Bob Axtell <engineer <at> neomailbox.com> writes:

> Can you provide some clues for instances where MOVs and TVSs are failing 
> for you? Seriously, I'd like to know
> why these are inadequate.

Back to 16f54 testing:

pic16f84 burned with 'blink led' code, only fast, for scope on portb.4, xtal
3.58 Mhz ceramic with internal caps, osc mode xt, wdt on or off (does not matter
for now - I don't know if the wdt is forzen when the -V is applied). Vdd 5.0V,
5V1Z and decoupling directly on Vdd/Vss. One 1 kohm resistor connected to
portb.0 . All other pins turned on as outputs after reset excepting this one.
The 1k resistor's free end was connected through 1k to a dc psu (+/- voltage
capable) and through a cap and a potentiometer to the secondary of a mains
transformer. Secondary was a few volts ac. With this I could set any ac or dc
voltage on the 1k resistor. A reset button and 10k pullup was provided on mclr.
scope points were portb.4 and xout. A dvm would meter portb.4 directly. There
was provision to connect a clamp directly at portb.4 .

test 1: ext ac and dc = 0. circuit works, pb4 = ok, xout = ok (4 Vpp)

test 2: ext ac = 0 ext dc = -5 V. circuit does not work, pb0 = - 0.5V. Using
1N4148 clamp does not help. temporarily setting dc = 0 starts chip, setting it
to -5V stops it.

with chip stopped xout shows lower osc. amplitude 'riding' on ground (i.e. Vdd
to Vdd+2V peak).

with -5V applied, used the follwing clamps to Vdd at pb0 (temperature about 15
deg C, all connections were on socket, no soldering temp.):

1N4148: pb0 = -0.56 V, chip stays dead
4.5V tvs (unidirectional glass axial leaded): pb0 = -0.64 V, chip stays dead
1N6xxx: Schottky: pb0 = -0.45 V, chip almost works (flutter)
BAT42: Schottky: pb0 = -0.37 V, chip works

test 3: as test 2 but with ac instead of dc and also clamp to Vss. dc was
supplied to center pb0 at 2.5V then ac was added until the chip stopped working
(periodically, in rhythm with ac). The noise immunity with dc = 0 was about
0.8Vpp (300 mV ish AC eff.). Similar things happened near Vdd. I assume that in
a working circuit the ac noise susceptance will cause 'clock theft' to occur and
may trigger a tightly timed watchdog.

The tvs would have to open in reverse at 0.4 V to be useful. It cannot do that
so it is useless here. According to my data here to make a zener or tvs open
forward at 1 mA and 0.4 V I'd have to heat it to 150 deg C. Thus the er a of
steam powered chips is returning.

Peter P.


-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist