MPLAB, CCS Compiler, & PICKit 2

SW · Joined: 09 Jun 2011 Posts: 10

I've worked quite a bit with the CCS compilers and Microchip's PICKit 2 over the last couple years. I recently moved and am recreating the lab I used while I was in college.

I've run into a problem that I haven't been able to solve, and it's probably one of those simple ones that makes you smack your forehead when someone shows you your oversight.

I'm working with a 16F877A (40 pin) and a 16F876 (28 pin). I have my little "hello world" with flashing LEDs but when I compile through MPLAB (version 8.63) using the CCS addon, and use my PICKit 2 (version 2.61) to load the .hex on the chips nothing happens. Using only a logic analyzer I can see distinct pulses that appear to be something other than noise on PGD and PGC. My code for the 16f876 is as follows:

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

Put in the #fuses line as shown below.

SW · Joined: 09 Jun 2011 Posts: 10

I updated as you suggested and still have nothing. My rails are showing 4.9 volts and after checking all legs I show no voltage pulsing from any pin. Vdd shows hot as it should, and the two Vss show 0 as they should.

To dumb it down for myself, I have a 20MHz crystal (DigKey model 20.0UY6K) and two ceramic capacitors with 33 stamped on them which go to ground. Each LED has a 470 Ohm resistor connected to it. On the 16f876 I do not have an MCLR setup since the PICKit 2 is supposed to take care of that. I do have one set up on the 16f877a though. Neither setup works.

Any other suggestions?

Thank you.
_________________
-SW

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

1. Put a 10K resistor to +5v on the PIC's MCLR pin. (That's on pin 1).
That's part of the ICSP specification for Pickit 2.

2. What's your CCS compiler version ? It's given at the top of the .LST
file, which will be in your project directory. Example of version numbers:
http://www.ccsinfo.com/devices.php?page=versioninfo

3. Post your current complete test program, including all the modifications
that I suggested.

4. Your crystal number is probably what you see stamped on top ot it.
It's not found with Digikey's search engine. What's the Digikey part number ?

5. Is your crystal placed across pins 9 and 10 ? Are the 33pf caps
placed from pin 9 to ground, and pin 10 to ground ?

6. Are the two ground connections on the PIC to pins 8 and 19 ?
And the Vdd (+5v) to pin 20 ?

7. Are you testing the program in Debug mode or in Release mode ?
Try it in Release mode. This is selected with the drop-down box at the
top of the screen in MPLAB.

8. You can upload images to imageshack.com and post the links to them.

SW · Joined: 09 Jun 2011 Posts: 10

What are the odds that my PICKit 2 is bad? Nothing looks out of the ordinary, and it seems to verify the code after it is written without error. When it is programming the chip, the Busy light flashes like normal too. I'm out of ideas here. I don't know how I could have two bad chips at the same time when both are new.
_________________
-SW

SW · Joined: 09 Jun 2011 Posts: 10

Okay, going through my parts list again. The Digikey crystals I bought are P/N 887-1023-ND, or "Crystal 20.000 MHZ 18PF HC49US". Could it be the built-in capacitance that is messing me up?

Thanks,
_________________
-SW

Ttelmah · Joined: 11 Mar 2010 Posts: 19552

Not really.
The 18pF, is the _load capacitance_ expected by the crystal to work at the specified frequency. It'll still resonate with the wrong capacitance, _but your capacitance looks about right_. The 'load' seen by a crystal, is (assuming the two capacitors are the same value):

Cl = ((C^2)/2C)+Cs

Which simplifies to:

Cl= C/2 + Cs

Where Cl is the load capacitance, 'C' is the capacitor value you use, and 'Cs' is the stray capacitance on the board and PIC pins. Typical value for this is 5pF.

So, with your 33pF:

Cl = 33/2 +5 = 18.5pF

Just about spot on, assuming there is no silliness with the board tracks, giving extra stray capacitance.....

As PCM Programmer says, post the modified code with the mods made as he describe, and verify the connections he lists. Until the 'HS' fuse is used, the oscillator is _not_ enabled to run in the right mode.
Check the invoice for the capacitors, or what you ordered, or if they have any extra markings. Assuming they are 33pF, they should work fine, but if (for example), they were 33nF, there would be a problem. Also look over the board with a magnifying glass, looking for solder whiskers or blobs bridging pins.

Best Wishes

SW · Joined: 09 Jun 2011 Posts: 10

I've made the suggested corrections with the exception of the build configuration as it was not an option for me for some reason. I have also verified my Vdd, Vss, and LED connections. The pictures of my setup can be found here:
http://imageshack.us/g/215/bb002small.jpg/

The first five pictures are of a larger breadboard I was using when I posted my first message. The two after that are of a smaller one I used to rule out any possible problems with the larger breadboard. The active green LED just shows I have power from the PICKit 2 applied to the board. The last picture is a screenshot of the CCS setup in MPLAB, along with my PICKit 2 software showing a valid load.

The capacitors in question read "33" on them. I understand this value is in the form of pF. If that is not the case, that could be a problem.

My current code for the 16F876A is:

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

SW · Joined: 09 Jun 2011 Posts: 10

Yeah, I forgot to mention that. I switched red and blue. I know that's backwards, but I didn't want to go digging around to find my jumper box. The blue and red are switched depending on whether I use the larger breadboard or smaller one. I can confirm that pin 20 has 4.83 volts, which is odd since I have the PICKit 2 set to 5.0. I also have the PICKit 2 connected to a powered USB hub. As an added check, the greeen LED won't function if the voltage was reversed.

I installed the CCS MPLAB plugin as described. I now have a debug/release feature, however the problem still persists.
_________________
-SW

Ttelmah · Joined: 11 Mar 2010 Posts: 19552

Er.
Don't see how this can work. Might explain why it doesn't!....
On the 876, you seem to have a wire connection to pin 8, and the same signal going to pin 9, and pin 10 (three parallel red wires below the crystal, in the last layout). Assuming this is '0v', the one to pin 8 is rcorrect, but then the one to pin 10, is shorting out the crystal output, and the one to pin 9, is shorting the crystal input to ground. Internal gate will then be fighting itself, and won't work. The crystal capacitors seem to be running vertically down the contact strip, which will short them out completely....

As a comment, can I suggest you get some black wire, and use black for your 0v connections, red or white for for 5v connections (standardise on whichever you want), and then use other colours for your other connections.

Both layouts show exactly the same fault.

Best Wishes

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

Per Ttelmah's comment,
Here's a photo of a 28pin PIC in a breadboard, showing the correct
colors for power and ground, and the correct crystal and cap layout
for the oscillator circuit:
http://www.voti.nl/blink/pics/bb-870-big.jpg

SW · Joined: 09 Jun 2011 Posts: 10

Hey guys thanks a bunch. Ttelmah was right. It was the capacitors! I can't believe I overlooked those all three times I rewired the board. Everything is working now!

The wires you see on the boards are color-coded by their lengths. They come from the jumper kit JW-350 by Elenco.

Thanks again for all the help!
_________________
-SW