Re: hacking a computer touchpad to return 2 resistances?

On 2007-05-18, Paul <nospam (AT) needed (DOT) com> wrote:

Your first step, is finding out the data format coming out on the
PS/2 connector. Convert the output byte codes for the X and Y
values, and store them in parallel registers. Connect the registers
to digital to analog converters. The output at this point is a
voltage. The voltage may be suitable for driving something
directly (with a little buffering, say).

I was thinking along much the same lines. All the documentation
you'd need is readily available on the web - I remember looking it
up a while back when I did a feasibility study into a PS/2<->ADB
adapter (quite doable).

However, the thought has just occurred to me that this won't work
as the OP wants. A mouse is a relative input device - it doesn't
care where it is, just what direction and how fast it is moving.
Those touchpads work in the same way for compatibility - if you
put your finger in the top right corner your pointer doesn't
automatically jump there on screen. What the OP wanted was that
kind of absolute behaviour which you can't get at without pulling
the device to bits at replacing its brains.

That isn't straightforward and I would guess from the style of the
question it is massively beyond the OP's capabilities. In any case
it means the problem is slightly less well-formed. What is the
resistance when the finger is absent from the touchpad?

I think a joystick with potentiometers on the X and Y axis, is
a *lot* less work...

I'd certainly agree. Something along the lines of
http://tinyurl.com/224ecq would do the job with minimal effort.
If something flush to the panel was needed I'd be thinking in terms
of four pushbuttons in an arrow configuration. Use a bit of extra
circutry to gain analog outputs that increase/decrease with each
button press.

I'd also be interested to know exactly what the OP is actually
trying to do. I suspect there's a better way of doing whatever it
is.

--
Andrew Smallshaw
andrews (AT) sdf (DOT) lonestar.org

In message <slrnf4s508.bkv.andrews (AT) sdf (DOT) lonestar.org> Andrew Smallshaw
<andrews (AT) sdf (DOT) lonestar.org> wrote:

Those touchpads work in the same way for compatibility - if you
put your finger in the top right corner your pointer doesn't
automatically jump there on screen. What the OP wanted was that
kind of absolute behaviour which you can't get at without pulling
the device to bits at replacing its brains.

That actually isn't the case for most modern touchpads. Check out
Synaptics' software line, which actually works on most touchpads today,
you can get a pretty graphic of where you are touching at any time.

In terms of how it appears on screen, you are correct, but the hardware
gives the location information to the computer, and the computer
translates the absolute positioning into relative movement.

--
"These, and other cliches will be available to you all for one more
day of training, with me"
-- Jack O'Neill

On 2007-05-18, DevilsPGD <spam_narf_spam (AT) crazyhat (DOT) net> wrote:

In message <slrnf4s508.bkv.andrews (AT) sdf (DOT) lonestar.org> Andrew Smallshaw
andrews (AT) sdf (DOT) lonestar.org> wrote:

Those touchpads work in the same way for compatibility - if you
put your finger in the top right corner your pointer doesn't
automatically jump there on screen. What the OP wanted was that
kind of absolute behaviour which you can't get at without pulling
the device to bits at replacing its brains.

That actually isn't the case for most modern touchpads. Check out
Synaptics' software line, which actually works on most touchpads today,
you can get a pretty graphic of where you are touching at any time.

I considered this possibility before I posted but dismissed it.
A PS/2 style touchpad (we can dismiss USB devices out of hand on
complexity grounds) will start up is PS/2 compatibilty mode, i.e.
relative positioning. The PS/2 standard will be supported by all
devices and is well documented. To get absolute positioning, the
device will need putting into a special (non-standard) mode and I
doubt you'll be able to get hold of the docs.

For a well-known brand name it's a possibility, but not for commodity
hardware. Since you mentioned Synaptics, I checked out the X.org
driver for those. There is a special one for them which gives
access to extended modes, but it lacks an absolute positioning
mode. Whether that's because the authors couldn't get the details
or because of the limited resolution of the device I haven't looked
into.

In terms of how it appears on screen, you are correct, but the hardware
gives the location information to the computer, and the computer
translates the absolute positioning into relative movement.

Not by default. As I've already mentioned, they start up in PS/2
mouse compatibilty mode. That's strictly a relative-movement
protocol.

--
Andrew Smallshaw
andrews (AT) sdf (DOT) lonestar.org