Searching \ for '[PIC] Final comments on oscillation' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: techref.massmind.org/techref/microchip/devices.htm?key=pic
Search entire site for: 'Final comments on oscillation'.

Exact match. Not showing close matches.
PICList Thread
'[PIC] Final comments on oscillation'
2006\04\09@191745 by Bill Kuncicky

picon face
Thanks for all the replies.  Everything seems to be working OK now.  Since I
did everything that people suggested (100 nanofarad caps across Vdd and Vss
close to the PIC pins, carefully checked for any intermittant shorts, and
put huskier caps on the 7805 and checked the output with the scope) it is
hard to say just exactly what did the trick.  Assuming the trick is did, of
course.  From what everyone has said there is no way that anything that I
did can hurt, though, in any case.

I have to say that I am amazed to learn that the caps on the regulator
(input to ground and output to ground) can be that large!  I have always
assumed that the specs of the manufacturer had to be followed pretty
closely!  Live and learn.  That is, IMHO, a Good Thing To Know!  I shall
file it with my many other Good Things To Know.

I also noticed that Bob Axtell said that the symptoms were "absolutely
classic" for the problem being oscillation, and Spero Pefhany has the
opinion that it sounds like something else (latchup, or intermittent short
of an output to supply rail.  So two of the experts disagree.  :-)  Of
course, I don't know what it was.  What I do know is that it seems to have
been cured, and I am grateful for the comments.  I tend to stay awake
thinking about these things when they do not get solved before bedtime.  :-(

Bill

2006\04\09@193516 by Jan-Erik Soderholm

face picon face
Bill Kuncicky wrote :

> it is hard to say just exactly what did the trick.

You could always pull each "fix" out again... :-)
One at a time this time.

> I also noticed that Bob Axtell said that the symptoms
> were "absolutely classic" for the problem being oscillation,
> and Spero Pefhany has the opinion that it sounds like
> something else (latchup, or intermittent short of an
> output to supply rail.  So two of the experts disagree.  :-)

Well, it takes to experts to realy disagree. doesn't it. :-)

Jan-Erik.



2006\04\09@194605 by olin piclist

face picon face
Bill Kuncicky wrote:
> I have to say that I am amazed to learn that the caps on the regulator
> (input to ground and output to ground) can be that large!  I have
> always assumed that the specs of the manufacturer had to be followed
> pretty closely!

Many regulators don't like really large caps on their output.  I think
people here were assuming you were using a 7805, which tend to be forgiving
but some versions don't want more than a few uF if I remember correctly.
Adding a larger cap to the output of a regulator can be a bad thing.  Always
stick to what the data sheet says.  If someone told you to do otherwise,
they were just plain wrong.

Think about it a bit too.  Large bulk caps are going to respond slowly,
which the circuit in the regulator can do too.  This is why you need large
caps on the regulator input.  That's where the regulator will get the extra
current from for the frequencies its circuitry can respond to.  However, the
circuit has high frequency limitations.  So the caps you add need to make up
for that.  Since the frequencies these caps need to repond to are high, the
caps themselves don't need to be that large.  If you think about it, a
regulator is a control problem, with all the associated instabilities as you
try to approach the edge of the circuit's speed.  A brick outhouse like the
7805 will be more stable over a range of conditions, but some LDOs are
unstable except with a narrow output capacitance range with a specific ESR.
Always do what the data sheet says.

All this has nothing to do with the bypass cap at the PIC.  The PIC can vary
it's current draw at fairly high frequencies, which can cause significant
voltage excursions at the other end of a wire from a regulator, due to the
wire's inductance and resistance.  That's why bypass caps need to be as
close as possible to the pins they are bypassing.


******************************************************************
Embed Inc, Littleton Massachusetts, (978) 742-9014.  #1 PIC
consultant in 2004 program year.  http://www.embedinc.com/products

2006\04\09@225740 by Bob Axtell

face picon face
Bill Kuncicky wrote:

>Thanks for all the replies.  Everything seems to be working OK now.  Since I
>did everything that people suggested (100 nanofarad caps across Vdd and Vss
>close to the PIC pins, carefully checked for any intermittant shorts, and
>put huskier caps on the 7805 and checked the output with the scope) it is
>hard to say just exactly what did the trick.  Assuming the trick is did, of
>course.  From what everyone has said there is no way that anything that I
>did can hurt, though, in any case.
>
>I have to say that I am amazed to learn that the caps on the regulator
>(input to ground and output to ground) can be that large!  I have always
>assumed that the specs of the manufacturer had to be followed pretty
>closely!  Live and learn.  That is, IMHO, a Good Thing To Know!  I shall
>file it with my many other Good Things To Know.
>
>I also noticed that Bob Axtell said that the symptoms were "absolutely
>classic" for the problem being oscillation, and Spero Pefhany has the
>opinion that it sounds like something else (latchup, or intermittent short
>of an output to supply rail.  So two of the experts disagree.  :-)  Of
>  
>
But they are probably related. The 100nF cap provides current when the
PIC needs
it (when the system clock changes state). Without it, almost anything
can happen.

>course, I don't know what it was.  What I do know is that it seems to have
>been cured, and I am grateful for the comments.  I tend to stay awake
>thinking about these things when they do not get solved before bedtime.  :-(
>
>Bill
>  
>


--
Note: To protect our network,
attachments must be sent to
spam_OUTattachTakeThisOuTspamengineer.cotse.net .
1-520-850-1673 USA/Canada
http://beam.to/azengineer

More... (looser matching)
- Last day of these posts
- In 2006 , 2007 only
- Today
- New search...