PIC is not working, sometimes working...

denkid · Joined: 17 Jan 2005 Posts: 10

Hello.

Thank you for reading.
My country is not use english.
So, My english might be very poor, sorry.

I made many Target Board using PIC18F6520
but, 10% is right.
left 90% is not working.

I think it is problem of OSC.
But it is just my thinking.

I am gonna show the pcb picture

I use paste when I solder this pcb myself.
And, I attached 8 LEDs
Sometimes, 2 LEDs is same working on following codes

Ttelmah · Guest

From what you show, the oscillator should not be a problem. What the picture is not quite large enough to show, is the decoupling on the Vss line (hopefully the parts on the edge of the picture a decoupling capacitors).
On the oscillator, what is the crystal, make, type, frequency?.
What is not working?.
How is the supply generated?.
You have a very large number of 'NC' pins. Are you programming these as outputs on bootup?. If not, some 'odd' effects can result.

Best Wishes

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

Your ground connections are not very good. Each Vss pin should have
a very short trace that goes to its own via, which goes to the ground
plane. The Vdd connections should also be kept short, and each Vdd
pad should have its own 0603 capacitor (100 nF) to ground.
However, I don't think your layout is so bad that it will cause 90% of
the boards to fail. They should still work.

1. Post the schematic of your MCLR circuit.

2. Post your #fuses statement and #use delay() statements.

3. Post your compiler version.

denkid · Joined: 17 Jan 2005 Posts: 10

Thankyou for your answer.

I posted the picture of schematic

Thankyou

SET · Joined: 15 Nov 2005 Posts: 161 Location: Glasgow, UK

MCLR has 10k pullup which should be fine. The grounds are bad, also the supply decouplers (which you can just see one end of) must ground even further away - whether this is bad enough to cause malfunction or not I'm not sure.

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

Please post these things:

2. Post your #fuses statement and #use delay() statements.

3. Post your compiler version.

denkid · Joined: 17 Jan 2005 Posts: 10

1. fuses configuration

Oscillator : HS
Osc.Switch Enable : Enabled
Power up Timer : Disabled
Brown out Detect : Disabled
Brown out Voltage : 2.5V
Watch Dog Timer : Disabled - Controlled by SWDTEN bit
Watch Dog Postscaler : 1: 128
CCP2 MUX : RC1
Stack OverFlow reset : Enabled
Low Voltage program : Disabled
Any Code Protect is Disabled

8Mhz OSC is used.
#use delay(clock=8000000)
I use ICD2 Programmer

2. Compiler version : 3.249

FULL SOURCE

Humberto · Posted: Wed Nov 21, 2007 1:40 pm

denkid · Joined: 17 Jan 2005 Posts: 10

right side of Vdd and Vss pins are near the OSC

The pins is this.

Vss
OSC1
OSC2
Vdd

So, How can I connect Vdd and Vss Shortly.
Because, It is placed with OSC1,2 very near

As you can see the PCB figure.

I think, if The Vss, Vdd should be shortly then the OSC parts must be placed bottom layer.

Thankyou so much

SherpaDoug · Joined: 07 Sep 2003 Posts: 1640 Location: Cape Cod Mass USA

I don't like the grounding of the upper 22pF cap on the crystal. It connects to ground plane, but how does that ground plane connect to the GND of the PIC? That whole loop from PIC pin through the cap and back to the PIC should be very short. I would like to see both caps grounded to a point between them.
_________________
The search for better is endless. Instead simply find very good and get the job done.

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

Try the program shown below. Your program is doing read-modify-write
operations directly on the Port E pins. The program below will do it
to the Latch register. This prevents any possible RMW problems.

Also, you are manually setting the Config bits in MPLAB, which could
easily allow some mistake to be made. The program below embeds
the fuse settings in the program. That's much better.

denkid · Joined: 17 Jan 2005 Posts: 10

Thank you so much.

I got much information from you.

Thank you

denkid · Joined: 17 Jan 2005 Posts: 10

[img] http://www.mindcolor.com/pcb3.jpg [/img]

Thank you for your help

I have modifed my pcb following your opinion.
(please click upper link to see the figure of pcb)

Please watch my pcb, again.

Have I still problem?

And, some Target Board seems to be right for moment.

but, The Target Board will be broken soon,

LED is very strange working.

I can not asume that working

Thankyou so much

Best regard.

Wayne_ · Joined: 10 Oct 2007 Posts: 681

You say 10% of boards are working and 90% not ?

How many boards have you made ?

As some boards work and others not answer the following:-
do the boards that work ALWAYS work ?

If YES then it is most likely a fault in the assembly of the boards, either bad soldering or faulty components etc.

If the good boards also fail then it may be a design fault.

Wayne

Humberto · Posted: Thu Nov 22, 2007 9:37 am

It is evident that there are many doubts from different problems, even in the title
"PIC is not working, sometimes working..."
According to the info posted in this thread, I guess that the poster is still in a learning
stage, trying to know:
1) How to draw electrical schematic with a PCB design tool
2) How to draw the lay out using the same PCB design tool
3) How to handle an array of LEDs using an MCU
4) How to write C code to handle such LEDs
5) How to make the project using an oversized and complex MCU
6) How to make the project using a 2 face PCB plenty of through holes
7) How to be proficient soldering SMD at once.
8) How to make a pile of co$tly PCB´s and that once mounted and programmed all
the set of components, it works successful in the first attempt

In my opinion, this is not the way.
The key is to get all the improvements step-by-step, not all together.

To get the needed skill level to solve all the proposed tasks in the first attempt, first
you must burn a lot of MCU and PCB to gain experience.

My $0,02.

Humberto
PD: A famous Argentine boxer defined the experience like:
"a comb that gives the life when one have lost the hairs".