polling

newguy · Joined: 24 Jun 2004 Posts: 1911

When the PIC's pins are configured as inputs, their input impedance is nearly "infinite". Therefore, if you place some charge on the input by connecting a 5V wire, it will take a long, long time to dissipate.

Add pulldowns to port d. Maybe 10k to ground or so.

spudtheimpaler · Posted: Wed Mar 16, 2005 3:05 pm

DragonPIC · Joined: 11 Nov 2003 Posts: 118

Guest · Posted: Wed Mar 16, 2005 8:38 pm

As I said ealrier, I think if you are using a mechanical switch you will get bounce.... ! Wire are far worse... Do not use wires to make simulate the sensor input.

Do the correct thing. it's good to do it right at this stage, if you learn to be slopy then it will get you in the end when you donl;t need it to.

Make a sensor simlator, use another pic and debouce the heck out of a REAL set of small rocker switches and then have that pic output a CLEAN signal.... problem solved..
Good luck
hans w

Guest · Posted: Thu Mar 17, 2005 9:48 am

newguy · Joined: 24 Jun 2004 Posts: 1911

Yes, it will read it thousands of times/second, but that's not the point. Connecting a switch (i.e. simply pushing a button connected to the pic) can result in several "events", all very quickly. It feels like one push to you, but in fact it can be connecting/disconnecting several times, all in the space of a few milliseconds.

The pic is quick enough to read each of these bounces, and will react accordingly. Say that you have a button that will increment a value every time you push it. Sometimes it may only increment the variable once, but most of the time you'll find that one push (what feels like one push to you) will actually result in 2, 3 or 4 being added to the variable.

Taking a bare wire and connecting it to the pic by hand is WORSE. It will result in far more bounce than a regular switch will do. At the very least, please rig up some switches and hardwire them into the pic!

Guest · Posted: Thu Mar 17, 2005 11:25 am

Guest · Posted: Thu Mar 17, 2005 11:52 am

_spudtheimpaler · Guest

I've been working on it solidly since saturday :(

I understand there is best practice methods, and options like plugging into an rs232 (which i have virtually no experience in). But as Hans W points out, the code is virtually bomb proof. I found that adding the 10k resistors help a lot. i put them liberally on the input port and its a lot more stable. I still have voltage discharge problems on the input port, takes about half a second to discharge the voltage, but we are compensating for that in code.

I honestly don't believe bounce is the problem. Though i am doing some software checking on the input port. I think a combination of the discharge problem and another issue of input voltage. The sensors are putting out 3.71V and i thought that would be enough, but apparently its around about the threshold. I'm desperately hoping that 3.7 is enough, else i dont know, i'll have to try and mess around with ADC or comparitors or the like.

However, with simulated sensors the robot is now following as it should, so thank you very much! I shall keep the forum posted on progress ;)

Again, thanks,
Mitch

rwyoung · Joined: 12 Nov 2003 Posts: 563 Location: Lawrence, KS USA

_spudtheimpaler · Guest

rwyoung · Joined: 12 Nov 2003 Posts: 563 Location: Lawrence, KS USA

Also, unless you have a good reason not to use the PowerUp timer, go ahead and use PUT instead of NOPUT. If you are using in-circuit serial programming, the power up timer is generally a good thing to have running.
_________________
Rob Young
The Screw-Up Fairy may just visit you but he has crashed on my couch for the last month!

_spudtheimpaler · Guest

Ttelmah · Guest

It is not a 'programming timer', but a 'power up timer'. It holds the chip in reset, for a tiny fraction of a second after power is applied. Starting immediately is fine, if your power rail rises very fast, but with the charging times of capacitors, and resistive connections, there can be a problem if the chip starts, and the supply rail has not fully established itself. The PUT, is a 'good thing', for 99% of installations, and really should default to 'on'. It is an exceptional circuit that needs to start as quickly as possible, and ha been designed to do this, that would not use the PUT.

Best Wishes

WhistleBlower · Guest

Warning to spudtheimpaler... CHEATERS IN THE COMPETITION WILL BE OUTTED. WE HAVE YOUR NAME AND ARE GOING ABOVE YOUR HEAD TO CR. Attempts to gain external help for free WILL NOT BE TOLERATED. As per the specification document, you may gain assistance from other groups for an agreed fee.