Issues on clockless uart

Shashi <shashi22k@rediffmail.com> wrote:

> I'm doing a project in clockless uart..as u know that the primary
> function of uart is parallel to serial conversion while transmitting
> and serial to paralel conversion while receiving..I was wondering if
> someone could tell me as how can i do a parallel to serial conversion
> and vice versa without using a clock.

I sense a serious case of "target conflict" here.  As in: trying to
make omeletts without harming roughly elliptical hulls of reproductive
products of certain species of birds (vulgo: eggs).

Serial protocol means that bits travel on the same line, separated by
nothing but time.  So you *will* need some gadget in your circuit that
measures or controls time.  Such a gadget is called a clock.  Ergo:
you *will* need a clock to generate any serial protocol.

That leaves only one reasonable way to sneak out of your conflicting
goals: you could have the clock elsewhere, i.e. in the communication
node at the other end of that serial line (and pray that your
clockless circuit is fast enough to keep up with the forced external
clock).

> Note: This is a circuit whoch doesnt use a clock at all/ Asynchronous
> Circuit/Clockless Circuit.

Then you'll have to put one in to drive the UART.

-- 
Hans-Bernhard Broeker (broeker@physik.rwth-aachen.de)
Even if all the snow were burnt, ashes would remain.

Shashi wrote:
> 
> Hi,
> I'm doing a project in clockless uart..as u know that the primary
> function of uart is parallel to serial conversion while transmitting
> and serial to paralel conversion while receiving..I was wondering if
> someone could tell me as how can i do a parallel to serial conversion
> and vice versa without using a clock.
> 
> Note: This is a circuit whoch doesnt use a clock at all/ Asynchronous
> Circuit/Clockless Circuit.

I belive this was also posted to comp.arch.fpga.  A UART sends data bit
serial with a given period per bit.  So the signal must sampled on a
timed basis in order to read the bits after synchronizing with the start
bit.  Normally a clock is used to provide the timing, but there are
other ways.  A tapped delay line can be used to delay the input signal. 
The start bit can be detected at the end of the delay line and the data
bits are read at each of the bit taps.  Once a start bit is detected,
the start bit detection must be disabled for one full character time.  

This would be a very difficult circuit to design.  But it could be
done.  

-- 

Rick "rickman" Collins

rick.collins@XYarius.com
Ignore the reply address. To email me use the above address with the XY
removed.

Arius - A Signal Processing Solutions Company
Specializing in DSP and FPGA design      URL http://www.arius.com
4 King Ave                               301-682-7772 Voice
Frederick, MD 21701-3110                 301-682-7666 FAX

On Thu, 22 Apr 2004 12:20:55 -0400, Ben Bradley
<ben_nospam_bradley@mindspring.example.com> wrote:


>>You could do it with a string of one-shots, but
>>I feel guilty just suggesting such a grotesque
>>kludge.
>
>   I thought of that too, it seems like an interesting if a bit
>bizarre design exercise, but I didn't mention it because the post
>smells too much like a homework/class assignment.

I guess giving "impossible" homework and watching what kind of stories
the students put on the internet and what discussion this will cause
might be fun to watch :-).

While a string of multivibrators or delay lines are simply to generate
a serial stream of serial stream of bits with some badly defined rate
(+/- 20 %)  and gating the start bit through a similar chain of
multivibrators (with as bad tolerances) on receive might give not so
impressive results.

Since it was not stated if this had to be done in silicon, I would
suggest looking at how the original Teletypes did it with rotating
shafts and electromagnets. These devices were fully electro mechanical
and in some more modern versions the only transistor was in the 20 mA
constant current source, but the serdes part was fully
electro-mechanical.

Of course, some may claim that the constant rotating speed of the
shaft was a clock, but on the other hand, as someone suggested using a
delay line to loop back the pulse, this also sounds like a delay line
oscillator and hence a clock.

Paul