Hmm. I suddenly got worried about my aging memory banks, so I just pulled down
the SN74HC595 datasheet from TI. It says +/-6mA output drive at 5V on the front
page! I think that's where I got my 5-6mA figure -- it stuck in my mind that
way.
It also gives an absolute package maximum of 70mA continuous and a continuous
output maximum of 35mA. But, as they also say, operating at these absolute
ratings is wrong-minded -- they are only to tell what maximum stress they can
tolerate before being damaged, not operate.
Did I miss something, then?
Jon
My numbers were based on typical curves in the Motorola HC handbook,
confirmed by some measurements made some time back. The smaller
numbers you quote are probably guaranteed values for sustaining a
legal high or low level.
Hmmm (stands on table to reach dusty bookshelf) here's the TI 1988 HC
databook. Voh is typically 0.2 below Vcc at 6 mA load, which gives us
an output resistance of 33 ohms in the pullup direction. That would
extrapolate to 150 mA into ground, except it's nonlinear so you really
won't get that much.
One appnote in the back shows an HC04 dumping about 42 mA into ground,
20 mA into 3.8 volts.
Oh, cool, page 4-50 has the output curves. Standard HC has an initial
pullup slope (near Vcc) of 56 ohms, and short-circuit output of about
42 mA. The bus drivers are more like 26 ohms and 75 mA, typicals of
course.
So these parts theoretically exceed their own max current ratings when
shorted. I guess long-term damage is theoretically possible,
electromigration or something. Chargin a cap once in a while should be
no big deal. You could charge a farad in a minute roughly.
John