Manchester Decoder

If you detect transitions of either direction, and fire a non-retriggerable one-shot of 0.75T duration, you'll recover an unambiguous clock. The rest is easy.

John

"transitions of either direction" ??

Now I'm lost :-(

Please elaborate for the analog guy ;-)

...Jim Thompson

Manchester encoding has:

The line always transitions in the middle of the bit. If some bit(n) is the same as bit (n-1) then do _not_ transition the line at the start of bit n. If bit(n) and bit(n-1) are different, then transition the line at the start of bit n.

As there's no guaranteed transition (low - high, high - low) at any position, the way to recover the clock is to capture both transition edges.

So to paraphrase John, set up a non-retriggerable one-shot with 0.75T and make it trigger on a high-low _or_ low-high transition direct from the data.

Good picture at

There are one-shots with dual triggers available. Try this one for size:

:)

Cheers

PeteS

I'll admit to a major brainfart

The encoding is:

If zero, then the bit will transition from high to low in the middle of the bit, and a one will force a transition from low to high in the middle of the bit.

My apologies (my mind was on something else I am in the middle of doing)

Cheers

PeteS

'Pure' Manchester code can have pathalogical groups of zeros or ones making decoding impossible. You also need to either use either Async (start and stop bits), or Sync (leading known sync byte) to be able to decode reliably.

Luhan

Now that I recall, my earliest one loaded an assembler program into a wirewrapped 8080 machine using a version of Manchester. The only hardware was a single comparator connected to the headphone jack of a portable cassette player. This fed a variable width bit directly into an input port. The decoding (as well as the encoding) was all done in software.

Luhan

try

for manchester to NRZ conversion, then NRZ to binary is easy.

Steve Roberts

See abse.

John

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

Over-sampling:

"Jim Thompson" schreef in bericht news: snipped-for-privacy@4ax.com...

Well, Jim,

Suppose the digital logic is below your imagination. To make things easier an every day example. It's like an airplane: When it's down it may go up, when it's up it will go down. No other directions available :-)

petrus bitbyter

John Larkin a écrit :

If that's too simple, just think it at the transistor level :-)

Wowie Zowie, 20 pins, needs an externaly provided clock, and cost just $17!!!

Little early in the season for this turkey.

Luhan

On 30/10/2006 the venerable Jim Thompson etched in runes:

Here's some application notes:

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

Check OpenCores.org. I believe they have an open Decoder in VHDL.

Cheers

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

Heres an xilinx appnote you might fine usefull.....

What's VHDL? Does it have a VBE ?:-)

...Jim Thompson

-- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona Voice:(480)460-2350 | | | E-mail Address at Website Fax:(480)460-2142 | Brass Rat | |

Very Hard to Deceipher Language. It's a way to make a simple schematic into scores of pages of gibberish.

John

...Jim Thompson

In the early days, Tarbel Electronics produced cassette interface boards for S100 computers. They used a modification of the Manchester code. 1's produced a full cycle at the clock frequency, 0's produced a half cycle at half the clock frequency...

Decoding was more deterministic. You got a 'carrier' of all ones (clock) followed by an initial zero (start bit) which was easily detected being twice as long (independant of phase). The data was decoded as async bytes with 2 stop bits (to give your software time for storage and other overhead).

This was the system I duplicated on my 8080 system. It got one hard error in the first year of operation using a standard audio cassette deck.

Luhan