A dining table with a touchscreen computer in it always seemed kind of a good idea – I sketched a coffee table with a touch screen in it a while ago. Dining tables are where you sit and read the paper when you eat your breakfast. Tables are what you sit around and share food, ideas and discussion. I thought an big central interactive screen would add to this experience.
Anyway, that was all back when touch devices – their software and hardware – were pretty embryonic (Palm, Psion etc) so I didn’t really dwell on it. The operating systems didn’t really exist, I am not sure ‘flat screen’ TVs existed and touch screens were resistive and the size of a stamp.
The original Microsoft surface the big Samsung table one got me thinking about it again and the following spurred me to get started:
- A broken Daewoo (Yes, they made TVs) 42″ TV at home that became a spare when it broke, we got a new one and then I replaced some components in the Daewoo and it sprung back to life.
- I saw a Microsoft Surface (the old one pre iPads, they are called Microsoft PixelSense now and they seem to have died a death in 2012/3) at uni and being a little underwhelmed. It was just a big touch screen computer with legs (not a useful table so kind of missed the point). I saw one at the Microsoft New zealand headquarters in a coffee table too, but it was a novelty item rather than something useful.
- Access to cheap CNC: I was working in the architecture dept at Auckland university at the time.
The design is intended as a ‘family dining table’ (though it could I suppose be used as a meeting table in an office). It seats four very comfortably. It has with a 42″ touchscreen in the middle of it linked to computer. The screen (and all its gubbins) spins and locks to make a ‘regular’ table top. I thought this was a nice touch for when you’re having dinner and don’t what a great 42″ screen glaring at you when you want to put your roast.
17mm ply for most part. Just used the lowest of the low with one good face and one rough. Think I ended up buying three or four sheets of 1.2 x 2.4m ply. No veneer or anything which was mostly cost and ignorance driven. For the support frame (legs etc) this is triple laminated to make a nominally 50mm wide structure – stiff enough. Probably overly so, that is the structural engineer in me. Table is as sturdy as an oak though.
Screen (and cooling fans)
I used a newly bought 42” Panasonic screen (the original Daewoo screen I mended bit the bullet and was only ‘HDReady’ rather than full HD. 720 vs 1080 and windows 8 does not play with HDReady.
I took plastic case off the TV and mounted it in a new timber case. The original plastic case had vents in the top and bottom and I imagine air is drawn through these by a stack effect when TV is operating upright, cooling the internal componentry. This effect is lost if the TV is horizontal. To remedy this I used three 5V computer fans (wired in parallel) to mechanically draw air through the TV box. Usefully TV USB port only gets power when the TV is ‘on’ so I used this to drive a relay which turns on the circuit to fans. The fans operate on an independant power supply (good quality phone charger) because the TV USB port, though 5V, is underpowered has a really low current, not enough to drive the three fans adequately. The fans are loud unfortunately and they do blow cold air into your nether regions but they are piece of mind. The screen does not heat up at all when using the table.
Touch Screen Overlay
A 42” infrared 3-point touch device with black alu anodised bevel. This is the weak-link in the whole system. Imported it into New Zealand from an outfit in Korea. It cost a fair bit – I specified toughened glass to make it toppling-over-bottle-of-wine-proof.. From my googling large capacitance touch overlays, the same technology as in smartphones and tablets, are prohibitively expensive. The overlay I have is infrared – it bounces light into the glass using internal reflection. The reflectance angle changes slightly when you touch the glass and this is measured by sensors. It is VERY sensitive and detect finger touches and crumb touches. Unfortunately Windows 8 puts a lot of emphasis on edge gestures (swipes in from edge). Historically manufacturers (including the one for my overlay) put their efforts into sensitivity in the screen centre, not the edges so my overlay falters in this regard. Also, the overlay has a bevel that sits proud of the glass so this does anyway with ‘edge swipes’ which are a key functionality in Windows 8. I do have plans afoot to remedy this.
The overlay is essentially a plug-and-play HID device.
Mains power runs up inside one table legs to a four-way power thingy which is mounted to underside of table. Plugs for computer, TV and cooling fans plug into this and go their separate ways.
Modelling/ Design Software
Mostly used SketchUp for the 3D stuff. I know it’s simple but it’s great. Its is the pencil and paper of computer modelling software. I worked with an architect once who also tutored at an some Architecture school or other who sent students home if they turned up with a Sketchup model. Boo to him. Sent them home if they are wearing glasses with no lenses in I say or have a wooly hat on I say (I tutored too).
Rhino for arranging cut patterns on ply 1200×2400 sheets efficiently and exporting. Rhino seems to clean up in terms of file formats it can import and export. I thought about writing a script to do this laying out but a task which humans do on a whim are a Phd in computer science.
I made two small scale mocked from cardboard – I got them laser cut. Very cheap and very useful. Main lesson I learnt? ‘Do (two) mockups’. Its essentially analogous to ‘measure tiwce cut once’. Would have been a costly f”!k up it I had gone straight to full scale.
Mockup #1: Caused major design changes in how legs and support frame came together. What I was working to before had no lateral stiffness. School boy error.
Mockup #2: I originally had the table top corners as sharp 90 degree angles. This was the second-rate engineer in me: Maximise table area. Maximise plywood use. I decided to round off the corner in finished product. It meant profanities were not wasted on every time I walked into a sharp table corner and I can hence mutter saved swear words at people in sunglasses or with moustaches instead. so unnecessary. No. you do not look cool. No you do not look like a beautiful celebrity or sophisticated cad. You look like a person wearing sunglasses when the sun is not out (i,e. an idiot or a blind person) or a friend of Jimmy Saville’s in the case of the facial hair. Table corners do look nicer too but I regret my choice a curves. I chose a quadrant of an ellipse, could have gone within something a little smoother. Any-who. Food for thought if you ever need a curve to transition for one straight line to another. The palette is wide – B-splines, Beziers, clothoids, circles , eclipses, trigometric functions – and they are all different.
The models also served as useful tool to show paint-job people what they were working on and what needed to be painted what colour/ finish. .
Construction and Key Lessons Learnt
Gluing (a lot of it!)
In total table composed of around 80 timber components.
The construction generally was from 17mm ply laminated together,for example 3 layers for the legs and main framing. The original intention (what is the difference between intent and intention? Need to look that up) was that pretty much all joints were to be notched/ sliding with no mechanical fixings or glue, like the Ikea furniture and a Shinto shrine. Ended up gluing EVERYTHING. Actually ended up reinforcing with epoxy and steel (bits broke when I was hammering together becaouse of poor fit from glue thickness).
What really caused my problems was that I failed to take into account the thickness of the glue layer when home-laminating. I took very careful account of tolerance on the ply thickness in my joints but assumed the glue thickness would be negligible. It is not. Sure you clamp the death of of the bits when laminating them together but the glue still retains a finite thickness. This throws all your carefully crafted Shinto joints out of kilter.Nothing a mallet didn’t fix.
My main concern was potential complexity in the getting power and wiring across the rotating axis which allows TV screen and its gubbins to be spun over (the party trick for the whole TV). These wires are: the power to the TV, the power to the cooling fans, the USB cable for the touch overlay and the HDMI cable. In the end just used a galv steel tube (~30mm dia found on the roadside – thing it was a part of a swing originally) as an axle and ran wires inside it.
Really simple and works super smoooooth with no bearings or anything. So long as one more or less rotate screen the same number of ways one was as the other I imagine that twisting of the wires should not manifest as an issue.
On/ Off Switch
Another issue was how the TV would be turned on and off. In a perfect world there would be a single switch that powered on the TV and the computer but settled for second best and remounted the TV buttons (the one that are typically hidden around the side on flat screen TVs) on the top of the table. Handmade and sprayed a little aluminium surround to plastic console buttons are mounted on. Turned out real nice in a 80s consumer electronics sort of way.
The TwistTable (like it?) been in operation for over a year now. Its pretty good. As a table it is great, though perhaps a little agricultural in its styling but big and comfortable. Its sturdy and the rotating ‘mechanism’ works perfectly.
Nevertheless the system really is let down by the touch screen, windows 8 and Sony laptop combo I have running. Have looked into running Android and ‘something’ on a raspberry pi2 but both underwhelming so current combo is best of a bad bunch. Dont get me wrong, its very useable but erratic and a little clunky. what could I have done differently in the design to remedy this? Not a lot at the time – wait for Windows 10? Wait for affordable capacitive large touch screens?