A bistable multivibrator has two stable states, and must be triggered (mechanically, or by the leading or trainling edge of a waveform) in order to change states. Once triggered, the circuit changes states and remains that way until it is triggered again.
Bistable multivibrator circuits are used as memories, latches, and storage and in clocked-logic blocks.
A simple bistable multivibrator can be made by cross-coupling two NAND gates:

Note: The following has the assertion levels written out so as not to break my brain.

```                                      next stable
+------+                  now       state
S ---| NAND |                S R T B |    T B
|      |o+---- T        L L - - |    H H *bad*
+-|      | |              L H - - |    H L <-set
| +------+ |              H L - - |    L H <-reset
+------------+            H H H L |    H L <-rest
| |            H H L H |    L H <-rest
+----------+ |            H H L L |  \/none\/
| +------+   |            H H H H |  /\none/\
+-| NAND |   |
|      |o--+-- B
R ---|      |
+------+
```

Let's look at what happens to this device under each of the input (S,R) combinations. When SR=LH, the "set" state, the value of T is set to H. When SR=HL, the "reset" state, the value of T is reset to L. When SR=HH, the "rest" state, the value of T stays the same. Setting SR=LL is bad because, upon returning to the "rest" state, the device becomes astable and rapidly flips between the two states marked "none". Thus, there are only two useful states: T=L,B=H and T=H,B=L, and we can refer to the state of the entire device by merely referring to the state of T. We automatically know that B is the opposite.

Note that the device does not necessarily go directly to any of the states; it may travel through up to 3 intermediate states before finding the stable one. See also propagation delay.

Also note that both "set" and "reset" occur when the inputs S and R, respectively, are brought to the low level. They are thus low asserting, and in mixed logic notation, a much simpler state transition table describing the device's useful states of operation is:

```S(L) R(L) | T(H) B(L)
----------+----------
0    0   | rest rest
1    0   |  1    1
0    1   |  0    0
```

This device is sometimes called an unclocked SR Flip-Flop or an unclocked SR Latch. It is the basic component in both the (clocked) SR Latch and (clocked) SR Flip-Flop, and thus occurs in other types of Flip-Flops, in registers and in memories. See also synchronous sequential network, digital systems, feedback, switch debouncing.

Log in or register to write something here or to contact authors.