Maker Pro
Maker Pro

Microcontroller or Memory

W

Wong

Jan 1, 1970
0
Hi,
One of my friend told me that this is a microcontroller,

Device type: EPROM/UV EPROM
Company: ST
Part number: M27C801
Datasheet at: http://us.st.com/stonline/books/pdf/docs/4377.pdf

Although I clarified to him that it is just a memory device, but he
is still insist that it should be a microcontroller.
Please tell me and my friend about the truth, any help will be
appreciated.


Best regards,
confusing people
 
D

Dana Raymond

Jan 1, 1970
0
Uhhh, I don't think so. You need a latch to store the current state so that
it will be presented in a stable way to the ROM to transition correctly to
the next state. That also implies the need for a clock of some sort.

Dana Frank Raymond
 
D

Dana Raymond

Jan 1, 1970
0
I'm sorry you are confused by this NG.
When asking the question "Is an EPROM a MICROCONTROLLER" the answer is NO,
and there is no room for confusion. A EPROM can form a part of a
microcontroller but not a whole one, by any definition.

The state machine that others have mentioned can be made from any type of
ROM, but at least two other elements (a latch/FF and a clock) are required.
In fact, the roles that an EPROM plays in a microcontroller, microprocessor,
or state machine circuit are synonymous - it holds the programmed pattern of
behavior, also known as the algorithms. But, no one would ever point to an
EPROM in any of these circuits and say "That's the
microcontoller/processor".

In fact, microcontrollers and processors contain many state machines, the
most complicated one, called the microcode execution unit, uses a ROM.
Again, no one ever points to it and calls that ROM a microcontroller.

The following site I found is interesting as it shows in a very clear way
the three elements of a state machine.
The ROM Memory map has as its address inputs the input terms A and B and the
output from the state register. The state register latches the ROM output
(state) and feeds that back to the ROM. A button on the page allows you to
generate a clock pulse to 'step' the state machine.

http://www.eng.dmu.ac.uk/~dinu/FSM.html

I hope this serves to put this beast to bed.
Dana Frank Raymond
 
R

R. Steve Walz

Jan 1, 1970
0
Dana said:
Uhhh, I don't think so. You need a latch to store the current state so that
it will be presented in a stable way to the ROM to transition correctly to
the next state. That also implies the need for a clock of some sort.
 
R

R. Steve Walz

Jan 1, 1970
0
Dana said:
This is the problem with that idea, as I see it Steve.

Asynchronous memory is unlike synchronous memory (as adding a latch and a
clock would make it) in that the databits during a read don't all change at
the same time. Its combinatorial logic, after all. This means the data bits
representing the current state will transition through other, invalid,
states before settling on the final one. However, these transition states
are tied directly to the address lines, and therefore will cause the state
machine to go off into the weeds.

Fundamentally, its the difference between the ideal and real behavior of a
circuit that makes an asynchronous ROM-based state machine impractical.

That's why state machines are always build with synchronous logic, so that
the combinatorial portion of the circuitry has time to stabilize into its
final value before the new state is considered valid.

If you can point me to an actual ROM-only state machine, I'd be very
interested in seeing it.

Dana Frank Raymond
---------------------------
State machines are built with clocks and latches merely to make them
dead-easy to program. Otherwise countless race and ambiguous conditions
have to be avoided and it takes more memory and many more steps. Also,
they have to be naturally clocked by their states, which doubles or
triples their size, and their vulnerability to the former problematic
conditions. It can indeed be done, but it is more difficult. A little
thought will enable you to make a simple one yourself, and to explore
the increase in size even a simple addition would require.

I helped write one once that with two eproms could speak my voice
out of binary storage on one, and would speak the two digit temperature
as fed it by an ADC and thermistor circuit, which answered and spoke
repeatedly when called on the telephone. But that was 15 years ago.
I'd have to hunt it up, and then I wouldn't publish it, it may STILL
be patentable.
-Steve
 
W

Wong

Jan 1, 1970
0
Hi,
Dana sounds more reasonable to me. By any definition,
microcontroller is formed by a processor and memory. It's hard for me
to accept that you can get a *stand-alone* EPROM and make it a
processor. If so, can you tell me how fast is your processor/EPROM
without a clock, Steve? I am interesting too if Steve able to show us.
 
R

R. Steve Walz

Jan 1, 1970
0
Wong said:
Hi,
Dana sounds more reasonable to me. By any definition,
microcontroller is formed by a processor and memory. It's hard for me
to accept that you can get a *stand-alone* EPROM and make it a
processor. If so, can you tell me how fast is your processor/EPROM
without a clock, Steve? I am interesting too if Steve able to show us.
------------
So, you have NEVER seen a speed rating on an EPROM or SRAM??
Think about it...

Of course the speed it will run on its own depends on the way it's
programmed.

-Steve
 
Top