I wonder if there is an area to post a hardware related question. I understand if the answer is No due to the fact it is such a large scope.
Printable View
I wonder if there is an area to post a hardware related question. I understand if the answer is No due to the fact it is such a large scope.
Just shoot with your question... you never know it might be of interest to somebody somewhere... this section is pretty good in which to ask it...
After laying out and etching two boards and populating them with parts I now remember a problem I had a year ago trying to use two PIC16F84A's together on the same board. I finally got so frustrated with the business that I switched to another device with more pinouts (a 16F877, I think.)
Last night I finally connected a secondary board with a PIC16F628A to the ground and 5VDC buses of a primary with another 16F628A. The primary has an LM7805 on it to source both PIC's. When tested apart, both boards run and test out perfectly.
When I connect the secondary to the power buses of the primary, the PIC on the secondary board fails to operate. I have checked the Vss and Vdd pins on the secondary and I get 5VDC on Both!
Here are the design facts:
- Both 16F628A's are set to run on their respective internal clocks at 4MHz
- Both 16F628A's have !MCLR enabled (pulled up with a 4700ohm external resistor) to be used as a manual pushbutton reset.
- Comparitors ar disabled on both PIC's
- I always solder a filter cap right under the socket across Vss and Vdd, in this case a mono. ceramic 1uF, 50v and did so on both PIC's.
- Finally, I did not make the error of putting 5VDC on both pins of the power connector, and there are no circuit paths on the secondary board which would cause that. I am certain that all ground traces/pins end up at the 7805 center pin, and there are no shorts. The filter caps are the only bridges.
This is really bewildering. I can't believe an industrial application of plc's wouldn't have a situation where multiple devices would be used together in the same sub-system so device isolation must be an art.
Does anybody know why my PIC's don't ever play well together? Is there some harmonics problem between the two clocks that the filter caps. don't overcome? I've often wondered about this. I don't have the years of experience to recognize things like that (the art.)
Sounds like a problem with the ground trace to the second board.Quote:
Originally Posted by afbecker
You say it works fine on its own, so is the terminal you are taking ground from on the first board somehow rising to +5v when the first board is running?
Regards
Kieth,
Well,
I'll have to check that out this evening. It should be continuous to the 7805 center pin with no bridges but I could meter it nonetheless and see where that leads.
I have had a reply from the MicroChip forum from a fellow who paralleled 4 PIC's together on the same bus and had no problems.
My difficulty showed up at 1:00 a.m. so that may be part of it!
If it was me, I'd establish one single point as ground, probably the center pin of the LM7805. Then I'd check all the possible points on both boards with a multimeter to see if there's any DC bias. I agree that it sounds like something is wrong in the power distribution or grounding.
You were all quite right.
I searched through the whole thing and found the ground trace was incomplete. The condition was no doubt the result of last minute circuit changes made at 1:00 a.m.
I learned several things here with everybody's contribution. First don't panic about fault conditions occurring in the wee hours. Second, it is possible to see Vss floating at 5VDC with no ground and a connection at Vdd. Finally it appears that several PIC's can live happily together on the same bus.
I appreciate the help. Next time I hope the subject of discussion will be a PicBasic programming one.
Allan
Glad you found itQuote:
Originally Posted by afbecker
Slight correction to that statement.....Quote:
First don't panic about fault conditions occurring in the wee hours.
Dont panic about fault conditions occuring after you have been working for too long ! :)
I doubt there is a single person on this list who hasnt made an error due to tiredness or distraction.
I think my worst error was when I was repairing a turntable that had a faulty transformer. I was wiring in a temporary replacement when someone distracted me. When I continued I joined up the wrong wires and stuffed 240V mains straight into the PCB that was supposed to take 20V AC.
Not ALL the IC's exploded... some just blistered... but the turntable had to go to that great disco in the sky :(
My favorite was accidentally reversing the power leads on the pc board for a new project. I (and a whole room full of people) hit the floor thinking it was gunshots when the electrolytic capacitors exploded.Quote:
Originally Posted by keithdoxey
Its also a good idea to use .1uF caps at the pic instead of the 1uF, .1uF seems to be the industry standard for decoupling. I use them on every ic on my boards and i also use them between any external i/o and gnd. I find this filters out all ac noise I receive on cables going to sensors etc. I also use ferites on the input power and power lines between boards. They are probable not necessary but I don't like to take any chances when building Idustrial Controllers because you can get a lot of noise from the motors etc.