Searching \ for '[PIC]:crystal reliability' 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/devices.htm?key=pic
Search entire site for: 'crystal reliability'.

Exact match. Not showing close matches.
PICList Thread
'[PIC]:crystal reliability'
2005\07\14@065329 by Mark Chauvin

picon face
I'm experiencing about a 3% failure rate on my PIC-based product.  Except
for one time that a wire was cut, the rest of the time it's been the 20MHz
crystal that's died.  I'm using a 16C58B chip and the crystal is an Abracon
ABL-20.000MHZ-B2.  Is this a normal failure rate for this type of crystal?
Also, I'm not sure what the two capacitors should be.  Right now, I'm using
33pF capacitors with the crystal, but from what I read in the PIC manual
that may not be right.  The PIC manual recommends 15pF, but there's a note
that it should be 30pF if VDD > 4.5 (I'm using a 78L05 so VDD is 5 volts),
and Abracon lists "Shunt capacitance" for this crystal at 7pF & "Load
Capacitance" at 18pF.  Being almost totally ignorant of circuit design, I'm
not sure what the right capacitors are, so I'm blindly using what someone
else told me to use, but could using a capacitor that's too big cause these
failures?

Thanks
-Mark

2005\07\14@072219 by Jan-Erik Soderholm

face picon face
Mark Chauvin wrote :

> I'm experiencing about a 3% failure rate on my
PIC-based
> product.

> Is this a normal failure rate for this
> type
of crystal?

*Any* crystal would break if hitted or
shaked hard
enough... :-)

What type of environment is it used in ?

I'm using 22pF
as my "standard" setup.

I don't claimt to "know" this, but I guess
that a
crystal can be "over-driven" if used with to large
caps...

Jan-
Erik.



2005\07\14@072715 by Vasile Surducan

face picon face
On 7/14/05, Mark Chauvin <spam_OUTchauvin555TakeThisOuTspamsbcglobal.net> wrote:
> I'm experiencing about a 3% failure rate on my PIC-based product.


 Are you so kind and tell us on how many pieces has been computed
this 3%. I'm very curious because I found a quite similar rate in
other circumstances.


Except
{Quote hidden}

It could be possible that failure comes from the PIC-XTAL resonant
circuitry and not from the crystal itself. Do a check by inspecting
the shape of your XTAL-OUT pin using a scope and at least 1:10
(verified) probe. Check the signal level for the faulty crystals.
Check it again for a good crystal. A too small capacitor on the OSCOUT
will reveal a high Vpeak to peak signal, looking like a square signal
with some resonant shapes on every high logic level, immediately near
every rising edge. A too high capacitor will create a smaller
amplitude, looking more like a sinusoidal shape. Essential is the
level of that signal.
There is some signal with faulty crystals or there is no oscillation ?
Answering this will show you who is quilty.

best regards,
Vasile

2005\07\14@072902 by Wouter van Ooijen

face picon face
> I'm experiencing about a 3% failure rate on my PIC-based
> product. (snip) Is this a normal failure rate for this
> type of crystal?

I am not into big numbers, but last week I got a 'dead' Wisp628 returned
that was caused by a non-working crystal. This has never happened before
on ~ 1k Wisp628's. Currently I use a resonator instead of a crystal for
the prefab Wisp628's (the kit still has a crystal to avoid customer
confusion).

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\07\14@080327 by Mark Chauvin
picon face
Jan-Erik:
> What type of environment is it used in ?
Automotive scan tool

Valise:
>  Are you so kind and tell us on how many pieces has been computed
this 3%. I'm very curious because I found a quite similar rate in
other circumstances.

A few hundred units over the last couple of years

Wouter:
>Do a check by inspecting
the shape of your XTAL-OUT pin using a scope

Well, even if I did understand what all you said, I couldn't do it because I
don't have a scope.  I'm sure it's good information though.

>Currently I use a resonator instead of a crystal for
the prefab Wisp628's

So is a resonator more reliable than a crystal?

Can I assume in the crystal mfg's literature that "Load capacitance"
specifies the recommended values for c1 & c2?  That would mean I'm using
capacitors that are 83% bigger than recommended.

Thanks
-Mark

2005\07\14@081321 by Fr10

flavicon
face
Have you tried a series resistor

"Rs may be required in HS mode as

well as XT mode to avoid overdriving crystals with low drive

level specification."

Regards

Francois



> I'm experiencing about a 3% failure rate on my PIC-based product.  Except
> for one time that a wire was cut, the rest of the time it's been the 20MHz
> crystal that's died.  I'm using a 16C58B chip and the crystal is an
Abracon
> ABL-20.000MHZ-B2.  Is this a normal failure rate for this type of crystal?
> Also, I'm not sure what the two capacitors should be.  Right now, I'm
using
> 33pF capacitors with the crystal, but from what I read in the PIC manual
> that may not be right.  The PIC manual recommends 15pF, but there's a note
> that it should be 30pF if VDD > 4.5 (I'm using a 78L05 so VDD is 5 volts),
> and Abracon lists "Shunt capacitance" for this crystal at 7pF & "Load
> Capacitance" at 18pF.  Being almost totally ignorant of circuit design,
I'm
> not sure what the right capacitors are, so I'm blindly using what someone
> else told me to use, but could using a capacitor that's too big cause
these
> failures?
>
> Thanks
> -Mark

2005\07\14@082449 by Wouter van Ooijen

face picon face
> So is a resonator more reliable than a crystal?

It is 'know' to be more shock-resistant, and it includes the capacitors,
so that is one (or actually two?) things less to worry about. And much
cheaper too. The only catch is that a resonator is much less accurate: I
think 1% worst case, 0.5 typical. A crystal is more like 50 or 100 ppm.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\07\14@093137 by Mark Chauvin

picon face
No, I haven't, but that may be something to look into
I suppose.  It would require a redesign of the circuit
board though.  Couldn't hurt anything could it?

--- ".....Fr10KILLspamspam@spam@mweb.co.za" <fr10spamKILLspammweb.co.za> wrote:

{Quote hidden}

> --

2005\07\14@094519 by olin piclist

face picon face
Mark Chauvin wrote:
> I'm experiencing about a 3% failure rate on my PIC-based product. Except
> for one time that a wire was cut, the rest of the time it's been
> the 20MHz crystal that's died.  I'm using a 16C58B chip and the crystal
> is an Abracon ABL-20.000MHZ-B2.  Is this a normal failure rate for this
> type of crystal?

No, not if it is used correctly and not abused.

> Also, I'm not sure what the two capacitors should be. Right now, I'm using
> 33pF capacitors with the crystal, but from what I
> read in the PIC manual that may not be right.

It's not up to the PIC manual.  The load capacitance is a function of the
crystal design and is specified by the crystal manufacturer.  There seems to
be a lot of confusion and dead fish waving regarding load crystal load
capacitance on this list.  To some extent, the Microchip manuals don't help.

It's important to understand why both the capacitors on either side of the
crystal are there.  Each is there for a different reason.  Microchip might
have a spec on the maximum capacitance on the oscillator out pin at a given
frequency.  They have very little to say about the capacitance required on
the oscillator in pin.

> Being almost totally ignorant of circuit design, I'm not sure what the
> right capacitors are, so I'm blindly using what someone else told me to
> use,

Since you're using the dead fish method, you apparently need a smaller fish.
After all the fancy analysis, 22pF to ground on either side of the crystal
actually works quite well most of the time.  But remember most of the time
isn't all of the time.  For short leads with nothing else connected and
20MHz crystal rated at 18pF load, 22pF caps should work well.

Despite all that, you still need to understand your high failure rate.  Is
the unit being subjected to shock and vibration?  If so, maybe a crystal
isn't appropriate.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\07\14@095113 by olin piclist

face picon face
Mark Chauvin wrote:
>> What type of environment is it used in ?
>
> Automotive scan tool

That sounds like it's regularly subjected to shock and vibration.  Crystals
hate that.  The simplest answer is to use a ceramic resonator if you can
live with the lower frequency accuracy.  Resonators are more rugged
mechanically.

> Can I assume in the crystal mfg's literature that "Load capacitance"
> specifies the recommended values for c1 & c2?

No.  But your 33pF caps are definitely too high.  That by itself would
probably not damage the crystal, but it might be oscillation close to the
edge.  Small changes could make knock the loop gain just below 1 so that it
doesn't oscillate anymore.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\07\14@112613 by Spehro Pefhany

picon face
At 07:07 AM 7/14/2005 -0500, you wrote:

>Can I assume in the crystal mfg's literature that "Load capacitance"
>specifies the recommended values for c1 & c2?  That would mean I'm using
>capacitors that are 83% bigger than recommended.

No, you're using about the correct value, from that pov. The capacitors
are effectively in series in a Pierce oscillator and there is some
stray/input capacitance, so 33pF is about right for the 18pF load
capacitance. That means that the oscillator will be nominally on the marked
frequency, not a bit off. This is not your main problem.

HOWEVER, drive power and starting are different requirements from getting
the frequency spot-on... most crystal MANUFACTURERS (the guys who you
will complain to if you start getting 1% failures in 1E6 crystals)
recommend that you use a (expensive) 'scope current probe and calculate the
actual drive power of the crystal with different series resistors (plot it
on a graph). And you look at what maximum value of resistor will
cause the oscillator to fail to start, over temperature and allowing
for aging, unit to unit variation, power supply variations and so on.
Then you pick a value (if possible) that allows reliable and fast
starting and keeps the drive power within the recommended range.
Too much drive power can lead to accelerated aging, deterioration of
characteristics, and outright failure. Problems are more common with
tuning-fork style crystals, high frequency crystals, and the tiny
SMT crystals that can only handle 50-100uW rather than 1mW or more that
the HC49 crystals can handle. The crystal manufacturers don't typically
give you the information except as typical values of the series
resistance. The IC manufacturer doesn't give the range of the gm of the
oscillator section. So, there is not really a guarantee from anyone
that it will work.

If you can use a larger crystal (full height HC49) or a resonator (we've
used many thousands of 20MHz 3-pin resonators with zero problems),  or
drop the supply voltage to 3.3V, you'll probably get much higher reliability
without major changes. The ones you have out there now may be ticking time
bombs waiting to fail. 3% is a VERY high reported failure rate.

>Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
.....speffKILLspamspam.....interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2005\07\14@120839 by olin piclist

face picon face
Spehro Pefhany wrote:
> The capacitors
> are effectively in series

No, this is a common Crystal Myth.

If the crystal is driven with a 0 impedence sine wave on one side, the
specified load capacitance is that value to ground on the other side that
will cause the desired phase shift at the rated frequency.  Therefore, if
the PIC oscillator output pin has a small impedence compared to the load
capacitance, then the capacitor on the oscillator output pin doesn't enter
into consideration.  In that case, the best oscillator input pin capacitor
value would be the rated load capacitance minus the inherent existing
capacitance due to the leads, circuit board, etc.

If the PIC oscillator output had a very high impedence, then both capacitors
need to be considered in series as you said.  In practise, the PIC output is
somewhere in between.  A 18pF capacitance at 20MHz has about 440ohms
impedence.  The PIC output impedence is probably less than that, certainly
not so large compared to 440ohms that it can be ignored.  The effective
drive impedence seen by the crystal is the oscillator output in capacitance
plus stray capacitance all in parallel with the pin output impedence.  All
these parameters can't be know that accurately, but in any case the drive
impedence will likely be no more than half of just the added capacitor in
the 18pF to 22pF range at 20MHz.  At lower crystal frequencies, the PIC
output driver impedence is even less compared to the capacitor.

Also keep in mind the purpose of the oscillator output pin capacitance.
It's not there to provide the crystal load capacitance, although it effects
it in some cases.  It's purpose is to attenuate the harmonics in the crystal
drive waveform.  Excessive harmonics can damage the crystal and can even
cause the oscillator to run at a harmonic instead.  It's important to keep
the loop gain below 1 for harmonics.

There are a lot of subtle interacting effects here with unknown exact
values.  This is why I said earlier that after all the fancy analysis you
still don't really know the right answer, and 22pF on both sides of the
crystal is about right with reasonable assumptions and typical crystals
rated for about 18pF.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\07\14@122202 by Bob Blick

face picon face
On 14 Jul 2005 at 5:57, Mark Chauvin wrote:

> Also, I'm not sure what the two capacitors should be.  Right now, I'm using
> 33pF capacitors with the crystal,

> and Abracon lists "Shunt capacitance" for this crystal at 7pF & "Load
> Capacitance" at 18pF.

That is probably just about perfect choice of capacitor. The capacitors are effectively in
series, so the load capacitance is 33/2, or 16.5pF. Add some stray capacitance, and you
are probably between 18 and 20.

If you are having failures it is most likely mechanical shock or overdriven crystals, but I
have very little experience there. But it seems to me that overdriving would be a long-
term problem so if the device is used 24 hours a day it would be more of a problem than
in something used occasionally.

Adding a resistor in series with the crystal will reduce the drive, but unless you can
actually determine that overdriving is the problem, it may make startup less reliable.
Also if you have the choice of oscillator on your PIC you may choose the lower power
"XT" rather than "HS".

If you decide to switch to a resonator rather than a crystal, be forewarned that many
brands of resonator will not start up with the PIC at 20MHz.

Cheerful regards,

Bob

2005\07\14@122339 by Dave VanHorn

flavicon
face
At 10:32 AM 7/14/2005, Spehro Pefhany wrote:
>At 07:07 AM 7/14/2005 -0500, you wrote:
>
>>Can I assume in the crystal mfg's literature that "Load
>>capacitance" specifies the recommended values for c1 & c2?  That
>>would mean I'm using capacitors that are 83% bigger than recommended.
>
>No, you're using about the correct value, from that pov. The capacitors
>are effectively in series in a Pierce oscillator and there is some
>stray/input capacitance, so 33pF is about right for the 18pF load
>capacitance. That means that the oscillator will be nominally on the marked
>frequency, not a bit off. This is not your main problem.

Sounds ok to me, as long as the oscillator is on-frequency.

{Quote hidden}

You got it. Overdriving the crystal is easy to do, and will cause
some level of failures off the bat, with more coming later..



2005\07\14@123237 by Mauricio Jancic

flavicon
face
Great explanation Olin...

Mauricio Jancic
Janso Desarrollos - Microchip Consultants Program Member
EraseMEinfospam_OUTspamTakeThisOuTjanso.com.ar
http://www.janso.com.ar
(54) 11 - 4542 - 3519

{Original Message removed}

2005\07\14@134342 by Spehro Pefhany

picon face
At 12:08 PM 7/14/2005 -0400, you wrote:
>Spehro Pefhany wrote:
>>The capacitors
>>are effectively in series
>
>No, this is a common Crystal Myth.

Not a myth, but so common that it's in all the technical literature.

>If the crystal is driven with a 0 impedence sine wave on one side, the
>specified load capacitance is that value to ground on the other side that
>will cause the desired phase shift at the rated frequency.

Sure, but that's nothing that remotely resembles the real situation.

>Therefore, if
>the PIC oscillator output pin has a small impedence compared to the load
>capacitance, then the capacitor on the oscillator output pin doesn't enter
>into consideration.  In that case, the best oscillator input pin capacitor
>value would be the rated load capacitance minus the inherent existing
>capacitance due to the leads, circuit board, etc.
>
>If the PIC oscillator output had a very high impedence, then both capacitors
>need to be considered in series as you said.

The output acts like a *current* source. This is where you go astray. The
crystal is acting like an inductor in a Pierce oscillator, and resonating
with CL (the series of the capacitors, plus in/out/stray capacitance), which
is why the frequency is always a bit higher than the series resonant
frequency of the crystal.

The only hard thing is that the output capacitance of the current source
and the input capacitance need to be accounted for, so really
a few pF (5 maybe) should be added to each capacitor before the
usual CL = Cin*Cout/(Cin+Cout) + Cstray calculation.

More to the point for Mark, the drive power of the crystal can be approximated
by this forumula:

Pdrive ~= Rs * 0.5 * (2*pi*f * CL * Vdd)^2  (CL is the actual load
capacitance as calculated above using 5pF for input and output
capacitance and 3pF for stray)

So, reducing the supply voltage, using lower value capacitors
(resulting in a small shift in frequency), or using a lower Rs crystal
will all reduce the power dissipation of a given frequency of crystal.

BTW, running those numbers with the *maximum* Rs of Mark's crystal (taken
from the datasheet) and a calculated load capacitance with 33pF caps
(calculated CL = 22pF) yields an estimated maximum drive power of 3.8mW, well
above the datasheet 1mW maximum rating of the crystal. That's the problem,
Euros to donuts.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
speffspamspam_OUTinterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2005\07\14@150835 by olin piclist

face picon face
Spehro Pefhany wrote:
> The output acts like a *current* source.

That's the same as saying it has inifinite impedence.  As I said, I agree
that the capacitors are in series for calculating the load capacitance in
this case.

However I don't agree that the driver has infinite impedence.  It has CMOS
high/low output drive transistors, and will look roughly resistive.  And, I
think there is a good chance that this resistance is about the same or lower
than the impedence of the capacitor on the pin at that frequency.

Do you at least agree that if the driver has finite impedence, then just
adding the two capacitors is not valid?


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\07\14@160153 by Bob Blick

face picon face
Olin writes:
> Do you at least agree that if the driver has finite impedence, then just
> adding the two capacitors is not valid?

The driver does not have finite impedence, especially the further from DC
you get. And this is at 20MHz!

The Z of a 33pF cap at 20MHz is somewhere around 200 Ohms. The output Z of
the PIC at that frequency is at least that. And it's not capacitive.

Certainly it can be complicated if you want to make it complicated.

Nice try though.

Cheerful regards,

Bob


2005\07\14@161345 by Wouter van Ooijen

face picon face
> BTW, running those numbers with the *maximum* Rs of Mark's
> crystal (taken
> from the datasheet) and a calculated load capacitance with 33pF caps
> (calculated CL = 22pF) yields an estimated maximum drive
> power of 3.8mW, well
> above the datasheet 1mW maximum rating of the crystal. That's
> the problem,
> Euros to donuts.

OK Sherlock, you convinced me, thats the problem. But what's the
solution?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\07\14@161756 by Spehro Pefhany

picon face
At 03:08 PM 7/14/2005 -0400, you wrote:
>Spehro Pefhany wrote:
>>The output acts like a *current* source.
>
>That's the same as saying it has inifinite impedence.  As I said, I agree
>that the capacitors are in series for calculating the load capacitance in
>this case.
>
>However I don't agree that the driver has infinite impedence.  It has CMOS
>high/low output drive transistors, and will look roughly resistive.  And, I
>think there is a good chance that this resistance is about the same or lower
>than the impedence of the capacitor on the pin at that frequency.
>
>Do you at least agree that if the driver has finite impedence, then just
>adding the two capacitors is not valid?

Sorry, I don't agree, at least prima facie, that some resistance would
affect the validity of the equations-- I think the effect would be to
lower the Q somewhat, not to change the resonant frequency directly.
The output capacitance is in parallel with one of the load capacitors so it
just adds to it.

Probably the best bet to investigate in depth would be to find some very
good MOSFET models (including subthreshold) and crystal models and model
the steady-state (large signal) operation of an oscillator (and compare
with reality in terms of the external voltage swing and so on). The
theoretical analysis usually assume linear operation, which is not really
valid. I asked Microchip to consider specifying the minimum and typical gm
(transconductance) for their clock oscillator amplifiers, since it can be
used directly to predict whether the oscillator will start or not, but
they didn't seem to think that was worthwhile (I guess nobody else is
doing it...). It's easy to see that if you lower the *impedance* at
the operating frequency at the input and/or the output, that a higher
gm is required in order for the circuit to oscillate (they are roughly
in parallel for that purpose, ignoring the crystal impedance).

BTW, speaking of deviations from theory, I should add that the
equation I gave for drive power assumed worst-case voltage swing across
the crystal, which is going to be on the pessimistic side. If the actual
swing is known (say from a very high impedance probe), it can be substituted
for Vdd in the equation.

Your thesis could also be checked empirically with a frequency counter.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
@spam@speffKILLspamspaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2005\07\14@162513 by Spehro Pefhany

picon face


>OK Sherlock, you convinced me, thats the problem. But what's the
>solution?

Use a more tolerant crystal, a crystal with lower Rs, lower the
Vdd, add a series resistor, lower the load capacitor values.
All have side effects (the last three will prevent the oscillator from
starting if taken to extremes).

The easiest solution, if accuracy isn't much of a concern, is to slap a
20MHz resonator in there, as, I think, you suggested earlier. More rugged,
starts faster, accurate enough for reliable RS-232/485 communication, and
really very good typical temperature stability. But not every application
can tolerate +/-0.5% or whatever error.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
KILLspamspeffKILLspamspaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2005\07\14@171834 by Dave VanHorn

flavicon
face

>
>BTW, running those numbers with the *maximum* Rs of Mark's crystal (taken
>from the datasheet) and a calculated load capacitance with 33pF caps
>(calculated CL = 22pF) yields an estimated maximum drive power of 3.8mW, well
>above the datasheet 1mW maximum rating of the crystal. That's the problem,
>Euros to donuts.

I commonly include a series resistor on the output pin,  for
adjusting drive level, and checking oscillator margin.

2005\07\14@175357 by Wouter van Ooijen

face picon face
> I commonly include a series resistor on the output pin,  for
> adjusting drive level, and checking oscillator margin.

but how do you determine the value?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\07\14@175358 by Wouter van Ooijen

face picon face
> Use a more tolerant crystal, a crystal with lower Rs, lower the
> Vdd, add a series resistor, lower the load capacitor values.
> All have side effects (the last three will prevent the oscillator from
> starting if taken to extremes).

I agree with the resonator, but for the applications that require the
accuracy of a crystal, what would be the practical solution? Of the
things you mention only the series resistor seems practical to me (the
average tinkerer will more likely stock a suitable range of resistors
than a suitable range of capacitors, and you can't lower the capacitors
below the PCB and other stray capacitance). So what would be a practical
value for a series resistor?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\07\14@181249 by Dave VanHorn

flavicon
face
At 04:51 PM 7/14/2005, Wouter van Ooijen wrote:
> > I commonly include a series resistor on the output pin,  for
> > adjusting drive level, and checking oscillator margin.
>
>but how do you determine the value?

The trick is not to put any more drive to the crystal than you really need.
Sub in resistors, till you find the one that reliably starts in all
your environmental conditions.
(you can use a pot to get close, just keep the leads short)
Cut that value in half.

Do a worst-case check on drive level, making sure that you still
aren't exceeding the rock's spec.

Various manufacturers have slightly different versions of this test,
but this is sort of the "average" of them.

2005\07\14@183657 by olin piclist

face picon face
Spehro Pefhany wrote:
> Sorry, I don't agree, at least prima facie, that some resistance would
> affect the validity of the equations-- I think the effect would be to
> lower the Q somewhat, not to change the resonant frequency directly.

I wasn't talking about changing the resonant frequency, only how to compute
the load presented to the crystal in a given situation.  Think about the
limiting case where the oscillator output had 0 impedence.  In other words,
it's a pure voltage source.  Hanging a capacitor on its output would not
change the waveform on either side of the crystal at all.  By definition,
the voltage source would produce the same voltage regardless of what current
it takes to do that.  In that case, the crystal output capacitor would be
the sole crystal load.  This proves that driver impedence has an effect on
how the load is calculated, at least when the driver impedence is low.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\07\14@190336 by Hopkins

flavicon
face

Here's a link that may help  

http://www.oscilent.com/spec_pages/PNDescrpt/Load_Cap.htm


Another one from ECS that's a little clearer

http://ecsxtal.com/pdf/oscir_des.pdf



_______________________________________

Roy
Tauranga
New Zealand
_______________________________________

> {Original Message removed}

2005\07\14@214659 by Spehro Pefhany

picon face
At 11:51 PM 7/14/2005 +0200, you wrote:
> > Use a more tolerant crystal, a crystal with lower Rs, lower the
> > Vdd, add a series resistor, lower the load capacitor values.
> > All have side effects (the last three will prevent the oscillator from
> > starting if taken to extremes).
>
>I agree with the resonator, but for the applications that require the
>accuracy of a crystal, what would be the practical solution? Of the
>things you mention only the series resistor seems practical to me (the
>average tinkerer will more likely stock a suitable range of resistors
>than a suitable range of capacitors, and you can't lower the capacitors
>below the PCB and other stray capacitance). So what would be a practical
>value for a series resistor?

The general theme on determining the value is to increase the value
of the resistor until it no longer starts reliably, then reduce it, allowing
some safety margin for all the variations possible, and check that drive power
is well within ratings at that point. If it requires too high a drive power
to start reliably then you might have to change one or more of the other
factors.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
RemoveMEspeffTakeThisOuTspaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com




2005\07\15@015039 by Nate Duehr

face
flavicon
face
Wouter van Ooijen wrote:

> I am not into big numbers, but last week I got a 'dead' Wisp628 returned
> that was caused by a non-working crystal. This has never happened before
> on ~ 1k Wisp628's. Currently I use a resonator instead of a crystal for
> the prefab Wisp628's (the kit still has a crystal to avoid customer
> confusion).

Customers buying Wisp's are confused by the substitution of a resonator
for a crystal?  Wow...

Nate

2005\07\17@130756 by Dmitriy Kiryashov

picon face
Hi Olin.

One can drive Xtal in with precise clock generator and load the
Xtal out with several different values of precise resistors with
low inductance and low capacitance ( say recommeneded for 1GHz designs )
Phase shift ( delay ) and output impedance can be measured for let's
say 20 samples for difference frequecnies, oscillation modes, etc.


WBR Dmitry.


Olin Lathrop wrote:
{Quote hidden}

> -

2005\07\17@131110 by Dmitriy Kiryashov

picon face
Hi Spero.

What would be easiest solution if you use crystals with initial 50 ppm
and temperature deviation +/- 30 ppm ( generic ones AT cut ) to calibrate
it with 10 ppm initial and +/- 10 ppm over -40 +85 *C temperature ? :)

I guess it is possible ?


WBR Dmitry.



Spehro Pefhany wrote:
{Quote hidden}

> -

2005\07\25@145555 by Wouter van Ooijen

face picon face
> Customers buying Wisp's are confused by the substitution of a
> resonator for a crystal?  Wow...

Of course not the ones who buy the pre-build version, but you would be
amazed by what the kit-building people can do wrong unless the
instruction is very precise. I have a rather detailed building
instruction on my website. I have can subsitute black colored 1uF caps
for the blue ones shown there, but any larger deviation will generates
emails. It might be that the Wisp628 attracts the less electrically
experienced PIC wannabees.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


More... (looser matching)
- Last day of these posts
- In 2005 , 2006 only
- Today
- New search...