PS2 keyboard interface to touch sensors.

If you type "PS/2 Keyboard" into a search engine like Google, you get a great deal of information. What information are you looking for that is not provided by the search engine results?
-Wayne

Google is very good at generic information, but rather poorly at specifics, unless you are lucky.
Also, trying to start a discussion that could be helpful to people who may also be attempting this.

Let me *repeat* the question -- what *specific* information are you looking for that was not found in all of the *generic* information that the search engine returned?
-Wayne

Well, there are a number of posibilities. The search phrase: "ps2 keyboard interface touch sensors" does't return anything usefull in the realm of rebotics.
If someone has done this I would like to hear about it.

PS2 gives serial data. You have 101 keys or more. Problem is that keys are stuck on matrix, and matrix is two plastic sheets with traces on them. The keys push the plastic sheets together, connecting traces, and the chip in the keyboard turns the signals into serial data..
If you can find a PS2 keyboard with microswitches it would be a great touch sensor. You could take each switch out, solder wires to them, and mount them wherever you like.
Commodore 64 keyboards have microswitches, I think Apple II comuters do too. Problem with them is they can only detect 1 key down with the keyboard encoder chip.
I think it is a good idea to use ps2 keyboard as touch sensor if you want to put 4 of them on your bot, one on each side.
You cannot easily seperate 1 PS2 keyboard into many switches. The time it would take would outweigh money saved.
The keyboard is a touch sensor already. It senses which fingers are touching firm enough to actuate the matrix.
Rich

yes, I'm sure we've all seen one ok true, but the PC keyboard has some keys that assert meta data. The shift, ctrl, alt, and maybe other keys assert.
I'm not sure if I agree. There are a number keys that that are not linked directly to other keys ( shft, ctrl, alt

[snip]
Indeed they do.
If you had bothered read the first link that is returned when you Google "PS/2 Keyboard" you would already know the answer to this question. Why don't you take a few moments to read up on the subject?
[snip]
-Wayne

Actually, that's not how it works. Each key press generates several codes which are transmitted serially to the motherboard. Conversely, each key release also generates a sequence. You get "make" codes and "break" codes for each key. Shift, alt, control, etc are just another key which generates their own unique sequence. Even caps lock are just regular keys. The driver is actually responsible for turning on the caps lock led by sending a command back down to the keyboard controller - they keyboard itself doesn't do it on its own.
The driver has to decode the key input data stream to figure out which key was pressed, whether shift is being held down, etc. Also, not all keys generate the same number of bytes on press and release. It's not as simple as one would think on first glance.
-Brian

Cool. Like I said, I have never done any keyboard work. Some of the research I did to find what I needed to implement the mouse encoder mentioned the basic communication similarities of the keyboard. Is anything?
So, it should be possible to dremel apart a PS/2 keyboard and wire bumper switches to the various matrix connections. I'll rip one apart this weekend and take a look and do some experimentation.

Someone asked why I bother to post, my reason is that I am building my robot. I am building my robot in a way that *anyone* interested in building one should be able to, with a limited budget but a reasonable skill set, build it out of commonly available parts. It is targeted toward a high school student level.
I am posting questions and information as part of this. It is good to get opinions of others, not nessisarily that you agree, but to hear concerns or facts as they relate to your question. Sort of "bouncing" the idea off the group. Isn't *that* what this news group is supposed to be about?
If someone is interested in building a robot themselves, they may come to this group or search google. They may find this and other discussions and the dialogs may be helpfull.
Certainly Brian Dean's response was helpfull.

Good luck! I've tried messing with them (IE soldering to them, cutting them up) and ended up very frustrated.
Rich

Or second glance, for that matter:
Many keyboard designs are capacitive, detecting keypresses by the increase in capacitance at that crossing in the matrix, rather than physical (electrical) contact. Direct connection of a switch may not work, or one may work but several activated at once, especially if they share a common matrix line, may not work the same as actual keyboard presses.
-----

[snip]
To Brian Dean's credit, he took the time to respond with information that is readily available on the net.
Just like you assume that people know what I2C means, I assume that you know how to do basic research > If you type "PS/2 Keyboard" into a search engine like Google, you would take a few seconds from your schedule to go to Google, type "PS/2 Keyboard" and look at the results.
Here's how it works:
Step 1: You go to Google.Com at:

Step 2: You type "PS/2 Keyboard" into the search window and click on the [Search] button.
Step 3: Google responds with a page of search results. The first result is to Adam's web page at:

Step 4: You click on Adams web page it brings you to a redirect page at:

Step 5: This brings you to a page of links:
PS/2 Mouse/Keyboard Protocol
PS/2 Keyboard Interface
PS/2 Mouse Interface
and a couple of translation links.
Step 6: Click on the first link and it brings you to:

which is a very complete description of the electrical interface to the PS/2 keyboard including connector pin-outs and signal timings.
Step 7: Click on the [Back] button to go back to step 5.
Step 8: Click on the second link and it brings you to:

which is a very complete discussion of the keyboard software issues. The discussion on scan codes occurs about 15% of the way into the document.
Brian's message is a synopsis of the information presented in Adams's document.
In short, you can use any key on a PS/2 keyboard except the Pause/Break key as a touch sensor. (You'll have to read the document to find out why the Pause/Break key doesn't work.)
For further comments on this thread I am going to *assume* that you have *read* Adam's two web pages on the topic.
-Wayne

mlw wrote: High school has advanced since my time! So now they are experts in electronics, operating systems and C++?
As I wrote before, we all enter this hobby at a different skill levels. Your level with respect to programming environments is high. If you don't provide an interface with your programs to a lower level of programming skill your software will be limited in accessibility for modification.
Any group is about whatever those in it at the time decide it is about, it is democratic by nature.
Sometimes it's quicker and easier to ask a question than to search google. If someone has already been there, done that, they can give a faster and better response than google. Google cannot compete with the human brain as a search engine. One day, maybe, with a bit of built in AI and knowledge of the needs and skill level of the person using it a search engine will do better. Wouldn't that be nice, an AI expert that can respond to your needs without telling you to go search Google :)
It is a bit like an expert refusing to answer a question until the person has done the hard yards in the subject as they have done. Perhaps accuse them of not wanting to learn :)
If someone asks me how a TV works I don't tell them to go do an electronics course, I give them an answer commensurate with their knowledge base.
Also it can sometimes be just as easy to give the answer as to tell them to go ask the Google God.
Search engines are great. But people are greater. And so can their books on a subject.
And even if we are on different wave lengths it can be motivating. After all your business about building your PID I was motivated to give my motors PWM speed control and design a suitable encoder.
- John

Well, not experts, obviously. If it is done right, they should be able to build the basic system with the level of skill and expertise you would expect from a high school student taking electronics in school.
Then, it is just installing Linux. That may be a hard part. (they should have some challenges, I suppose.) Then install my programs and run "robot." At that point it should be usable from any language via a web services interface.
If they want to do more, they can delve into the specifics of the program.
I am thinking that each function, motor control, object avoidence, speech, will all be implemened in a UNIX shared library (like a Windows DLL) and spun off into its own thread.
If you take a look at my code (I posted some proof of concept code that maintains speed via PID) Actually, what a usenet group is supposed to be about is described in its charter.
[snip]
I see no point in responding "go search google." I'm sure we've all done it, but it isn't really very polite. One assumes they've searched using some engine, but if someone knows something and is willing to contribute a few minutes of their time, why not ask?
Even though I search google and alltheweb, I like to pose the topic to others for a number of reasons. Not the least of which is it lets other people benefit from the process.
Very cool, that's what makes it fun.

look around they have articles and discussions relating to interfacing buttons to the PS/2 keyboard port, both through hacked keyboards and through boards designed for the purpose.
If you're up to serious microcontroller programming to roll your own interface, is a good reference.

... You might get lucky - I've had some apart that have the controller on a small PCB that goes to a connector into which the "printed" cables of the key matrix plug. I reckon that if you had one of those and desoldered the connector (after tracing the matrix!), you could hook a ribbon cable to it to a more "friendly" connector, say a bit of nylon terminal block (connblock).