Because of the way the chip handles the signals when the switches are closed, it is very difficult to get the right character with one key press, for example to get '1' we have up then left with up still held, to simulate this with all three signals connected to one switch gives the result '.' this is not the result we want, it is clear that a delay is needed for the chip to understand that I wanted a one and not a stop. There are ways around this, and I look forward to contributions to articulate this, I had considered 555 timers, transistors or very low resistance capacitors.
I adopted a method that brought the shift key and the action key together, The thinking was to place both switches close together on above the other soldered on a strip board, to ensure a real-time delay is achieved between both keys I selected two heights of tactile switch, the upper of the two to be in contact with the finger first, shortly after the second switch is closed by the rest on the operators digit.
There are a few differentials that you can select, I chose a difference of 0.2 mm - it worked for me.
2x 130 x 95 mm strip board. ( go bigger if you can )
drill bit for breaking tracks and a stanley knife for clearing and cutting valleys.
coloured wire 4m lengths ( you will use more purple than yellow) this was available in a fun bag. colours
36 RVFM tactile switch 6x6mm 2.4 rapid electronics part # 78 - 3812
36 RVFM tactile switch 6x6mm 2.8 rapid electronics part ~ 78 - 3806
8 tactile switches of you choice for the shift keys.
Because of the size of the board I used and the spacing on the switches I was very limited from the outset to where to attach my wires without causing shorts on the switches. I adopted to break the track where the entry wire was to go, and strip enough wire to go through the hole and loop over to the leg of the switch, all the other connections snake from the entry connections, an off cut of strip board is used to couple up all the entry wires of the same colour and then one is taken of that to the underside of the fignition computer.
here are the diagrams of how I connected each switch.
and here is how it looked during construction.
fitting it in the C16 case I used for the first attempt.