Smoking Allegro and blanking time

Well guys I have a progress report on getting my PWM drive going.
Well for a while anyway, good while it lasted. I am going to use them
to power 3.4 amp steppers for a wood router I made. If you recall Jon
gave me some good info to get the right delay between the bridges.
I got a drive put together and it seemed to work pretty good, turned
a motor about 60 rpm before starting to jump around. Then I thought I
would play around with the off time of the drivers. The fragrant
aroma of fried chips instantly wafted through the air. Whoops! That's
right, zapped all 4 of those puppies. Good thing I had ordered
extras. So I thought maybe there was something else here I should ask
about.
Allegro UDN2961W's can be set to run in slow decay mode or fast
decay mode. In slow decay mode only the source drivers do the
chopping.
I set them for slow mode because the documentation says device power
dissipation is minimized in this mode. This to me says it will cause
less heating of the chip. Is this right or do I have this backward?
In case anybody wants to read about this it is here:
http://www.allegromicro.com/datafile/2961.pdf
Originally I set the off time in the middle of the range, 470pf and
56k going to the RC pin,(pin 8 on the 12 pin sip), and it seemed to
work ok. Then reading further in the documentation on the blanking
time it says "The frequency of the PWM current
control is determined by the time required
for the load current to reach the set peak
threshold (a function of the load characteristics
and VBB) plus the OFF time of the
switching driver(s) (set by the external RC
components).
To prevent false resetting of the flipflop,
due to switching transients and noise,
a blanking time for the comparator can be
set by the user where tB » 3600 x C in the
slow-decay mode or tB » 2400 x C in the
fast-decay mode. For C between 100 pF
and 1000 pF, tB is in ms."
Ok so I read this and thought false resets would mean missed steps
so I better try this. I plugged in 3.6k and 470pf and damn if it
didn't ruin my whole day when things went up in smoke. So I think I
am learning from this that off time and blanking time are 2 different
things, and this was just the calculation, not what the actual value
for the resistor should be. But I am still confused as to where to set
the RC values for best performance, slow decay or fast, and no false
resetting of the flip flop(I think that would be a bad thing).
Help!
Lonnie