Difference between latch and buffer

Your right about a latch, it is a digital circuit that hold a state as long as power is applied. They are usually made from "D" flops and must be clocked. This type changes state with the clock edge depending on the input. They do not change state with the input state only. In other words, if the output is high and the input goes low, the output will only change to low when the next clock edge comes. Some latches use "R-S" flip-flops that chang state with a set or reset signal and do not require a clock.

A buffer may or may not have a latch in it. It's purpose is to drive a line or to drive a number of other circuits, to apply more current to a digital signal. In addition many buffers can operate tri-state where they have a high impedance mode that doesn't load other connected driving circuits. Many times they are bi-directional and can transmit or receive. When there is no internal latch, the output follows the input without a clock, some invert the signal.

Analog buffers are high input impedance amplifiers and are used for sample and hold as well as to drive lines or other circuits. The usually have unity gain. Their purpose is to reduce loading and provide isolation in analog devices. Bob

Yes, you are wrong. :-) A latch does hold the value indefinitely - they come in 'transparent' and 'D-type' or 'edge-triggered'. The transparent kind transfers its input to the output all of the time that the clock is active, and holds when the clock goes inactive; the d-type or edge- triggered holds what's there at the active clock edge.

A buffer is just a straight-through inverting or noninverting amplifier, that has greater drive capability than the circuit it's buffering.

HTH! Rich

i'll give it a shot. Latch something that holds data that was latched in at some time interval from a source that may or has already changed its value of the source since the latch state took place. in other words, once a value is latched from a source, it does not care what the value of the source changes to, the latched value will be what you see until you relatch again.

--- Buffer. I can think of 2 cases here. one, a unity gain amplifier to represent a source value with out effecting the characteristics of the source. what this means is like Load impedance or capacitance effects that could change the behavior of the source.

-- now in the digital sense. A buffer is like a cache, info is backed up in memory/registers to be read later. this way, you don't lose any data while what ever is used to read it at the most convenient time is performed. normally in these types of buffers, when data is read out of them they usually do so in either FIFO(First in First out) or LIFO (Last in First out); normally the FIFO is use in most apps., but you do have the LIFO in cases of Stack related type functions.

in this type of system, a Latch is not needed because the buffer is capturing all data for you to read when able. of course, there are limits to buffers! a good design should not allow this to exceed the limit. Latches can hold data when the latch state is done but, data will be ignored until you relatch again.

hope that helped.