Erratic behaviour joystick button on soundblaster

Posted by cyberfreak on August 21, 2008, 12:45 pm

Hi everybody,

I just installed an old PC that has a sound card(isa with sound
card+joystick interface) that i use as an automation device(programmed
in Pascal).
What it does is read for many hours the joystick port for buttons
pressed(a digital sensor connected to a relay closes the circuit and
generates the event).
The problem is that is generating erratic button presses( with
absolutely no aparent
reason - it works for days with no problem then it fails) - it might
be the sensor who sends the wrong signal..electric noise...but
i believe the problem is on the side of the computer.
What should i do - might it be a defective sound blaster compatible
card?

The sensor is trying to sense a metal(the metal is glued to a 35mm
cinema film)..when
it senses the metal(very rapidly..in a fraction of a second) it
activates a "relay" that closes
the circuit for the joystick button generating an event.

The code i use to read the bit In pascal:

FUNCTION button_1 : Byte;
BEGIN
button_1 := ((Port[$201] AND $10) XOr $10) ShR 4;
END;

Isn't there a way to force the event to last longer than the time
the relay is active to allow me to diferentiate between real events
and hipothetical electricity contamination?

Some sort of a timer..everything happens very quickly...

Any help would be welcome.
Thank You.

Posted by Gordon McComb on August 21, 2008, 1:33 pm

cyberfreak wrote:

> Isn't there a way to force the event to last longer than the time
> the relay is active to allow me to diferentiate between real events
> and hipothetical electricity contamination?

Google "switch debouncer" and you should find a number of circuits for
this type of thing. If you are getting fleeting noise, something like a
Schmitt trigger on the input may help. Again, a G search will turn these
up. Basically this type of gate only responds to voltages above a
certain level.

In the modern world all of this can be programmed with a simple
microcontroller. That allows you to play more with the values, just in
software.

The old game port on a sound card or a stand-alone card is not
instrumentation grade. If adding the noise filtering to the switch
inputs doesn't help you should consider a better interface. You can buy
these for ISA and PCI slots for maybe $200 (some are more, with added
features), ready to go. Depending on how critical it is to sense the
foil mounted onto the film this small investment could be well worth it.

-- Gordon

Posted by Si Ballenger on August 21, 2008, 5:04 pm

On Thu, 21 Aug 2008 09:45:18 -0700 (PDT), cyberfreak

>Isn't there a way to force the event to last longer than the time
>the relay is active to allow me to diferentiate between real events
>and hipothetical electricity contamination?
>Some sort of a timer..everything happens very quickly...

Searching google finds sites like below. An NPN transistor could
be used in place of the relay to do the joystick button push. You
could do something similar with the serial port. A voltage blip
on the rx line will leave jibberish in the input buffer, which
can be read on a repeating basis by software.

http://home.cogeco.ca/~rpaisley4/LM555.html#29

Subject
Posted In
Date

Replace an inductive joystick with a micro controller?
General Robotics Forum
2006-11-07

Machina Speculatrix and Emergent Behaviour
General Robotics Forum
2005-12-22

Watson / Jeopardy 'robot', whats so cool?
General Robotics Forum
2011-02-16

MindForth Programming Journal (MFPJ) 2010 August 30
General Robotics Forum
2010-10-14