David -
Crystals and capacitors work this way:
The two caps appear in series ACROSS the cap, and two equal
caps in series have a net value that is one-half either.
But, looking into the IC, there is a stray capacitance of
about 5pf.
So, if the crystal wants a load capacitance of 18pf (this
is the desired cap ACROSS the crystal) of 18pf, therre is
already 5pf there from the chip. That leaves 13pf to be
made up. To have a net 13pf from the two caps, each needs
to be 26pf (27pf is closest standard value).
I think (2*C)-5pf is an error and should be 2*(C-5pf).
Jim
On Fri, 15 Sep 2006 18:31:30 -0400
"David VanHorn" <microbrix@gmail.com> wrote:
> >
> > Uh-oh on the crystals and caps!!! I have always been
> using the cap
> > rating as speced by the Xtal. For instance if it says
> 18pF that is
> > what I use - I don't know enough to know otherwise. I
> would like to
> > learn more about that but don't know what "the Fox
> site" is?
>
>
> The Cl spec is NOT the value of the caps to use.
> The fox electronics site has some good engineering info,
> like this file:
> http://foxonline.com/pdfs/xtaldesignnotes.pdf
>
>
>
> Can you clarify what this means " (2C)-5pF is a good
> starting "
>
>
> See "Calculation of load capacitance" In the case where
> the caps are equal,
> the left term simplifies to 2*Cl
>
> Sorry, I am pretty much new to electronics and such and
> have only
> > been stumbling my way through some relatively complex
> projects. I
> > have made some good progress, but electronics is far
> more complex
> > than I ever imagined :-(
> >
> > Thanks again!
> >
> > Chris
> >
> >
> >
> >
> > --- In AVR-Chat@yahoogroups.com, "David VanHorn"
> <microbrix@...>
> > wrote:
> > >
> > > >
> > > >
> > > > 1) ISP, I have never used ISP yet. It appears as
> though I just
> > place
> > > > a 6 pin header on my board - wire that to the
> correct pins (SPI)
> > and
> > > > plug the cable in for programming. I will NOT be
> using SPI on my
> > > > project so it looks like those pins will only be
> used on my
> > project
> > > > as PCINT input pins. Any comments or suggestions
> about ISP for a
> > > > first-timer?
> > >
> > >
> > > You can use the pins, just put your programming
> connector so that
> > it's
> > > directly connected to the chip, and resistors in the
> signals that
> > feed other
> > > devices. You need to make sure that the programmer
> "wins" any
> > contention,
> > > easily. Also watch that activity on those pins
> won't do anything
> > bad in
> > > your hardware, like activate motors or something
> hazardous.
> > >
> > >
> > > 2) Coupling, someone suggested there could be
> coupling problems
> > with this
> > > > chip and the mounting method. I don't know much
> about
> > > > electronics so I have no idea. Should I just add
> some 1 uF caps
> > at
> > > > every power entry point to prevent this problem?
> > >
> > >
> > > Oh boy.. Power Decoupling is something you need to
> take seriously,
> > and 1uF
> > > caps are not a good choice. This depends on your
> frequency of
> > operation,
> > > and the actual caps you use, but generally 0.047 at 8
> mhz is about
> > right.
> > > If you have any choices, locate the caps at the
> ground pins on the
> > chip, and
> > > take VCC to the cap, and from there to the chip.
> (forms an
> > L/C "tee" filter
> > > using the track inductances)
> > >
> > > 3) Is there anything "Different" to watch out for
> when using this
> > > > chip as compared to the mega48,88,168 chips that I
> am familiar
> > with?
> > >
> > >
> > > Something you might not be aware of, I mention it a
> lot because the
> > > consequences are SO nasty.. The CKOPT fuse has a
> default state
> > that enables
> > > a "vittoz mode" oscillator. You definitely do NOT
> want this fuse
> > in the
> > > default state! If you scope the crystal with it in
> "vittoz
> > mode", you'll
> > > see that the amplitude is about 0.5 VCC. The chip
> will skip
> > cycles, baud
> > > rates will be a bit off (always slow) and sometimes
> the chip will
> > execute
> > > instructions wrong. Or, if you got "lucky", you'll
> have just
> > enough
> > > amplitude that everything will work. Until you get a
> new batch of
> > chips, or
> > > crystals, or have 10,000 units in production.
> > >
> > > Also, crystal caps are determined by the crystal, NOT
> the chip,
> > and a 22pF
> > > crystal does NOT use 22pF caps. (2C)-5pF is a good
> starting
> > point, see the
> > > app notes at Fox for details and a more rigorous
> treatment.
> > >
> > >
> > > [Non-text portions of this message have been removed]
> > >
> >
> >
> >
> >
> >
> >
> >
> >
> > Yahoo! Groups Links
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
>
>
> --
> Feel the power of the dark side! Atmel AVR
>
>
> [Non-text portions of this message have been removed]
>
---------------------------------------------------------------
The Think Different Store
http://www.thinkdifferentstore.com/
For All Your Mac Gear
---------------------------------------------------------------