[sdiy] Semi-OT: PIC trouble

[Fredrik Carlqvist]

I noticed the display pulls a lot of current (>100mA), very much
depending on the number of pixels lit. Does the Vcc on the PIC change
when you pull the HDLG? 

If the PIC has brown-out reset enabled, it could reset as soon as the
voltage falls below for example 4.5V (voltage set by configuration
bits). Remember that a 100ns glitch could be enough (although the data
sheet states 200us). This could make the LED blink irregularly when the
display is in-circuit. 

Even if brown-out reset is not enabled, power line problems could
influence your RC-oscillator. Maybe a lower voltage will push the
frequency above the max freq for the chip, since you are normally on
48MHz?

Since the display unit has no outputs and the inputs are CMOS, I see no
way for it to influence the PIC on a logical/interface level. The only
possible communication from display to PIC is via the power line. 
 
Separate the power lines all the way to the PSU and add capacitors at
both chip Vcc/GND until the problem goes away :-) If that does not help,
try a crystal oscillator.

Some PIC18 derivatives have a problem with the interrupt consistency,
the shadow registers does not behave well. PIC18F2455 does not seem to
be one of them though. 


Fredrik C



-----Original Message-----
From: owner-synth-diy at dropmix.xs4all.nl
[mailto:owner-synth-diy at dropmix.xs4all.nl] On Behalf Of Karl Ekdahl
Sent: den 31 oktober 2005 22:56
To: synth-diy at dropmix.xs4all.nl
Subject: [sdiy] Semi-OT: PIC trouble

Hi list, i'm having trouble interfacing a HDLG-2416
(http://www.datasheetcatalog.com/datasheets_pdf/H/D/L/G/HDLG-2416.shtml)
with a PIC18F2455. I seem to get some feedback from the HDLG back to the
pic that makes the oscillator unstable, i have a eternal loop flashing a
LED and it's *not* flashing in a perfectly sequencial manner.
I have a 2.4A PSU and the HDLG is connected via two 4094's so i don't
really understand what's going on. I have verified that it's really the
HDLG by simply pulling it out of its socket and observing the LED.
Though since the pic is fast, i do run it on 48Mhz (oscillator, not
crystal/resonator) which i guess might be a problem while not using 1st
grade components....

hrmm.. now when i come to think about it, haven't people here been
talking about resonance/noise in LEDs?

Ideas (or directions to a good pic forum) highly appreciated!

Karl