FMP - Seven (9/03/11)

As promised here are the photos from what I mentioned in yesterdays update. To create a fillet with lasercut sandwiched pieces, the top and bottom layers have to be slightly offset inwards from the outer edge. This is what they start as, with the two thinner pieces sandwiching the larger piece:

Once lasercut and attached to each other, they are covered in filler:

And then finally sanded back to create the fillet and give them an overall more circular appearance:

The larger piece is what attaches the lower moving base to the top, they slot into each part:

Moving onto the new CAM, here it is:

As you can see it still needs to be sanded down and filled, but it has fixed the alignment problem thankfully.

FMP - Six (4/03/11 - 8/03/11)

A lot has happened since the last update, on the one hand I got the carousel working at the end of last week, but then on the other there's been a serious alignment problem.

To get the bad news out of the way, somehow, even though I used templates straight from my digital model, the centre MDF piece is lopsided, and as a result this has had a knock on effect of everything which surrounds it, most importantly the CAM. The now realised off centre CAMs and MDF piece:

As it is lopsided, the above slot for the vertical poles to slide down doesn't accurately follow the CAM, and in other places falls off it. Being only 6mm wide, there isn't enough room for this leeway, so I've had to extend it's width by sandwiching the original CAM with two 5mm lasercut and then heatbent acrylic pieces, I've yet to get a photo of this, but I will for the next update. 

The problem doesn't end there, as the original CAM was only 6mm wide and sandwiched between two higher matching styrene profiles, to create a trench, I could use circular vertical poles, as the wheel they'd be attached to would prevent it from going on its side (a massive problem with my mock up). But now the width has been extended to 16mm, this prevention is useless, so I've had to redesign them so that they are now more square thus unable to rotate. This does have its upsides though as due to the cartoonish nature of the carousel, these poles were going to be heat bent, and for the wheels to be attached they'd have to be milled at one edge. But now these poles are made up of a 3mm profile sandwiched between 1mm thinner profiles, which will be filled so that it has somewhat rounded corners. Its cutout in the base is slightly larger than this, but due to the square sides and edges it prevents it from rotating. Photos to better describe this fix will be uploaded tomorrow.

Moving onto the good news, getting it to work. This started with lathing the aluminium bar to create a hole in one edge for the rod from the motor to slide up to. A perpendicular hole was then drilled to penetrate this hole, and this hole was threaded in turn so that a screw could be fitted. The rod from the motor has one flat side (which doesn't show up too well in the photos) which is meant for an external screw to go up against. All in all this prevents the bar from freely rotating while attached.

The bar attached to the motor:

The hole in the edge where the rod slots up:

The blurry flat edge of this rod:

In celebration of it working, and the top piece successfully rotating the bottom one I recorded a short video. Even though this was the moment that highlighted the alignment issue, it is still worth noting.

FMP - Five (3/03/11)

Progress is moving quickly with the baseboard. Today marked the most important part of my model, the CAM, in that getting it into shape and attaching it to the base. After having figured out the circumference issue to get it perfectly wrapped round, I lasercut the hilly design in 6mm acrylic:

To transform it into a circle I needed to heatbent it, having learned the hard lessons of heatbending from the previous project, I made a negative and positive form for the job. Once the acrylic was flexible enough it was pushed into the negative former:

In order for it not to warp and cool back on itself the positive former was then pushed in:

I needed two formers due to the different diameters of the two CAMs. They were made in the same way as my base ring, that is using a router. Even though I had the correct length for these heatbent pieces, there is a gap where they meet due to the rounding down of circumference values when halving it; I tried filling it using the formers but it didn't work as well as I hoped, so I plan on doing this when they are attached in place.

I also made the supporting MDF cylinder for the mirrored part (as shown in the last update), which had to be at a set height so as not to be visible. The 25mm acrylic tubing sits on top of this block and then runs up inside of the mirrored part. 

It's important that the CAMs are in the correct position, or else everything above it goes to waste. To do this I lasercut styrene doughnuts as seen in the two below photos:

The outer/larger CAM is not yet attached due to a lack of styrene doughnut but this will be fixed in place tomorrow, and then both can be filled and sanded. Along with that I've also got to cut out the 0.5mm styrene runners which sit either side of both CAMs to prevent the wheels going off.

To end on a bad note I realised tonight, when assembling it that my lasercut pieces (the actual parts which rotate) were taken from an old model and as a result the holes cut out for the actual vertical poles which run over the CAM were out of place.

FMP - Four (21/02/11 - 2/03/11)

First things first, I've started on the final model, working from the base up. This means the baseboard of course came first, followed by the base structure. Last week was also spent waiting for things to arrive, such as my new springs, wheels and most importantly my aluminium square bar (which everything rotates off).

To have something to build from I needed a digital model so I begun working on that, using all the information I gathered from my mock up models, such as gaps to avoid collision and the below CAM. The baseboard is also more complex than a usual one, considering I needed to fit the motor, switch and battery pack into it. The switch (like most off the shelf ones) has an inbuilt tension/locking device which allows it to be slotted into a hole that matches its size and the device then holds it in from the back. The MDF I used for the baseboard was far too thick for the device, so I milled out a slot in one side and place two lasercut acrylic layers for it to sit into:

The most important part of the baseboard is the location of the gearbox motor, in that it has to be perfectly in the centre or else everything will be off. To achieve this I used two runner MDF pieces (the double thick MDF edges on the right and left of the below photo) and a printout to make sure it was located correctly:

Going from my initial designs I also digitalised the centre part of the carousel, which is fixed to the base. In order to make it more carousel like (I originally was going to have Dr Seuss photos in the frames) I opted for mirrors, the mirrors used are 2mm mirrored styrene with 1mm styrene cladding in irregular shapes to give that 'cartoony' look:

To hide the CAM and to allow the carousel something to slide over an outer base ring was made from two MDF pieces done via a router, with 2x4 (all cut to the same height) sandwiched between to achieve the overall height of 50mm. An acrylic top sits on this base with holes cut out for acrylic rod which goes on to support the above stationary structure. To hide the exposed sides, it will be cladded:

After spending a week waiting for my sqaure bar to arrive I decided to make me own, sadly that evening upon returning home I found it waiting for me, after being delivered during the day. Needless to say I was happy to see it arrive, but I wasted an afternoon on taking down a larger bar:

In more recent news I'm looking into possible colours to match the Dr Seuss theme:

The solid colours for the base don't work, so I'll probably add a streak design into it, something to break up the colours, such as this (which is from the Dr Seuss island at IOA):

FMP - Three (18/02/11 - 20/02/11)

As promised I made a similar working model to that in the last update, but with a bar instead of a rod. The same principle applies, but to highlight the problem of only using one CAM I had one of each design along side each other (one CAM and then two CAMS):

As I was using a bar instead of a rod, to achieve the maximum grip between parts, I had to lasercut my own gears. This bar was slotted through acrylic rod which was in turn threaded and glued to the vertical dividing pieces, to strengthen and maintain the bar at an equal height along its length:

Back to the CAMs; the two different CAMs next to each other:

As you can see, the CAM device on the left is comprised of two CAMs; a secondary one fitted between the outer CAMs which freely rotates on a rod between these two outer ones. The vertical pole which is attached to the horse is also attached to the other end of this inner CAM and is again, allowed to freely rotate. So that when the outer CAMs are rotated up the inner one is also brought up, but as it's allowed to rotate freely, the end where the vertical pole is attached can maintain its vertical stance (due to it remaining over its centre point) while being lifted up:

And vice versa when it comes down:

The same can not be said for the other CAM device (the one on the right), which angles the vertical pole while going up and down, and as a result, needs a larger hole to be slotted through. Going up:

And coming down:

This angle at which this vertical pole becomes when completing its journey is obvious when compared to that of the outer CAM device. And as I mentioned before, the idea that the horse goes back on itself when brought up is clear in the following comparison photos. The nearer horse is attached to the two CAM device, with the further horse being attached to the one CAM device:

That is them going up, and it's clear to see the idea of the horse going back on itself.

And coming back down, notice the angle at which the vertical pole is at compared to the nearer one. Now the same comparison but from a different angle:

This has lead me to believe that even though I don't have concrete proof that in actual carousels they use this secondary CAM device, they must incorporate something incredibly similar as to prevent the horse going back on itself, which could possibly slide a rider off. It's also interesting to note that the secondary CAM horse doesn't go nearly as high as the other one.

This design is also titled as the most reliable method of getting a carousel to work and provides the smoothest ride. But from my perspective, via the making of these working models, it's also the most complex and fragile, requiring more engineering than lasercut acrylic pieces. For this reason I'm going with the below CAM design which I used on my first working model.

In other news, my gearbox motor finally arrived:

It's smaller than I thought, so I plan on attaching it upright instead of using bevel gears. Also from my working models the idea of using a bar is far more reliable than that of a rod, so I'm going to mill the rod coming from the motor square, and do the same with a larger vertical rod which will be the main centre point which everything rotates from. To attach the two I plan on milling out a square indent in one end of this centre point rod so that rod from the motor can slot up into it.

Along with the motor, and in preparation of using the below CAM design I've brought a small tyre (I had to order in 11 more as this is all they had, they should be available by Wednesday) and some springs:

FMP - Two (15/02/11 - 17/02/11)

First things first, I fixed the first working model:

I had trouble again trying to get the length of the heatbent acrylic right, but then I realised what I was doing wrong. The circumference I was getting was found by using the diameter of the inner circle that I was bending it around, I hadn't taken into consideration the width of the acrylic, which then knocks the diameter up by 10mm, having a great effect on the circumference. However this new circumference came up too large and struggling to think what was wrong I tried using the diamater from the middle of the acrylic, so increasing the diameter by 5mm this time (2.5mm both sides), so that the radius of the circle I'm using to get the circumference would slice the acrylic in half. This worked a treat and came up perfect but due to warping from the heat, I had to fill the tiny gap as seen in the photo below:

In regards to the trouble of the poles becoming stuck, I swapped them for thinner ones with rounded off edges to help them glide the CAM. Also pieces of rod were glued in the middle which the above base now sits on and can easily rotate. All in all there is no trouble with it, but it has highlighted a few issues. The major one being that these poles, as they hole they sit in was cut for a larger diamater, spin a lot. It's hard to keep them from doing, but I've thought about attaching small tires onto the ends which together with guides either side should prevent them from spinning.

I've also thought of placing springs into the top acrylic houses which the poles go up into, so that they're constantly forced down, thus hopefully eliminating them getting caught mid air. This is the acrlyic housing:

Moving onto a second possible design of getting to work, is the above CAM design that I mentioned in my before post. This works by having horizontal poles coming from the centre point (which is turning) which have a bevel gear attached which rotate due to a horizontal bevel gear. So that when the centre point rotates, the poles rotate along with it as well as rotating along their length. Attached to their length is an offset CAM device which, while rotated, creates a greater radius than that of the pole, this is where the vertical pole along with the horse are hung. If that sounds too confusing the following photos should clear it up:

So as the structure is rotated around the wooden base the vertical gear is rotated too:

This in turn rotates the CAM device:

I was initally having trouble making the rod and gear one piece as the hole in the gear was too large (it was store brought). To fix this I attached a copper arm which is soldered to the rod as well as slotted through the gear:

A problem I raised in my research was that the CAM device would cause the pole to go back on itself when it was rotated. The vertical pole should just move up and down with as little forward and back motion as possible. So to keep it centred but also allow it to raise, I added a secondary CAM (the inner parts in the photo below) which when rotated by the outside ones would go up or down, but being loose on one end it allowed the pole to stay somewhat centre:

The idea works fine but it's the model that is fragile and prone to these bits falling off. Just like the vertical gear, attaching these CAMs to a rod isn't as easy as I intially thought as there is little to grip onto from anything circular. For this reason I'm going to try using a sqaure bar tomorrow, which will turn anything slotted over it regardless of how tight the joint is.

FMP - One (14/02/11)

Well today marked the first day that I physically start my FMP, after having spent too much time bogged down with my External Brief referral.

For my FMP I'm building a working carousel based on the works of Dr Seuss, this entails that the overall design will be asymmetrical, wonky and colourful. Having become bored of the perfect finish needed for product models, this freedom to allow certain angles or corners be off is a welcome addition. After putting some serious thought into the chosen mechanism I'll be using I opted for a curved CAM, as you'll see in the photos. Choosing this over the other options which include a CAM above the figures, and a similar curved CAM at the top, this design involves less, which of course means there is less to go wrong.

Having only drawn up these designs, I needed to put it to physical test so today I focused on a working mock up. The initial design is incredibly straightforward, a curved/hilly CAM sits underneath the rotating base which has poles running through it which meet the CAM. As the base rotates these poles are in turn rotated and move up and down according to the height of the CAM underneath it at that time, as seen in the photos: the curved CAM was lasercut and then heatbent around a former of a similar circumference (a mistake in finding the circumference of the circle resulted in the piece coming up short, this will be fixed tomorrow):

The rotating structure on top which houses the poles and figures:

The extra rings at the base keep the poles on the CAM so that they don't slide off either side, a new design will be looked into to fix this as these rings sometimes lipped the pole resulting in it not following the CAM as seen in the two following photos, firstly it working how it should:

The pole becoming stuck at a height:

One way to prevent the pole from sliding off either side is to attach a secondary base disc which fits tightly around the centre point (the larger pole), as you can see in the above photos the the middle gap in the base disc is a lot larger. The new disc is the black one:

Currently this mock up relies on me turning it for it to work, of course on the final one it will be powered by a motor. To get the ideal rotations per minute I've had to buy a gearbox motor, sadly the one I brought isn't powerful enough and I'm waiting for the larger model (which RMP is dependent on voltage) to be delivered, For now though this motor will do and allows me to design ways of how to connect it to the centre pole which rotates the entire thing. The circuit design is final, just a straightforward power source, switch and motor: