Searching \ for '[PIC]: XTAL startup issues on PIC 73B' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/microchip/time.htm?key=xtal
Search entire site for: 'XTAL startup issues on PIC 73B'.

Exact match. Not showing close matches.
PICList Thread
'[PIC]: XTAL startup issues on PIC 73B'
2000\06\09@083259 by D. Schouten

picon face
Hi All,

In one of my applications I'm using a PIC16C73B 20MHz part
running at 18,432MHz. The xtal is connected the standard way
i.e. with two 33pF caps (loadcapacitance spec of xtal is 32pF)
to ground and layout wise everything as close as possible to
the PIC chip.

The problem is that from the 1000 boards I test, about 25 aren't
running at all. Replacing the xtal usually solves this problem.
Sometimes the board works well in the pretesting phase, but fails
at the final test. Current xtal brand is TXC, but we've had similar
problems with other brands too.

My question is, is this 2.5% failure rate normal, or is something
wrong with my application?

Is the PIC16C73B sensitive in HS mode, or do I have to add small
series
resistance with the xtal, or large parallel resistance?

The application which has his problem is quite price sensitive, so
there isn't any room for varicaps or other tweaking stuff.

Thanks for any help.

Daniel...

2000\06\09@084751 by Jim P

flavicon
face
Daniel -

Short of fully characterizing a crystal that works against a
crystal that doesn't work in a crystal text fixture using
appropriate test equipment -

Does a change to the parallel cap values have any effect
in getting a bad crystal to work?

Can you change just one of the parallel caps and get
the crystal to work?

By now you should have a pile of 'bad' parts - when a
difference that 'cures' the problem is found - try several
different bad xtals. I have used small machine pin sockets
with xtals successfully in the past to 'socket' xtals.

These are the kind of 'general' things I would try in order
to 'see' what works then go from there ...

Jim P


{Original Message removed}

2000\06\09@090248 by Andrew Kunz

flavicon
face
I've heard there are many reports on the '73B (and '77 and F87x) having problems
starting up crystals.

We have had to go to HS mode even with 4.000 MHz crystals.

Haven't had problems on my products with ceramic resonators, though.

Andy

2000\06\09@092745 by stouchton

flavicon
face
Primarily, all crystals are different, and all oscillator circuits are
different.  If these differences line up the wrong way, no oscillation.
Most manufacturers of oscillator circuits never fully characterize its
parameters at wafer level, and unknown variances go into the marketplace. I
went around and around with Microchip on this one several years ago, and
could not find one engineer that could state exactly what the crystal
specifications are.  They primarily spec'd load capacitance, but this is
only a small part of the crystal spec.  Equally important is the loss of the
crystal (Rs).

Here is a trick that worked for me with RF Microdevices:   Contact the
crystal manufacturers and let them know what you are up to.  After
contacting about 10 of them, I lucked onto the people that supplied the
original spec to RFMD for their design spec (Bomar Crystal).  you might get
lucky too.



ALSO:   Power supply is critical for xtal startup / run.  Make sure your DC
is reaaaaal good.  Make sure you have a GOOD decoupling cap, and limit lead
length.  (High quality ceramic, short heavy leads, avoid 90 degree bends).
Harmonics running around on the supply seem to really effect the PIC's
oscillator circuit.


My bet is that the Rs spec on your xtal is marginal.  If you have a network
analyzer (or spectrum analyzer, or scope and sig gen) you can compare good
to bad xtals and determine what spec is pushing the limit.  I would first
primarily focus on the loss through the xtal.

{Original Message removed}

2000\06\09@110726 by Don B. Roadman

flavicon
face
On 9 Jun 2000, at 14:29, D. Schouten wrote:

> n one of my applications I'm using a PIC16C73B 20MHz part
> running at 18,432MHz. The xtal is connected the standard way
> i.e. with two 33pF caps (loadcapacitance spec of xtal is 32pF)
> to ground and layout wise everything as close as possible to
> the PIC chip.
>
> The problem is that from the 1000 boards I test, about 25 aren't
> running at all. Replacing the xtal usually solves this problem.
> Sometimes the board works well in the pretesting phase, but fails at
> the final test. Current xtal brand is TXC, but we've had similar
> problems with other brands too.

I'm posting this reluctantly because it is not a solution. Some
others had some good Ideas, and I thought I'd submit this little
piece of info for what its worth. I was playing with some 16F84s. I
had a 10 Mhz part that I bought back in 98, I believe, and had been
using it with some 8.7Mhz crystals that came in a grab bag from
Dan's. I tried a bunch of them and they all worked fine with this 10
Mhz pic. About a month ago, I yanked this pic to use temporarily
untill some new ones came.  The vendor screwed up and sent 4
Mhz parts, and I tried them anyway using the 8.7M crystals. None
of them worked and I tried several crystals, so I figured that pics
arent capable of much overclocking. They did work with some 4.7
Mhz crystals.

Just recently, I got very bored and hooked up a 10 Mhz crystal to
one of these 4 Mhz pics, and it ran fine with a simple little led
flashing program. I was amazed, so I stuck a 20 mhz crystal on,
and it still worked! I tried several of both 10 and 20 Mhz crystals
and it always worked. Also tried 3 or 4 of the pics, and always, it
worked. I didnt have any higher freq. crystals to try, so I still dont
know where these chipc would crap out. I didn't have the presence
of mind to examine the oscillator ouput with a hi impedance probe,
so I dont know if the osc wasnt running or the chip crapped out.
Now I suspect the oscillator didnt start. I  surmise that there are
some crystals that some pics just don't like, but I don't know why,
and I guess I never will short of doing fully instrumented detective
work.

I posted a querry looking for test programs that are designed to
wring out a pic and put it through its paces to see if it is a good
part as far as possible, but I still havent found anything done just
as a thorough chip test program. I've only played with the '877 parts
a llittle, so I am interested in this thread and its eventual outcome.

2000\06\09@113435 by JP.BROWN

flavicon
face
Closely related to this problem, I had trouble with unreliable start up
some time ago using the 16C61 chip (running at 20MHz), I built a number of
these circuits and they where all a bit iffy on start up. If I started to
probe around the reset pin the circuits would spring to life, I cured the
problem by using the reset circuit in the PIC data book (the one with 2 Rs
a cap and a diode) after this I had no problems at all with start up.
(before I had been using a simple pull up R and PWRTE enabled)

On Fri, 9 Jun 2000, Don B. Roadman wrote:

{Quote hidden}

         -----  John P. Brown      spam_OUTJ.P.BrownTakeThisOuTspambradford.ac.uk ----
          \            --- Witty remark goes here ---         /
           --------------------------------------------------

2000\06\09@125807 by Harold M Hallikainen

picon face
       I also went round and round trying to get resonators that were
guaranteed to work on a 16c74 at 16 MHz. I finally found muRata
characterizes their resonators for specific chips and frequencies (see
http://www.murata.co.jp/search/ic-e.html). Microchip could not give me
characteristics on the oscillator (gain and phase versus frequency plot),
and the resonator companies generally could not do anything without that.
One resonator company told me the resonator was fine if it worked in
their test circuit (CMOS gate with feedback resistor). The fact that they
could not supply a resonator (with internal capacitors) that worked in MY
circuit (a 16c74 with the resonator hung between two pins, center pin of
the resonator grounded) lost them a sale for a few thousand resonators.
       Another issue I noticed was on one board I thought I'd "improve" it by
increasing the width of the very short traces from the 16c74 pins to the
resonator from 12 mils to 25 mils. Worked fine... 'til Microchip came out
with the 16c74b. Would not oscillate. We had to continue to use the
16c74a on that board. All my other "unimproved" boards worked fine with
the 16c74b...
       So, a corollary to Murphy's law (this one from the broadcast industry)
is that amplifiers will oscillate, and oscillators won't.

Harold



FCC Rules Online at http://hallikainen.com/FccRules
Lighting control for theatre and television at http://www.dovesystems.com

________________________________________________________________
YOU'RE PAYING TOO MUCH FOR THE INTERNET!
Juno now offers FREE Internet Access!
Try it today - there's no risk!  For your FREE software, visit:
dl.http://www.juno.com/get/tagj.

2000\06\09@132120 by D. Schouten

picon face
Thanks to all people who responded on my question sofar.

To be more detailed about the application environment, the
73B is powered by +5V +/- 2%, and all standard design rules like
bypassing the supply line, groundplaneing and short board traces
are all applied. This combined with using SMT for the circuit means
minimal parasitics.

Here's what I've got up to now.

I've downloaded the app note 'FACT001' from the Microchip
website which describes xtal oscillator design.
Due to the limited time space I had today I haven't tried
everything yet.

I've checked the clockout pin of the PIC to see if the
oscillation signal had sufficient amplitude and if there
was any clipping which would mean overdriving the xtal.
My little 18,432MHz sinewave wasn't clipping, so I assume
I don't need any series resistance to temper the gain circuit.
Also the amplitude was 4.3Vpp, all measured with 25 degrees Celsius
ambient temperature.

The loadcap spec of the xtals I use is 32pF (=CL). I have use two
33pF caps to ground on the oscillator pins, which is slightly
low in value according to the following formula I discovered today :

CL = (C1*C2)/(C1+C2) * Ccircuit

The Microchip appnote took 12.5pF as an average circuit capacitance,
so I did that too. With C1=C2=33pF, the loadcapacitance is 29pF.
This value is 3pF under the xtal spec.

I know that the ceramic caps I use are having a tolerance of +/-10%
and
Ccircuit is guessed, but I think I have to increase the 33pF caps a
bit.
What I'm going to do is enlarging the cap on the OSCOUT pin from
33pF to 47pF, so that I come closer to the 32pF CL spec.
Also the Microchip appnote advised to enlarge this specific oscout cap
in case of startup problems.

I'm going to try it a.s.a.p. and keep you up to date.

Daniel...


> {Original Message removed}

2000\06\09@134607 by Dan Michaels

flavicon
face
Daniel Schouten wrote:
.....
>What I'm going to do is enlarging the cap on the OSCOUT pin from
>33pF to 47pF, so that I come closer to the 32pF CL spec.
>Also the Microchip appnote advised to enlarge this specific oscout cap
>in case of startup problems.
.......

I haven't tried the 'B' chips myself, but in the past others
have also expressed similar troubles. For reference, Scenix chips
are notorious for having startup problems, and don't seem to
want to run unless a feedback resistor is wired between OSC1
and OSC2, and the 2 caps are "different" in value. You might
try experimenting along these same lines. Rs might be as low
as 33K, high as 1M. Caps 5-10 pF different. [Maybe piclist can
discover the magical formula, and bill Mchp for its time and
discomfort - <:-))].

best regards,
- Dan Michaels
Oricom Technologies
http://www.sni.net/~oricom
==========================

2000\06\09@135618 by Dan Michaels

flavicon
face
Harold wrote:
.........
>        Another issue I noticed was on one board I thought I'd "improve" it by
>increasing the width of the very short traces from the 16c74 pins to the
>resonator from 12 mils to 25 mils. Worked fine... 'til Microchip came out
>with the 16c74b. Would not oscillate. We had to continue to use the
>16c74a on that board. All my other "unimproved" boards worked fine with
>the 16c74b...
..........

Harold, did you ever try using a xtal with the 'B parts, or only
resonators ?????

Cheers,
- Dan Michaels
==============

2000\06\09@140709 by Martin SchŠfer

flavicon
face
Daniel,

I think that the parallel resistor (you called it feedback resistor) is the
best way for a secure start of an oscillator. This has nothing to do with
tricks concerning the 73B in XT mode. This is known as basics in EE. For the
PIC and an 20MHz crystal I suggest an value of about 680K. Before you have
not tried this, you cant exactly say, where the problem is.

Martin


;**********************************************
;**   name:      Martin Schaefer             **
;**   company:   elektronik 21 GmbH, Germany **
;**   e-mail:    .....schaeferKILLspamspam@spam@elektronik21.de    **
;**********************************************


> {Original Message removed}

2000\06\09@140913 by Dan Michaels

flavicon
face
Daniel Schouten wrote:
.....
>What I'm going to do is enlarging the cap on the OSCOUT pin from
>33pF to 47pF, so that I come closer to the 32pF CL spec.
>Also the Microchip appnote advised to enlarge this specific oscout cap
>in case of startup problems.
.......

Harold wrote:
>        I also went round and round trying to get resonators that were
>guaranteed to work on a 16c74 at 16 MHz. I finally found muRata
>characterizes their resonators for specific chips and frequencies (see
>http://www.murata.co.jp/search/ic-e.html). Microchip could not give me
........

Just for the heck of it, I checked out the link Harold gave, viz-a-viz
resonators for Scenix chips:

www.murata.co.jp/Ceramy/owa/ICLIST.showiclist?sKeyHin=SX28AC-HS&sKeyM
ak=SCENIX&sLang=2&cntmax=50

It gives following values for use with their resonators. Rd is series
resistor.

                C1  C2   Rf   Rd
CSA12.0MTZ       30  30   1M    0
CSA20.00MXZ040   10  10   1M    0
CSA4.00MG040    100 100 100K    0
CSA8.00MTZ       30  30   1M    0
CSTCC4.00MG0H    47  47 100K  470
CSTCC8.00MG0H    47  47   1M  470

Of note is the smaller C values used at higher freq. Daniel, maybe
your caps values are going in the wrong direction, and should be
smaller - despite what the Mchp datasheets say. After all, keep in
mind that following Mchp's advice hasn't produced the results you
want. I have always used 18-27 pF with 20Mhz xtals [although never
tried 'B series chips], and never had a problem.

- Dan

2000\06\09@141954 by Dan Michaels

flavicon
face
Martin Schaefer wrote:
>I think that the parallel resistor (you called it feedback resistor) is the
>best way for a secure start of an oscillator. This has nothing to do with
>tricks concerning the 73B in XT mode. This is known as basics in EE. For the
>PIC and an 20MHz crystal I suggest an value of about 680K. Before you have
>not tried this, you cant exactly say, where the problem is.
......
.......
>> try experimenting along these same lines. Rs might be as low
>> as 33K, high as 1M. Caps 5-10 pF different.
........


Oops, I meant to say "Rf might be as low ....."
Feedback resistor goes from OSC1 to OSC2 [or "parallel" as usage
goes].

Martin, the PIC datasheets I have show an Rf internal to the
PIC chip, so supposedly an external one is not required. However,
it is possible that adding a correctly-valued external one will help
with the newer 'B parts. Only trial and error will tell.

best regards,
- Dan Michaels
Oricom Technologies
http://www.sni.net/~oricom
==========================

2000\06\09@145309 by Dan Michaels

flavicon
face
[back again]. Found the following in doc# DS31002A, which is the Oscillator
section of the Mid-Range Reference Manual on the '97 CDROM:

"A method for improving start-up is to use a value of C2 greater than C1.
This causes a greater
phase shift across the crystal at power-up, which speeds oscillator start-up".

Note that C2 is shown on output pin OSC2.
============

Also note the following re "... if the circuit worked best ...", which may
provide
some guidance too:

.....
"Connecting the probe to the OSC1 pin will load the pin too much and
negatively affect performance. Remember that a scope probe adds its own
capacitance to the circuit, so this may have to be accounted for in your
design, i.e. if the circuit worked best with a C2 of 20 pF and scope probe
was 10 pF, a 30 pF capacitor may actually be called for. The output signal
should not be clipping or squashed. Overdriving the crystal can also lead to
the circuit jumping to a higher harmonic level or even crystal damage".

2000\06\09@150759 by David VanHorn

flavicon
face
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

At 12:52 PM 6/9/00 -0600, Dan Michaels wrote:
>[back again]. Found the following in doc# DS31002A, which is the Oscillator
>section of the Mid-Range Reference Manual on the '97 CDROM:
>
>"A method for improving start-up is to use a value of C2 greater than C1.
>This causes a greater
>phase shift across the crystal at power-up, which speeds oscillator
start-up".
>
>Note that C2 is shown on output pin OSC2.

You'd also have to tweak the other C value, as the loading C for the xtal
needs to stay at the specified value   Cl=(C1*C2)/(C1+C2)


- --
Are you an ISP?  Tired of spam?
http://www.spamwhack.com  A pre-emptive strike against spam!

Where's Dave? http://www.findu.com/cgi-bin/find.cgi?kc6ete-9

-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 6.5.2 for non-commercial use <http://www.pgp.com>

iQA/AwUBOUFcIYFlGDz1l6VWEQKdywCfa4oYeOjVeEKwr7IlEO2BjHjGk/gAn1bh
PCs71Vn2iYv7uZnMxBbKixU8
=jdk7
-----END PGP SIGNATURE-----

2000\06\09@151350 by mike

flavicon
face
{Quote hidden}

Have you tried a high-value (1m+) resistor across the xtal - this can
sometimes help startup

2000\06\09@210724 by steve

flavicon
face
> I've checked the clockout pin of the PIC to see if the
> oscillation signal had sufficient amplitude and if there
> was any clipping which would mean overdriving the xtal.

Looking for clipping isn't a good method to check for overdriving.
An overdriven crystal will exhibit phase noise long before you'll see
anything on a scope. The best way to test this is to have a simple
program in the chip to toggle a pin and watch the output on a scope.
Stretch the display out so you can watch the edge you aren't
triggering on. The more it is overdriven, the more the frequency will
jump around.

Steve
======================================================
Steve Baldwin                Electronic Product Design
TLA Microsystems Ltd         Microcontroller Specialists
PO Box 15-680, New Lynn      http://www.tla.co.nz
Auckland, New Zealand        ph  +64 9 820-2221
email: stevebspamKILLspamtla.co.nz      fax +64 9 820-1929
======================================================

2000\06\10@065003 by D. Schouten

picon face
> It gives following values for use with their resonators. Rd is
series
> resistor.
>
>                  C1  C2   Rf   Rd
> CSA12.0MTZ       30  30   1M    0
> CSA20.00MXZ040   10  10   1M    0
> CSA4.00MG040    100 100 100K    0
> CSA8.00MTZ       30  30   1M    0
> CSTCC4.00MG0H    47  47 100K  470
> CSTCC8.00MG0H    47  47   1M  470
>
> Of note is the smaller C values used at higher freq. Daniel, maybe
> your caps values are going in the wrong direction, and should be
> smaller - despite what the Mchp datasheets say. After all, keep in
> mind that following Mchp's advice hasn't produced the results you
> want. I have always used 18-27 pF with 20Mhz xtals [although never
> tried 'B series chips], and never had a problem.

Well, it could be true that the values are going the wrong way. But
with lowering the C values I'm only moving even further away from
the CL spec of the xtal.

Up to now I haven't really tried something on the hardware, which is
the first thing I'm going to do Tuesday morning. Sofar I have only
studied the problem by looking for info on the web and ofcourse this
great list.
I have two boards which won't start, so I'm going to experiment with
these.

First I'm going to enlarge the C2 value as I quoted earlier. Besides
this
I will experiment with parallel resistance too like some of you guys
advised.

On the subject of 'A' parts and 'B' parts, I have to admit that the
number
of 'no startup' failures were getting much larger since the 'B' parts
were implemented.

Daniel...

2000\06\10@112809 by Peter L. Peres

picon face
Hi,

from what you are saying I think that 2.5% of the crystals have a bad Q. You
can sort/test your crystals before production using a GDO lacking a better
instrument. You will be comparing the incoming crystals with one slelected
for high 'enough' Q previously, from a group.

hope this helps,

Peter

PS: Do the 2.5% failed crystals work in other circuits ?

2000\06\10@150958 by Dan Michaels

flavicon
face
Harold wrote:
..........
>        Another issue I noticed was on one board I thought I'd "improve" it by
>increasing the width of the very short traces from the 16c74 pins to the
>resonator from 12 mils to 25 mils. Worked fine... 'til Microchip came out
>with the 16c74b. Would not oscillate. We had to continue to use the
>16c74a on that board. All my other "unimproved" boards worked fine with
>the 16c74b...
.........

D. Schouten wrote:
..........
>On the subject of 'A' parts and 'B' parts, I have to admit that the
>number>of 'no startup' failures were getting much larger since the 'B' parts
>were implemented.


Has it occurred to anyone other than myself that the 'B part appears
to be awfully finicky - so that simply changing trace width creates
a go-no-go situation?

I understand problems with Scenix, as they are running 50 Mhz, which
happens to be in the "critical" frequency range where RF effects start
to take over, ie transmission/reflections/harmonis/etc, and where
fundamental mode crystals are hard to come by, but 16 Mhz with a PIC?

Any considerations as to the real problem here ???

1. Crappy Mchp engineering.

2. Crystals and resonators today are just too different from each
other, due to a profusion of new xtal manufacturers, so every design
situation is unique.

3. Too simplistic design approach by users. We have come to a point in
time where we really need Smith Charts, Q and loop gain measurements,
transmission line calcs, etc/etc, just to duplicate what is easy to
do on other controllers - even at relatively low frequencies.

4. Physics has changed in the new millenium.

5. Something else [phase of the moon, karma, prayer, etc].

Cheers,
- Dan Michaels
==============

2000\06\10@155834 by Reginald Neale

flavicon
face
<x-flowed>    Daniel asked:

>The problem is that from the 1000 boards I test, about 25 aren't
>running at all. Replacing the xtal usually solves this problem.
>Sometimes the board works well in the pretesting phase, but fails
>at the final test. Current xtal brand is TXC, but we've had similar
>problems with other brands too.

    What crystal package are you using? With some packages
   it is possible for PCB manufacturing process residue to
   get trapped under the crystal. When you unsolder the old
   one, it boils off any trapped moisture and the replacement
   crystal works fine. Of course, the more margin your
   oscillator circuit has, the less likely this is to make
   a difference. Clues: modest application of dry heat from
   heat gun causes failed oscillator to start. Application of
   condensing moisture by directing breath through a straw at
   crystal footprint causes oscillator to stop.

   Reg Neale

</x-flowed>

2000\06\10@160048 by Jim P

flavicon
face
The simplistic approach of utilizing a single 'inverter'
stage to be used as a crystal oscillator has always
been bad design practice in the past and is always
fraught with it's own set of perils.

A little practical History.

I re-designed a 10.24 MHz oscillator years ago that
used a TTL gate. The original oscillator and crystal
were overly subject to frequency changes with
temperature. Temp compensation was attempted
on the crystal and only partial success was had.

A discrete 2N2222 design was drawn up, built and
the same crystal now showed 10.245 MHz in the
new circuit! This new circuit with the original crystal
was less responsive to temp change than the TTL
gate design leading to the conclusion that *inherent
circuit propagation delays* in the TTL gate over
temperature were responsible for the excessive
frequency change with temp. This lead to speculation
that the inherent properties of the TTL gate contributed
substantially to frequency instability versus temp and
the prop delay also contributed substantially to the
off-frequency operation (the proper freq of operation
of the cystal was actually 10.245 MHz but ran at
10.24 MHz using  the TTL gate). I also seem to
remember a problem with starting or 'squegging'
(improper freq of operation) of the TTL based
oscillator.

Unfortunately, the new design with the same circuit
was oscillating 5 KHz higher in frequency. Padding
a 10.24 MHz crystal down 5 KHz is not done (this
is/would be 'excessive warping').

A new crystal was found and all were pleased in
Mudville with the 2N2222 based 10.240 MHz osc ...

I don't know what MChip (or others) actually has
fabricated on the die as an 'oscillator' stage but it
probably isn't a well designed analog 'oscillator'
stage with suitable feedback or biasing.

The moral of the story: It's "potluck" with the
onboard oscillator and an external crystal all of
unknown specification.

Jim P


Dan wrote:

<snip>

I understand problems with Scenix, as they are running 50 Mhz, which
happens to be in the "critical" frequency range where RF effects start
to take over, ie transmission/reflections/harmonis/etc, and where
fundamental mode crystals are hard to come by, but 16 Mhz with a PIC?

Any considerations as to the real problem here ???

1. Crappy Mchp engineering.

2. Crystals and resonators today are just too different from each
other, due to a profusion of new xtal manufacturers, so every design
situation is unique.

3. Too simplistic design approach by users. We have come to a point in
time where we really need Smith Charts, Q and loop gain measurements,
transmission line calcs, etc/etc, just to duplicate what is easy to
do on other controllers - even at relatively low frequencies.

4. Physics has changed in the new millenium.

5. Something else [phase of the moon, karma, prayer, etc].

Cheers,
- Dan Michaels
==============

2000\06\10@175710 by Dan Michaels

flavicon
face
Jim P wrote:
>The simplistic approach of utilizing a single 'inverter'
>stage to be used as a crystal oscillator has always
>been bad design practice in the past and is always
>fraught with it's own set of perils.
.........
{Quote hidden}

OK, I hear what you are saying about the single inverter stage
oscillator. OTOH, there are how many BILLIONS of these things
out there? Can they really be so fragile that simply changing trace
width like Harold did is gonna have such a great effect?

Cheers,
- Dan Michaels
==============

2000\06\10@184612 by Jim P

flavicon
face
I can't believe that at 10 MHz (or so) changing
trace width will have that much effect - BUT,
how near or on the edge might a critical circuit
parameter be for a particular design and layout
and where an apparent change of trace width
has en effect - that is the $64K US question.
Designs as critical as that are to be SERIOUSLY
avoided.

I would look elsewhere for that magical, critical
factor makes or breaks 'the deal' ...

How desirable in some of these designs would a
simple external oscillator comprised of three
resistors, a couple of caps (value 39 to 120 pF)
and a single cheap bipolar NPN transistor plus
crystal be *in return for* oscillator reliability?

Some serious bench work needs to be done
with a couple of these PICs and detemine
whether these built-in 'osc' elements require
more (or less) bias to put them into a linear
operating region, whether the xtals are suitable
or not, etc ... Some simple adjustment on cap
parms (the two external shunt caps) I would
think would solve most of the 'incompatibility'
of all but the most stubborn of crystals.

My .02 US anyway ...

Jim P


{Original Message removed}

2000\06\10@190307 by David VanHorn

flavicon
face
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Do your xtal caps return to the chip ground directly, on their own trace?

- --
Are you an ISP?  Tired of spam?
http://www.spamwhack.com  A pre-emptive strike against spam!

Where's Dave? http://www.findu.com/cgi-bin/find.cgi?kc6ete-9

-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 6.5.2 for non-commercial use <http://www.pgp.com>

iQA/AwUBOULkt4FlGDz1l6VWEQKrwgCgpoT1uWf7fBsE7UaVrw2krmRs5ygAoMJF
yhCQMOXGD2+Eh47xrBe3oo2g
=VZQC
-----END PGP SIGNATURE-----

2000\06\13@134418 by D. Schouten

picon face
Hi,


I'm back from experimenting with the 2.5% failed PIC73B boards
that didn't run with the 18.432MHz xtals.
I tried several things to get the PIC clock running without changing
the xtal, but without any results sofar.

I tried the following things :
1) Enlarge only the 33pF cap on the OSC2 pin to 47pF, as advised on
  the Microchip app note.
  -> Nothing happened.
2) Kept the two xtal caps at 33pF, and tried several parallel (Rf)
  resistor values ranging from 220k to 1M.
  -> Nothing happened
3) Same as point two, but with Cosc2 = 47pF instead of 33pF
  -> Nothing happened
4) Lowered the two xtal caps to 15pF
  -> Nothing happened
5) Same as point four but with parallel resistance
  -> Nothing happened
6) Made Cosc1 = 47pF and Cosc2 = 33pF (just in case the Microchip
  app note had a typing error)
  -> Nothing happened

Note that I did the above things with three different no-go boards.
Also the circuit design around the PIC is done according to all
standard design rules.

In all three cases, modding the board to it's original values and
changing the xtal by a new one, brought it finally to life.

So, I still haven't got a clue what is wrong with these malfunctioning
boards. Also, how can I test the xtals? I was thinking of making an
identical PCB with machined sockets to test every xtal before sending
them to my pcb assemble house. The ones that fail can be send back to
the xtal manufacturer (which ofcourse are willing to help you.....
NOT).

The failed xtals I have right now are desoldered from a multilayer
pcb.
So I don't trust these anymore since they're probably already damaged
due to high temps for too long duration.

The only thing I didn't try yet, is adding some series resistance with
the xtal.


Daniel...

2000\06\13@140052 by Arthur Brown

flavicon
face
Daniel
Looks like xtals or the way they are treated in the assembly of pcb's are to
blame for the problems you have got.
Why not build a test rig using 16f84 plus LED and croc-clips to test xtals
before using them.

Regards Art
----- Original Message -----
From: D. Schouten <.....danielsKILLspamspam.....XS4ALL.NL>
To: <EraseMEPICLISTspam_OUTspamTakeThisOuTMITVMA.MIT.EDU>
Sent: Tuesday, June 13, 2000 10:40 AM
Subject: Re: [PIC]: XTAL startup issues on PIC 73B


> Hi,
> SNIP<<<<<<<<
>
> In all three cases, modding the board to it's original values and
> changing the xtal by a new one, brought it finally to life.

2000\06\14@083950 by Robert A. LaBudde

flavicon
face
<x-flowed>At 07:40 PM 6/13/00 +0200, you wrote:
>Hi,
>
>
>I'm back from experimenting with the 2.5% failed PIC73B boards
>that didn't run with the 18.432MHz xtals.
>I tried several things to get the PIC clock running without changing
>the xtal, but without any results sofar.

After all this discussion, this may seem too obvious, but are the crystals
you're using cut for the purpose intended?

If not, adding a large series resistance (e.g., 100k-1M) with the PIC may
solve the problem. Carbon composition versions are helpful because of the
triggering noise they generate.

You should also eliminate the possibility of too slow clock start-up: the
PIC may not be bringing the oscillator up fast enough if it's poorly
matched to the crystal. In this case larger capacitors may exacerbate the
problem.

Also, given the time you've already spend on this, I'd suggest you build a
simple 1-transistor oscillator with a couple of variable capacitors and two
socket pins to plug a crystal in. This is essentially done on a plug-in
breadboard. Then you could experiment with a continuum of capacitances and
series resistors and discover dead xtals without the complexity of the PIC
in the loop.


================================================================
Robert A. LaBudde, PhD, PAS, Dpl. ACAFS  e-mail: ralspamspam_OUTlcfltd.com
Least Cost Formulations, Ltd.                   URL: http://lcfltd.com/
824 Timberlake Drive                            Tel: 757-467-0954
Virginia Beach, VA 23464-3239                   Fax: 757-467-2947

"Vere scire est per causas scire"
================================================================

</x-flowed>

2000\06\16@163622 by Erik Reikes

flavicon
face
You also might try pulling the pics themselves, sucjing their program out
and taking a look at the config word bits to make sure you are in HS mode.
I've had a flaky programmer problem where some of my PICS were getting bit
errors here and there on programming, especially in the config word area.


At 08:37 AM 6/14/00 -0400, Robert A. LaBudde wrote:
{Quote hidden}

Erik Reikes
Senior Software Engineer
Xsilogy, Inc.

KILLspamereikesKILLspamspamxsilogy.com
ph : (858) 535-5113
fax : (858) 535-5163
cell : (858) 663-1206

More... (looser matching)
- Last day of these posts
- In 2000 , 2001 only
- Today
- New search...