Re: [avr-chat] avrdude-ATMega32 fuse bits issues

[Terry Karlson]

Neil Davey wrote:
> Hi Chris,
> You could be running into a couple of problems...  As Joerg pointed out,
> breadboards/proto boards and high speed devices generally are an unhappy
> mix if your not careful..
> They do generally have higher capacitancies than a pcb board, so you may
> have to lower, or even remove the 22pf caps on the crystal, as the extra
> capacitance in the board could be loading the crystal to much... if you

Hmmm...  let's try a quick calculation:

Take the common breadboard configuration of rows of 5 holes connected. 
Let's calculate the capacitance between 2 parallel rows and assume that 
figure at least indicates the scale of parasitic capacitances we're 
dealing with.

Yanked a row of contacts out of a breadboard: it's essentially 2 flat, 
parallel leaves (===== viewed from above), so between 2 rows we just 
have a parallel plate cap.  Separation ~ 0.5 mm.  Length of base = 12 
mm.  Now, the leaves within a row start out vertical at the base, but 
above 2 mm, deflect towards each other higher up, so here the separation 
between rows is greater than 0.5 mm (up to about 2.0 mm apart).  So, we 
could take the height as 2 mm and assume that accounts for most of the 
capacitance, but let's just double that to 4 mm just to be sure.  (The 
total height of the leaves is 7 mm.)  So, area ~ 12 * 4 ~ 50 mm^2.

I'm not sure exactly what types of plastic breadboard blocks may be made 
of, but a glance at a table of relative permittivities eps{r} shows a 
few plastics in the range 2 .. 3, so take eps{r} ~ 3.

Thus, C = eps{r} eps{0} A / d ~ 3 * 9e-12 * 50e-6 / 0.5e-3 ~ 3 pF.

(Anyone have an LC meter?)

Assuming I did that right, that looks like enough to certainly affect 
things a bit, but not necessarily fatally so...

(I have breadboarded a few micro designs, but of course ever since we've 
had internal RC clocks we haven't had to bother with the crystal until 
it was on a board.  But I _must've_ tried this at least once or twice 
(e.g. by also cutting the XTAL leads to about 10 mm), and don't remember 
it _not_ working...)

-Terry Karlson