are mounted on the rear side of the arm to act as a counterweight for the grabber and its load. Finally, the entire arm can be rotated through 360 degrees by a turntable at the top of the tower it sits on. It felt like something innovative to build, and had certainly a huge potential for various programs and arm’s modifications.

Photos:

Video:

Click here to view the embedded video.

Media reference:

Fascinating LEGO Model Of The Day, Robótica recreativa y educativa (Spanish only)

Share/Bookmark

Large model of a Caterpillar front shovel excavator designed for a public display. Features a drivetrain with a subtractor, superstructure rotation mechanism, V12 piston engine, arm with 3 joints controlled remotely by motorized linear actuators, pneumatically opened bucket, remotely elevated operator’s cabin, lights and custom stickers.

Datasheet:

Completion date: 11/06/2010
Power: electric (Power Functions) / pneumatic (fed from internal compressor)
Dimensions (with the arm in ‘resting’ position): length 80 studs / width 28 studs / height 35 studs
Weight: 3.5 kg
Suspension: pendular bogies
Propulsion: PF XL motor geared down 9:1, through a subtractor
Motors: 2 x PF XL, 6 x PF Medium, 1 x 71427

I wanted to build a front shovel excavator controlled by linear actuators for some time already. It took me a while to find a machine with a distinctive silhouette (as many large front shovel excavators are shaped similarly) and my choice was the Caterpillar 90-tonne machine with an unusual cabin elevation system.

The model would be built at much later time if it wasn’t for the people from the Focus popular science magazine, who invited me over to a Picnic of Science held in Warsaw (the largest popular science event in whole Europe). It was a kind invitation, but it also meant a fixed term to build something spectacular and, so to say, public-proof. I had to hurry to make it on time, I even used my absence of leave to finish the model, but I felt it was still not tested sufficiently.

The look of the model is accurate to some degree. The size and all the dimensions are kept faithful, and the colour theme is an exact rendition of one of the two Caterpillar liveries I’ve seen this machine with. The details on the top of the model, however, were intended to look diverse rather than accurate, the V12 piston engine model is completely unrealistic (but it turned out to be a great crowd pleaser, as anticipated), and the arm lacks a single actuator. Additionally, the way the bucket opens up has been changed slightly to make the bucket look better and be more structurally sound.

As for the functional side, it was largely affected by the unexpectedly huge weight of the model. While the chassis was providently built around a really massive studless frame, the drivetrain had to be additionally geared down to handle it. Even with the 9:1 gear reduction, some gears in the subtractor are still being grinded to dust. Not a single gear in the drivetrain was broken, but there was a plastic dust coming out whenever I opened up the subtractor. But in fact, the painfully low speed of the model is accurate as the real machine moves at 4.4 kmph.

There is a studless Technic turntable between the chassis and the superstructure. The stability is provided by four double rollers, consisting of two wedge belt wheels with rubber tires each, and integrated into the superstructure’s bearing frame (located, as usual, in its ‘floor’). The rollers are supported by large tiles on the top of the chassis. I have assumed that use of large rollers will be a convenient way of filling the gap between the superstructure and the chassis, and that the rubber tires will provide a degree of elasticity should the superstructure tilt while operating. Whether these assumptions were accurate or not is difficult to prove eventually. But I was right to only use four sets of rollers (front/rear/left/right one) as they provided sufficient stability for as long as they were part of a sound, rigid structure.

The major drawback of this model is the use of an 8-teeth gear wheel as the final part of the superstructure rotation mechanism. The weight of the superstructure coupled with the friction generated by the rollers’ rubber tires caused this gear to be severely damaged during final tests. Since it was very difficult to replace (read: it required to take half of the model apart) and there was no guarantee that the new gear wheel wouldn’t share its fate, I left if the way it was. Which means that the superstructure rotation works only when aided by hand. The video clearly shows that superstructure rotates as long as the rear part of the hull is being gently pushed (at some point I even use a single finger for that to demonstrate that there is indeed a mechanism working there and that I’m not simply rotating the superstructure by a hand). The use of the 8-teeth gear wheel, which worked fine in lighter models, was a mistake here to begin with. I should have used a worm gear instead, preferably one that would be mounted to the chassis and driven from the chassis.

The rear part of the hull houses the 8878 rechargeable battery which powers the whole model and which can be accessed by a small sliding lid. On its left, there is a big V12 piston engine driven by a separate PF Medium motor coupled with the drive motor (which means that there is no mechanical connection between this engine and the drivetrain whatsoever). The whole V12 is fully enclosed inside the hull, and can be displayed by opening up a two-piece lid. I always wanted to include a piston engine into an excavator model, and with a model that large I finally did. On the right side of the battery, there are four IR receivers and beneath them, there is a pneumatic autovalve which controls the opening/closing of the bucket.

The front part of the hull houses a massive frame around the arm’s main joint. The arm is 6 studs wide at this point, and there is a driveshaft, two pneumatic hoses and three electric wires going through the joint. On the right side of the arm there is a PF XL motor which drives two linear actuators that lift arm’s first section. There is also a PF switch which controls the lights on the arm, and which can be accessed by the means of a small door in the right side of the hull. On the left side, there is a cabin sitting on a scissor lift mechanism, which is moved by a single linear actuator driven by a 71427 motor located under the piston engine. The elevation range of the cabin is impressive, especially when you realize that the whole scissor lift mechanism is just 3 studs tall. The mechanism has its drawback too – it puts a very high load on the motor, so it has to be additionally geared down (hence the low speed of cabin elevating) and the weight of the cabin has to be kept low (hence the number of details is lower than intended).

The arm consists of two sections. The first one is controlled by the aforementioned PF XL motor which drives two linear actuators integrated into the front part of the superstructure’s frame. The second section is controlled by a PF Medium motor located in the hull, between the arm and the 8878 battery, which drives a single linear actuator integrated directly into the arm’s first section. There is another PF Medium motor inside the first section, which drives another two linear actuators that tilt the bucket. The second section of the arm is pretty much empty. The arm was built in the same way as it was with my previous excavator: it consists of a studless frame with directly attached actuators, and a studfull outer skin which provides extra stiffness but is not necessary for the arm to work. I am not quite pleased with the somewhat weird look of the pneumatic hoses on the arm – I could have enclosed these inside the arm, but I felt that an excavator’s arm ought to have some hoses on it. But due to the complex construction of the arm’s first section, there was no way to arrange the hoses in a more realistic manner.

The arm has some excellent movement range. In fact, it is somewhat better than that of the original machine – to the point at which it is not necessary to open the bucket to unload it, as it can be simply tilted.

As for the bucket opening mechanism, it is very simple and consists of just two small pneumatic cylinders. Some minor part of the load is usually left inside the bucket when it opens, but this is caused by the bucket’s inner shape which results from its internal studless frame rather than by the mechanism itself.

This has doubtlessly seen more practical use than any of my other excavators, as it was operated for a couple of hours straight during a public event. It has also successfully faced the devastating curiosity of little children present at the event, but I won’t describe the details lest I sound like a contraception promoter In fact, the failing superstructure rotation mechanism is the only thing that prevents me from calling it my best working excavator model ever. The model has been also highly acclaimed by LUGPol members – so highly, in fact, that I was genuinely surprised.

Photos:

Work in progress photos:

Video:

Click here to view the embedded video.

Media reference:

Hobby Media (Italian only)

Share/Bookmark

My third 2010 Truck Trial vehicle. Uses modified Jeep Wrangler’s chassis, features 4×4 drive, full suspension, opened doors, bonnet and trunk, headlights and rear lights with LEDs.

Datasheet:

Completion date: 21/05/2010
Power: electric (Power Functions)
Dimensions: length 41 studs / width 20 studs / height 19 studs (not including aerial)
Weight: 1.518 kg
Suspension: pendular, stabilized with 2 shock absorbers per axle
Motors: 1 x PF Medium, 1 x PF XL

My third 2010 Truck Trial vehicle was based largely on the previous Jeep Wrangler Rubicon. It took me over a year to come up with this model since the first plans for it appeared in early 2009 and were postponed since then. It turned out to be good for the model, as I was able to build it entirely in dark gray and using race-proven chassis components. The basic chassis concept remains the same as it was for the Jeep: two pendular studless axles held together by a studfull frame embedded into the cabin’s floor. All the components taken from the Jeep have been modified though: the front axle has a modified steering system that makes the wheels protrude forwards, the rear axle is narrower, and the frame is wider and fully covered with plates from below.

My goal from the beginning was to make this model accurate, even at the cost of its off-road performance. Unlike the Jeep, this model is scaled down accordingly to our Truck Trial regulations, which makes it much smaller than usual TrTr vehicle. Therefore it was clear that it share some disadvantages with the real Hummer, such as relatively small wheels and large wheelbase with low ground clearance. The latter resulted in numerous ‘getting stuck on belly’ accidents during tests.

The Hummer is much heavier that the Jeep was, which was exactly my intent. Based on the performance of the Jeep, which had enough torque to be practically unstoppable, I have estimated that its drivetrain and chassis components should handle weight of some 1.5 kg without problems. Since our rules take power-to-weight ratio into account, heavier vehicles get more points. Therefore I was trying to achieve greater weight while maintaining Jeep’s speed and performance. The speed is almost identical, despite smaller wheels (8 studs in diameter compared to Jeep’s 10 studs), as the gearing in the drivetrain was changed from 5:1 to 4.5:1. In order to obtain this exact gear ratio I have used a double chain between the drive motor and the driveshaft. A single chain would be broken instantly, while a double chain survived all tests I’ve ran. The use of a chain had one extra advantage: it required much less reinforcements in the surrounding structure that the gears alone would, since the gears were, well, chained together. Eventually, there were some minor inaccuracies in the model as compared to the real H1. The bonnet had a cut-out section in the middle of it, which was necessary to accommodate for the upper ends of the front shock absorbers. The air intake, which in reality is fixed to the chassis and goes through an opening in the bonnet, was attached to it in the model. I have also reduced the size of the front bumper and gave up some extra elements such as the famous “cow catcher” in order to achieve a better approach angle. Finally, there were openings in the sides of the cabin that allowed the doors to open up easily, as there was not enough internal space to use opposing hinges and I didn’t want the doors to require force to open.

The model participated in a Truck Trial race on May 22nd, and it generally performed above my expectations. It had sufficient torque and displayed some excellent manoeuvrability. The stability was satisfactory, except for one incident when I managed to flip the model over somehow. There were some traction problems, but not larger than those experienced by other vehicles (we keep having problems when driving on a dry forest bed). After the first section of the race Hummer placed 4th out of 11 vehicles, which was partially thanks to the power-to-weight ratio more favorable than it was for the Jeep. During the second section it easily climbed over a large branch that seemed too big for its ground clearance. Then during the third section both chains in the drivetrain broke at once. Since it was impossible to fix it on the field (large portion of the cabin had to be taken apart to access the chains, and I needed tweezers), I was forced to give up the rest of the race and thus the model ended up on the last place. It seems that the model had a considerable potential and could effectively compete against larger vehicles, but the idea of using chains in the drivetrain was clearly wrong. The thing I’m happiest about is that I’ve finally managed to put the damped shock absorbers to some practical use, even if it didn’t look as good as simple plates on hinges would.

Photos:

Videos:

Click here to view the embedded video.

Click here to view the embedded video.

Share/Bookmark

Sariel.pl has a Downloads section now. The list of files available for downloads is not a long one at the moment, but starts with a complete, high quality step-by-step instruction for my acclaimed Jeep Wrangler Rubicon model. You can get it for free and try to build one on your own.

When I’ve completed my Jeep Wrangler Rubicon model and declared that I’ll publish instruction for it if it works well in the Truck Trial, I did not expect much feedback for this unsophisticated and rather small model. Yet the Jeep has shown its charm while working its way through the Trial Truck track. A lot of people have become very enthusiastic about building their own copy of it, which flattered me but also meant that I have to make sure the instruction will meet their expectations. First, I’ve created 3D models of front and rear axle, so that builders interested only in suspension wouldn’t need to wait for the complete instruction. Then I proceeded to create the complete instruction by gradually disassembling the model and taking photos of it to eventually publish them in reverse order. This is a well-known and relatively time-efficient method, used very well by builders such as e.g. Nico71. However, the photos are usually less readable than 3D renders.

I was unwilling to create a complete 3D model of the Jeep, and this is when I received an unexpected, yet valuable help from a reader called Jegge. Having only the aforementioned photos as the reference material, he managed to model the entire Jeep is just a few days, and he was kind enough to release the model for my use. I was deeply impressed by the quality and accuracy of his work, and I felt that I should do something to let other people benefit from his effort. Jegge has agreed to let me correct some minor inaccuracies of his model and create a complete instruction based on it, which would be released free of charge. Jegge’s work means a lot to me, and his selflessness is a great example of how amazing the community of Lego fans can be.

Releasing such a professional instruction was a good reason to add another section to my website – the Downloads section. Right now there is just one instruction and two tutorials converted to PDF format for your reading convenience, but hopefully more will be added later.

As for the instruction itself, my goal was to make is as easy to use and as intuitive as possible. This is why it introduces a different building order than the earlier photos-based version, while being much longer as well. There is a complete list of parts included, but keep in mind that this is a software-generated list and may not be 100% accurate. Moreover, many parts can be substituted with different Lego pieces, and it’s up to your creativity to try out some changes or modifications. Honestly, I would be genuinely pleased to see some severely upgraded version of this model built by someone else

I hope the instruction will be useful to you, and that building the model will be a rewarding experience. Note that I have currently no plans to create more instructions – which obviously doesn’t mean that I’ll never do that again.

Update: the ZIP package now includes also a plain text file with parts list and MLCad models of the whole model, the front axle and the rear axle.

Download the instruction in a single ZIP file or browse instruction:

Share/Bookmark

The Polish edition of the Focus popular science magazine has featured my Robotic Hand in their May issue. Scan included.

It seems like every hobby has its own surprises. Personally, I never expected to become known for something I’ve built over a single weekend just to keep myself occupied while waiting for the parts needed to finish my S Tank 2 model, but here it is – the Robotic Hand, a construction I considered so insignificant that I almost neglected publishing it at all, was featured by at least 15 various websites including even the mighty Engadget.com, and has recently made its way into the Focus magazine. It has been mentioned as one of four constructions in an article entitled “Klocki mają moc” which translates as “Bricks have the power” (note the unintended pun – the Polish word for power, “moc” is written just like “MOC”, an abbreviation for “My Own Creation” commonly used by the LEGO builders). The other three constructions include the LEGO house built by Top Gear’s James May, the Dispatchwork artistic happening using the LEGO bricks as a medium, and the 1:1 statue of Christ displayed in Vaesteras, Sweden. Not that it makes me a real star, but but it’s still a remarkable company to find oneself among. And the Focus magazine is not your ordinary newspaper, as it’s being released for 15 years now, with monthly circulation currently at 170,000.

On another note, I have been invited by the Focus magazine to participate the Picnic of Science which will be held at Warsaw on June 12th. I’m supposed to bring some model along and present it to the public. We are still settling some details and I’m not yet sure which model to choose.

Here’s a scan of the Focus article. I apologize for the quality, but my scanner is pretty much antique.

Share/Bookmark

Datasheet:

Completion date: 18/04/2010
Power: electric (Power Functions)
Dimensions: length 47 studs / width 24 studs / height 23 studs
Weight: 1.224 kg
Suspension: pendular, stabilized with 2 shock absorbers per axle
Motors: 1 x PF Medium, 1 x PF XL

After the disappointing failure of my Tatra T815 truck, I was left with less than two weeks time until the next race. Re-using Tatra was out of the question, so my only option was to build some simple and small 4×4 vehicle. Since our current rules allow this type of vehicle into competition, I decided to model the iconic Jeep Wrangler in the acclaimed Rubicon version. It was built in roughly 4 days.

The basic rule of this construction was to keep everything as simple as possible and to reduce the weight to minimum. This is why the body was built with the new Technic panels, which provide the best combination of stiffness and lightness. Both front and rear axle are built entirely with liftarms, but the frame that connects them is built with Technic bricks. The reason to use bricks was that I wanted to build a roofless version of the Jeep in order to obtain low center of gravity, and since the sides of the cabin are made of single panels, the whole model had to be kept together by the cabin’s floor. Therefore there are bricks in the floor, that provide excellent stiffness and robustness while keeping the floor thin.

The chassis, which weights just 0.75 kg, consists basically of the studfull frame connected with studless axles. Since there is very little space in front of the front axle, both drive and steering had to be transferred to the front axle from behind. This was achieved by placing a 16-teeth gear with a clutch on the driveshaft – this gear works as an idler gear between two other gears, thus transferring the steering independently to the drive. Front axle is compact but massive, strong enough to handle much heavier vehicle, and it’s bottom is entirely covered with liftarms to prevent it from getting stuck on an obstacle. It should be noted that both axles use knob wheels instead of differentials, and both are portal axles with an integrated 3:1 gear reduction. Each axle is stabilized by a pair of short shock absorbers located between the axle and the frame for maximum structural integrity. In order to minimize the effect of central driveshaft’s torque tilting the body, which occurs when only pendular axles are used, the axles are stabilized with the hardest shock absorbers available. The rear axle is built around the 7×5 liftarm frame and is not covered from below.

The model is driven by a single PF XL motor with a total 5:1 gear reduction, and the motor’s location is strictly related to our current rules. Our rules state that every model has to be equipped with a piston engine whose set-up and location are consistent with the original engine. This rule, however, can be omitted if the drive motor is located exactly where the original engine is. This is Jeep’s case – the PF XL motor occupies exactly the same place that is taken by the original engine in a real Jeep, hence there is no need for the piston engine and the model can be simpler and lighter. You can see the top ends of the front axle’s shock absorbers next to the motor. It should be noted that the PF XL motor actually touches the front axle, but because it’s located in the center of the chassis and because it has a round shape, the axle can still oscillate under it. Such a trick was necessary to fit the XL motor under the bonnet.

The steering is controlled by a PF Medium motor with a 9:1 gear reduction. The motor is located between the seats in the cabin, and the steering wheel is connected to it. The cabin has a safety cage built on top of it, which is strong enough to survive a turnover without any damage. Finally, there is the 8878 rechargeable battery box and the IR receiver located over the rear axle, as a counterweight against the PF XL over the front axle.

One special feature of the bodywork is a droppable modules design. The obstacles used in our races are often larger than the ones the real vehicles are designed for, so the usual practice to handle this difference is to increase the model’s ground clearance. I wanted to avoid it, however, to keep the Jeep’s center of gravity as low as possible. It meant that some parts of the body are very likely to touch the ground e.g.  when the approach angle proves to be insufficient. Therefore I designed a number of body elements as independent modules that would fall off when stressed, on assumption that it’s better to drop something than let it get the whole vehicle stuck. The list of droppable elements includes front and rear bumper, rear spare wheel and front mudguards. The sides of the cabin are somewhat elastic too. This design is also helpful when the front wheels’ steering lock makes them collide into some parts of the body.

The model turned out to be pretty stable, with a well performing suspension and an excellent amount of torque available. It was my goal to prefer the torque over speed, since I did not want the model to get stalled under any circumstances while its speed didn’t seem crucial. This is second model of a Jeep used in our races, and it’s much different from its predecessor, the Jeep Willys built by Atr. Among other things it is heavier, larger and slower. Observing the two models compete against each other and comparing their performance will be certainly a great experience.

The main disadvantage of this model is its size. The standard scale used for vehicles in our races is 1/13. I was in hurry while building the Jeep, so I picked up the wheels I found most suitable and scaled the whole model accordingly to their size. Eventually it turned out to be in 9/1 scale which is much different from the required one, and will probably result in letting this model race outside the official ranking (in a so-called ‘open class’). Still, building and testing this model was a valuable experience, and driving it in a race will be so too. It was also acclaimed by other builders and race contestants for its look, its authenticity and its performance. If it turns out to be successful in the race (meaning that it suffers no fatal malfunction and doesn’t end up last), I will prepare and publish a complete step-by-step instruction for it, as well as schemes of front and rear axle along with their individual instructions.

Update: since the Jeep performed well at the race, the instructions are under the photos.

Photos:

Full instruction with parts list:

Photo-instruction for the whole model:

Videos:

Click here to view the embedded video.

Click here to view the embedded video.

Share/Bookmark

The tool that has been available at my website for a long time – the gear ratio calculator – has been released in an entirely new version. It also includes an extra handy feature called the Gear Coupler.

Calculating ratio of two LEGO gears is fairly easy. It gets a tad more difficult when some less common types of gears are used and when there are multiple pairs of meshed gears. To provide an easy solution to this, I have created a simple on-line calculator designed specifically for LEGO gears some time ago. It was a very basic tool and now it’s finally replaced by a newer version.

The new gear ratio calculator provides a convenient graphic interface with pictures of gears (included are all currently used LEGO gears from 8-teeth gear to Hailfire Droid Wheel, along with all types of differentials, turntables and with a worm gear too), an ability to calculate the total ratio of up to 25 pairs of gears, and it works entirely in real-time, without reloading the page. It also comes equipped with a handy new feature called the Gear Coupler which shows gear combinations available at a given spacing between the input and output axle. It works in two modes, one for studless constructions (based on liftarms) and another for studfull ones (based on LEGO Technic bricks) – this is because in case of bricks the vertical spacing between the axle holes is different than the horizontal one. The list of gear combinations is provided along with pictures, gear ratios and explanations, and is ordered from the combinations providing the highest gear reduction to the ones providing the smallest one. Please note that this tool only works for parallel gears, hence no single bevel gears, no differentials, no turntables and no worm gears are taken into account. Still, it should be more convenient than having to manually check what gears fit into a given spacing.

Update: at readers’ suggestion, the calculator was given an ability to clear the selected driver or follower gear, as well as to remove an entire pair of gears. The total gear ratio of multiple pairs is automatically updated upon removal of any of these.

Update: the option to select the drive motor and see the theoretical output speed was restored. The calculator is also available via a new address, gears.sariel.pl.

The calculator requires a JavaScript-enabled browser and may not work properly under Internet Explorer in versions prior to 8.0. This is because of extensive use of the jQuery library.

The gear ratio calculator 2

Media reference:

Robótica recreativa y educativa (Spanish only), TechnicBRICKs, Tecnoloxia.org (Spanish only)

Share/Bookmark

My first Truck Trial vehicle for 2010. Features 6×6 drive and full independent suspension.

Datasheet:

Completion date: 10/04/2010
Power: electric (Power Functions)
Dimensions: length 66 studs / width 25 studs / height 31 studs
Weight: 2.624 kg
Suspension: full independent with swinging half-axles
Motors: 1 x PF Medium, 2 x PF XL

My first 2010 Truck Trial vehicle had a rather difficult creating process. Even though I started to build it more than a month before the race, the final version was created under two weeks time. It was mainly because of the chassis, which wasn’t stiff enough to handle the torque of two PF XL motors, and got completely rebuilt three times.

The goal of the Tatra was to have a realistic suspension with swinging half-axles which are typical for all Tatra trucks. It was achieved by removing differentials and transferring the drive with the knob wheels. The modern Tatras have their half-axles offset a bit – this was not modelled in order to keep the steering system simple and sturdy.

The steering system was located entirely between axles 1 and 2 and consisted of two complex sliding beams with links. The beams were sliding in different directions for axle 1 and 2 and were controlled by a driveshaft whose direction of rotation was inverted in between these axles. Because of insufficient internal space both beams were given the same gear ratio, resulting in an identical steering lock of both axles.

The drivetrain was kept as simple as possible, and the main portion of it, including the drive motors, was located between axles 2 and 3. Both drive motors were located between two stringers of the vehicle’s frame, one behind the other, and were not connected directly but by a common driveshaft located beneath them. Since the space between axles 1 and 2 was occupied by the steering system, an extra driveshaft was needed to transfer the drive to axle 1. It was located above the axles, under the cabin, and it transferred the drive to a V12 piston engine located inside the cabin as well. Thus the transmission of the axle 1 was located in front of it and in order to prevent such a vital mechanism from damage, a cover made of liftarms was installed before the axle.

The body of the truck was rather simple, with the cabin built almost exclusively with the new Technic panels to reduce its weight. The cargo case was simple and shallow, and there was a transportation handle built into vehicle’s frame behind the cabin – a solution I’ve used in some of my earlier Truck Trial vehicles, and which proved very useful as every Truck Trial race involves a lot of carrying the vehicles around. The cabin had a safety cage built on it, which was meant to give it extra stiffness and a more Tatra-like look, as the real Tatras participating Truck Trial races are always equipped with distinctive external safety cages.

The Tatra suffered from the same thing that most of my earlier Truck Trial vehicles did – insufficient amount of testing. The problem with this kind of vehicles is that they can be only fully tested once finished, because test are most revealing at full vehicle’s weight. As you can guess this implies that a finished vehicle has to be taken back apart if it fails tests, and to avoid this arduous procedure it is a common practice to test just the chassis alone with extra load on it simulating the anticipated body’s weight. It doesn’t always work as it’s difficult to precisely anticipate the final weight of the body if it wasn’t built yet, and it’s almost impossible to simulate its exact weight distribution. I have to admit that being short on time, I have only given a minimal amount of testing to the final version of the Tatra – this was also because I was afraid of damaging it right before the race.

During the race, the model experienced two problems – the first was a failing transmission to the front axle, which snapped. There were also minor problems with axle 3, which meant that only the middle axle worked at full efficiency, and it didn’t really help while climbing obstacles. The reason was a faulty construction of the chassis – the suspension itself worked fine, all the snapping occurred in the transmission system inside the chassis. The second problem was sideways stability. While building this model, I have given a lot of attention to make sure its weight distribution prevents it from falling forwards or backwards – and I succeeded. The sideways stability, on the other hand, proved much below what could be expected. It seems that the problem was not related with this particular vehicle but with this type of suspension in general, as exemplified by another Tatra model, the T813 built by Lomero and used in the same race. His Tatra had the same suspension type while keeping a considerably lower silhouette which suggested a better stability. Surprisingly, it kept tumbling over a couple of times per each section of the race. It convinced me that it’s practically impossible to model the Tatras’ suspension system in such a way that the resulting model is stable.

Finally, the model’s drivetrain was pretty much obsolete in terms of the rules used in Polish Truck Trial. The rules take power-to-weight ratio of each individual vehicle into account in order to prevent builders from winning by simply adding more drive motors. Each vehicle is given its own points conversion rate based on its power-to-weight ratio, which basically means that you can win with a vehicle that can race despite having a very low power-to-weight ratio (resulting in high points conversion rate) or with a vehicle that has a high power-to-weight ratio but is very fast and can complete the race in short time. Tatra was neither of these cases – it had a relatively high power-to-weight ratio because of two PF XL motors, while not being really fast. The current tendency among the top Truck Trial builders is to use complex gearboxes with 3 or even 4 gears available. Thus even a truck which is much heavier than Tatra can be driven by a single PF XL motor whose gear ratio is changed as needed by the means of the remotely controlled gearbox. This results in a models that are slow but still capable, while having an extremely high points conversion rate. Two good examples of such trucks are Emilus’ GINAFs and Atr’s Volvo – both being large, heavy, and driven by a single motor.

Upon observing the race, I’ve decided to abandon the Tatra project. While the transmission problems could be solved given a sufficient amount of testing, the suspension system would still not provide satisfying stability. With the next race being only two weeks away, I decided that I would focus on building a new, small, lightweight and efficient vehicle rather than fixing Tatra. This is also the reason not to document the Tatra’s technical solutions – I think it’s better to build something more successful and then document it once it’s solutions are proven in a race. Right not I have a complete front suspension of my next Truck Trial vehicle, which looks very promising, and if it succeeds in the race, I intend to publish a complete instruction for it.

Photos:

Videos:

Click here to view the embedded video.

Click here to view the embedded video.

Share/Bookmark

Re-edition of the model of the Swedish MBT, built two years after the initial version. Features full suspension, remotely tilted hull, remotely controlled dozer blade and lights.

Datasheet:

Completion date: 06/03/2010
Power: electric (Power Functions)
Dimensions: length 67 studs (including main gun’s barrel) / width 25 studs / height 19 studs
Weight: 2.15 kg
Suspension: pendular bogies
Motors: 2 x PF Medium, 2 x PF XL

After almost exactly 2 years, I decided to re-build my perhaps most renown tank model. The S Tank, created back in 2008, has been extremely well welcomed, and no matter how many bigger, more accurate and more complex tank models I’ve built later, there were always some people claiming that S Tank was the best. I, on the other hand, was aware of many shortcomings of this model, and it was my goal to overcome these shortcomings in the following models – a fact that has apparently escaped the attention of many S Tank’s enthusiasts. Eventually, I decided to build this model again using my current level of modeling skills, even though it was against my ‘never build the same thing twice’ rule.

The original S Tank, or Stridsvagn 103, is a fascinating machine all by itself. Designed in 50s, it was the main tank of the Swedish army for the next 40 years, until arrival of the Leopard 2. What was the key of its success and its long-term service? The fact that it was designed with Sweden in mind. Moreover, it was designed specifically for defensive roles. This is why the typical turret was removed completely, because the tank was intended to stay hidden waiting for the enemy to come. To operate in difficult Swedish terrain, the tank was fitted with then-advanced hydropneumatic suspension and a turbine engine (in addition to a traditional diesel engine), making it the world’s first MBT to use turbine engine. Aiming was done by moving the entire tank, using the suspension to tilt the hull as needed. In order to fire its main gun, the tank had to be stopped and its suspension locked – it was clearly a drawback, but it was acceptable given the tank’s defensive role. It should be also noted that in 50s and 60s the fire control systems were relatively primitive, and tanks rarely fired while moving. Finally, the S Tank was equipped with a front dozer blade, which could be used to entrench it quickly and without the need for crew to leave the vehicle.

Thanks to removal of the turret, the tank’s front could be protected by a single armored plate, set at an angle that made it very difficult to penetrate. Actually, at the time it was designed, the S Tank’s front armor was virtually impenetrable for any other tank’s armament. To make a full use of this advantage, the crew included two drivers – the second driver was facing backwards and had his own set of steering instruments. Thus the S Tank was able to drive backwards quickly, while being protected from the enemy by its front armor, as well as being able to return fire at any time.

The S Tank was clearly an unorthodox design, so specific that Sweden remained its only operator, but it could be pretty lethal to any advancing force. Why? Because it had to be spotted before it could be engaged. Now, take a good look at a fully entrenched S Tank:

The new model of the S Tank was supposed to be: accurate, performing well and dark gray. I believe it meets all these conditions. The hull, which was tilted pneumatically in the previous version, is now controlled by linear actuators. It takes much more hull’s internal space, but the stability provided by such solution is peerless. Since my first S Tank was driven by a single XL motor, and the motor was apparently working pretty hard, and now the PF speed control is available, I used two XL motors, each one driving a single track. The tracks’ suspension system consists of 4 large, realistic road wheels (6595 wheels without tires work very well with tracks) suspended on pendular bogies, two wheels per a single bogie. I wanted to use a torsion bars suspension for the first time, but it was impossible to make it work together with the tilted hull. The front dozer blade is operated by a single PF Medium motor instead of pneumatics, and it works better than I expected (this is an application where pneumatics usually performs better than motors do). It should be noted that front part of teh hull houses a traditional battery box rather than the rechargeable battery, and there are two reasons for that. Firstly, a battery box can provide more power for two PF XLs used simultaneously, and secondly I needed something heavy in model’s front part to acquire the proper weight distribution.

As for the visual details of the model, my goal was to model the most distinctive ones. Like many tanks with long service period, the S Tank has undergone several modifications and there are 4 variants of this tank, which differ mainly in details. Few details are fictional, just to fill some empty spaces. Please note that this model uses custom chrome sprocket wheels, rather than original ones that come only in yellow or orange at the moment. These wheels have been purchased at this Bricklink store.

The model took a lot of work and a lot of extra shopping for parts, but I think it was worth it. For me, it is clearly my best-looking tank up to date. It seems that models in dark gray look particularly well on photographs, which is why I attempted to make a couple of macro-like shots, even though I only had a general purpose lenses (macro lenses are expensive beyond human reason). The model performed very well too – it was my fastest tank up to date, and its large road wheels made the suspension’s work spectacular to watch. I have enjoyed driving it over a wrinkled blanket more than I enjoyed driving any other tank under any other conditions. However, there’s a good chance that it would perform better with a torsion bars suspension.

Photos:

Video:

Click here to view the embedded video.

Share/Bookmark