Morgan, another option is to use a 74C922 keyboard encoder. While it's
more expensive, it only requires two capacitors for debounce and scan
frequency. It's trivial to interface to a PIC with 6 lines which includes
a keypress strobe and an output-enable for bus applications. I've used it
in many projects.

   - Tom

At 12:58 AM 1/27/98 +0100, you wrote:
>Since I in different occasions had this question about how to simulate a
>keypress on an existing keyboard, I decided to post the answer to all.
>
>
>Original Question from andre@compufire.com: "using pic to control a vcr"
>>Hi to all engineers in this list.
>-snip-
>>I decided to drive them
>>from switches like pressing buttons. switches are connected together
>>in multiplexed method that means each switch when you press it
>>doesn't ground the only access i have is switches. my question is
>>to press a button electronically should i use optoisolater or can I use
>>electronic analog switch like 4066. any recommendations ?
>>
>>Andre                         thank you very much.>
>>==========================================
>>= http://www.compufire.com               =
>>= mailto:andre@compufire.com             =
>>= mailto:mcu-engineering@compufire.com   =
>>= Andre Abelian: Engine Electronics, Inc.=
>>= Tel 909-589-5485  Fax 909-598-5695     =
>>==========================================
>
>Cheapest (i belive) is to use a transistor resistor and a diode for each
>button. 
>
>What a coincidence: I actually today built a remote controlled PC keyboard
>(15 buttons) with this technology:
>
>The keyboard controller scans the keyboard by asserting one row at a time
>ti GND, others to Vdd. The horizontal lines it pulls high.
>
>For each button to be emulated I connected a NPN transistor«s emitter to
>the key«s scan row, and the collector via a diode to horizontal line.
>(diode kathode to collector) I drive the transistor with logic level from
>IR decoder via a 100 kohm R.
>
>>what is the idea of using a diode in each switch?
>
>If, when this transistor«s emitter scan line is high, and another key
>sharing the same feedback line as the transistor«s collector is pressed, it
>will sink the collector voltage. 
>
>If you are then holding the base high the transistor will lead current from
>emitter (now positive) to collector.  As the keys here have some high
>resistnace (rubber-type) the transistor will win and the keystroke not
>detected. 
>
>Anyway, in my application it would not matter, since only one transistor or
> switch is activated at the same time, but I want to make it reliable,
>since I will sell it. (For controlling engine test equipment when sitting
>in car.)
>
>Also beware that when your input to the transistor is high, the NPN
>transistor will most of the time have a negative Vbe. Almost any transistor
>can take -5V, but I don«t know about -Vcb, (which we will have if we don't
>use diode)
>
>/Morgan
>Morgan Olsson, MORGANS REGLERTEKNIK, Sweden, ph: +46 (0)414 70741; fax 70331
>-
>
>