TreadBot

[skimask]

If VSS and VDD are next to each other, try to allow more, 0.010 may not be enough in the long run. Solder will grow “whiskers” over time and cause the gap to close. The more amps the more gap and wider the trace.
Clean all rosin from the board. Rosin will become conductive at elevated temperatures. Then use a sealer when finished testing. Spray lacquer works good to keep the air off.
Think of you board as a capacitor. Components and traces from top to bottom may have a capacitor effect. The large fields of copper that mister e spoke of will add to this problem. For fun you can build high voltage caps from pc board.

I just got back from a circuit card repair school, 7 weeks, got my Class 3 CCR Tech certificate. I got some serious training on the effects of 'tin whiskers' and old Rosin. Those 'tin whiskers' aren't nothing to play with. Got really good pictures of whiskers approaching 1/8" long after only a year on a PCB. And we took ohm readings across old, warmish rosin, and they were LOW...less than 1K in a lot of cases. Cool class though...

i know no one is bashin me...

Far from it... I'm relatively impressed at the quality of that home-brew board.

upon closerr inspection, im pretty sure it has to do with my MCLR stuff.. it looks like i might have confussed Vdd and Vss, so i think my resistor goes to ground when i think it should go to Vdd right???

Could you post a picture with and without the components? Might be a lot easier for us to help you troubleshoot this...

[dragons_fire]

heres a couple pics of my boards:

first one, an old board from about 2 years ago, made with a sharpie marker.. PIC16F84A robot controller



second one, a board i made about 6 months ago using a presensitized board and on overhead transparency.... GPS interface for robot



third one, the one for this robot, using the toner transfer method...

[dragons_fire]

here is a copy of the board layout im using for this thing, i have modified it a bit since i used it.. im not happy with this one that i made..
looking top down, blue is silkscreen, red is top and green is bottom copper layers.

top left 9-pin header is motor controller (wired properly, not installed yet), big thing in the top middle with 7-pin header is an SD card writer (not installed yet), top right labeled "tilt" is a memsic 2-axis accelerometer (not installed yet), 3-pin header in middle left is switch, right below the switch is the vreg, to the right is a 6-pin and 10-pin and 2 x 2-pin headers to go to top board (not made yet), OSC on right side of Socket, 10-pin header below socket goes to top board. LCD and my ultrasonic sensors arent installed yet..

[Darrel Taylor]

What regulator are you using?

You said 5V @ 5A

The 5amp TO-220 ones I'm familiar with have Vout on the Tab.
It looks like your tab is connected to ground.
<br>

[mackrackit]

My earlier statement about a 470 from MCLR to VDD is not quite correct for this chip. I have not used this one and after looking at the data sheet for it things are just a little different. VDD does go to MCLR but MCLR on this chip can also be used as an INPUT and has some different characteristics.

If I am looking at your board layout correctly, you may want to look at page 34 of the data sheet. I see MCLR going straight to VDD on your board. Do not know if this is the problem and I am getting a little fuzzy. Been up for 48hrs.

[skimask]

first one, an old board from about 2 years ago, made with a sharpie marker.. PIC16F84A robot controller

Lots of fingerprints...gotta clean those boards good before you etch. Pretty nice for a sharpie.

second one, a board i made about 6 months ago using a presensitized board and on overhead transparency.... GPS interface for robot

Lots of extra flux, but no fingerprints I can see. And again, pretty good. 90 degree corners aren't the best though once you start dealing with higher frequencies.

third one, the one for this robot, using the toner transfer method...

Get some rubbing alcohol and those soft acid brushes ( http://www.tecratools.com/pages/asso...ics/15215l.gif ) to break down and clean off the excess flux when you're done soldering. That stuff will harden up over time and eat away at everything, as well as possibly cause high resistance shorts (low resistance opens? ) over time. In that class I mentioned, they stressed to go ahead and use loads and loads of flux while you're soldering, but also to clean the heck out of it when you're done.

[dragons_fire]

youre right, the vreg im using has an output on the tab. so thats been fixed now, but i still dont think that was part of my probelm, since everything works when the pic isnt in it..

i tired taking out all the MCLR stuff so it doesnt connect to anything, and it still heats up pretty quickly.. im starting to wonder if the chip is destroyed and that causing it to heat up now, or if there is still something wrong..

in the schematic, MCLR is connected through the diode to the resistor and cap. its also connected to the prog pin. i see where you thought it connected to Vdd, but it actually doesnt.. the data sheet i have, its a notes page on page 34. page 35 has stuff about resets and shows the same wiring that i have.

when i etched this board, there is copper on the top too, and some of my pins dont match up perfectly, so there could be a short under the socket. i checked before i soldered, but i could have missed something..

im going to play around with the toner transfer, and see if i can get it any better. if not, i will probably make a new photoetched board.
so changes to the board now are:
vreg tab connected to output,
no more ground planes
MCLR stuff layed out better,
copper pads on top are gone, (except for the top traces)
.01 gap between pads and traces, .02 betweeen traces
more gap between Vdd and standoffs (Vss)

thanks for the comments on the old boards i made.. when i started, i thought the sharpie was the only way to do it, and that it needed sprecial equipment to make boards any other way. the photoetched one, was actually something called RCAP, its an autopilot for RC planes. its an opensource project, and the board layout was straight out of their website. it runs at 2400 baud, so i dont think it will have too many problems with 90 degree corners.

but i will stop typing now and go play with my toner and see what i can do..

[skimask]

i tired taking out all the MCLR stuff so it doesnt connect to anything, and it still heats up pretty quickly.. im starting to wonder if the chip is destroyed and that causing it to heat up now

All of the MCLR stuff? resistor and all?
Chip destroyed - entirely possible it's shorted internally...

when i etched this board, there is copper on the top too, and some of my pins dont match up perfectly, so there could be a short under the socket. i checked before i soldered, but i could have missed something..

I wouldn't throw that board out just yet. There's plenty you can do with an exacto knife, fix-it-jumper-wires, and a soldering gun.

so changes to the board now are:
vreg tab connected to output,
no more ground planes
MCLR stuff layed out better,
copper pads on top are gone, (except for the top traces)
.01 gap between pads and traces, .02 betweeen traces
more gap between Vdd and standoffs (Vss)

Nothing wrong with ground planes, as long as they're connected to ground! It's the unconnected ones that may cause trouble. Also, a big slab of copper under a part also makes an effective heat sink, whether it's connected to ground or not. Good calls on the rest of the notes...

it runs at 2400 baud, so i dont think it will have too many problems with 90 degree corners.

It's not the communications at 2400 baud that you'll have trouble with, it's those signals running at 10mhz+ hitting that 90 angle, causing a reflection, that reflection bouncing back on it's input and possibly pulling an input back low somewhere else, or something this or something that.
Don't forget, just because your PIC is only running 10mhz or whatever, each edge of each signal (square wave) is a small chunk of a wave of a much higher frequency (depending on how fast it rises and falls). The higher the frequency, the worse the problems. Suffice to say, 90 bad, 45 good enough, at least it's worked for me up to about 80mhz.

[dragons_fire]

i needed a little sense of accomplishment, so here it is!! it running off of an 877A with a pololu micro dual serial motor driver..

[dragons_fire]

heres the .PCB file (PCBexpress) of my board as it is right now.. if anyone is bored and wants to take a look and see if they can see anything wrong, let me know...

if you have any suggestions on how to make it better, those would be appreciated too..

i think im going to remake the board.. i like making them, its half the fun..


i had to ZIP it or else it wouldnt let me upload it...


thanks for all the help guys.. i like playing with this stuff until i can get it to work... but this time its getting a little frustrating because i cant really seem to find the problem..