Non buffered keyboard element

[me-nkr]

So the keyboard component is a sequential one and buffers the input right?

Is there a way to use that component to create a component which outputs the current key being held or zero?

If not how can I have something like that?

Also having feature to have a custom key code map is also appreciated.

PS: I would like the component to give priority to key that was pressed last in case of multi key press, also here is the character set I want as of now

[mengstr]

When I need something like that it's usually just a hex-keypad or something similar. Those are easy to "build yourself" with buttons and a priority encoder. If you set a label on the button then it can be mapped to your computer keyboard. It might be a bit tedious to do this with a full ascii keyboard though ;-) and you'll have a problem handling shift and control keys.

keyb.dig.zip

[me-nkr]

out of topic how does the clock work like that? like the "wireless connection"?

[mengstr]

It just uses it. ;) The monoflop is a really annoying component since It can only be used if there is exactly one clock component present in the circuit. This clock component is used as time base to measure the time delay. and most of my a bit more advanced designs uses multiple clocks. So I tend to avoid it

[me-nkr]

what I was asking is there is a clock that is not connected to anything and there is no tunnel, how does that work? like is it a bug, is it a feature?

[mengstr]

The monostable doesn't have a direct input for the clock. That is probably why it requires exactly 1 clock no more, no less, in the design. It can then just look for the clock and use it internal in the software. The blue wires are just a human-visible representation of the "netlist" that the software is using internally while processing.

I have no idea why it was made that way, maybe so it would be a bit more similar to a real physical monoflop. They don't have a clock input either, but rather sets the delay in an analogue way using a resistor and capacitor. But since that is not possible here I guess it's a good middle way.

[hneemann]

That is mainly correct! The simulator knows no time. Only changes that take place one after the other or trigger each other. It is therefore not possible to simulate a component that triggers a change after a certain time - set with an R-C circuit. If you want to simulate a monoflop reproducibly, the time must come from somewhere. And the clock component in the circuit is used for this purpose. The delay time of the monoflop can then be determined in units of the clock period.

[hneemann]

What should happen if several buttons are pressed at the same time? Keyboards are normally connected in a matrix so that each key can be queried to determine whether it is currently pressed or not.

[me-nkr]

well in the case of nand2tetris, it's a bit ambiguous. But here's what I have made out

• there is no concept of multi key press
• when shift is pressed the alternate character of the key is used, but still there is no concept called shift, afaik
• when two keys are pressed together the last pressed key is chosen, if the last pressed key is shift, then shift + last pressed key to shift is chosen

[hneemann]

But is not everything in the Nand2Tetris project implemented by yourself? Then you can connect the keyboard however you like. And a computer game like Tetris can easily be implemented with the existing keyboard.
And if you need a control pad in which several buttons can be queried simultaneously, you can use the existing buttons.

[me-nkr]

In nand2tetris the scan code of button you press becomes the value of a memory address. once you release it the value resets to 0. For programs the value is queried every clock cycle. And I am trying to make the computer I make as true to the course as possible. Please feel free to mention anything I missed...