i found this fantastic old-skool website that explains the Electronics World from ZERO and noticed that the 74HC595 and 74HC165 are not in the list of "74 Series Logic ICs"
the 74HC595 was the first chip i started to learn using, and am now going to try the 74HC165 to get to the input-side.
Shift registers are not new.
Look at the date on the datasheet:
CrossRoads:
Shift registers are not new.
The very concept of the shift-register has been around forever. While I've never studied it, there had to be some form of a shift register (both for input and output) in early delay-line (mercury and otherwise) machines; heck, I think that technology original came from early radar systems.
I'm also pretty sure similar ideas were embedded in some of Korad Zuse's designs; I'd be willing to go back as far as Babbage's machines, mechanically (some kind of shift-mechanism was likely needed to transfer information back and forth between the mill and store, for instance).
okay, thanks - was just wondering why that guy didn't include the chips in his list.
Yes the 595 is newer than the original 7400 series, but not very much - the chips were designed in batches
in numerical order by Texas Instruments AFAIK. The 7400 series contains various
shift registers, some series->parallel, some parallel->series and some are both (74HC299). Some are
latched, some have tristate outputs, some are 4 bit, some 8, and probably some new ones are 16 bit...
(look for numbers like 74LVC16xxx where xxx matches 74HCxxx)
retronet_RIMBA1ZO:
was just wondering why that guy didn't include the chips in his list.
Maybe because the shift register was not used as much in those days.
I have been using the 74165 ( note no letters after the 74 ) since 1974.
It was much more common in the 70's and 80's to work with 8 and 16 bit wide Data and Address buses, which is why you see those types of IC's getting more focus in the past. The 74x244 74x240 74x373 and similar chips were in nearly every CPU design. Since they were doing things 8 bits at a time... there was much less of a need to serially shift data unless you were working with a peripheral device such as a Switch array or LED display. The grunt work was being done with BUS devices.
appreciate your responses as always, Mark & Mike !
(helps me maintain some semblance of sanity ... 
)
pwillard:
It was much more common in the 70's and 80's to work with 8 and 16 bit wide Data and Address buses, which is why you see those types of IC's getting more focus in the past. The 74x244 74x240 74x373 and similar chips were in nearly every CPU design. Since they were doing things 8 bits at a time... there was much less of a need to serially shift data unless you were working with a peripheral device such as a Switch array or LED display. The grunt work was being done with BUS devices.
i guess that's in keeping with printers using 'LPT1' back in those days.
'parallel' was still faster than 'serial' - hmm, i suppose it still is - but not as cost efficient ?
'parallel' was still faster than 'serial' - hmm, i suppose it still is
No, these days serial is faster. The problem with parallel is called "stuffing" and it referrers to the fact that all the lines take a different time to make a transition and this restricts the maximum transfer rate. This does not apply to serial data and so you can run it faster.
I'd disagree that serial is faster. If you can run 1 bit at 8 MHz (SPI) than you can certainly run 8 bits at 8 MHz. The problem with parallel is that it ties up a lot more IO pins and needs a lot more wiring.
I made a digital delay line when I was fresh out of college, it was all parallel data.
Sampled with a 12-bit ADC, stored it in memory, read it back from memory at 4 different times put it in 4 DACs, with pots to control the delay mix, and pots to feed back into the input for repeated echo that faded out. All parallel (serial parts not around yet then) and with a pretty fast clock. Doing that serially would have required a much, much faster clock rate. 12 bits stored with 1 clock vs 12 clocks.
Have to dig up my schematics ...
No shift registers tho, just parallel address & data busses.
If you can run 1 bit at 8 MHz (SPI) than you can certainly run 8 bits at 8 MHz.
Yes but 8MHz is quite slow, it is not what I am talking about.
Interesting discussion here.
Not all of the contributors are correct.