Test of ASCII waveform display

I prefer the second version.

...Jim Thompson

It's more honest. Never trust a data sheet with sharp edges on the transitions...

And it helps the youngsters understand that there are indeed indeterminate regions of digital waveforms.

...Jim Thompson

Edges!

This is OK, too...

John

I'm growing fond of the slanted edges. They are actually the most realistic, and look smooth too.

That's not what has been revealed recently, which is that some newsreaders may add or subtract leading spaces from a line. Thus, to ensure that doesn't happen, leading non-space chars prevent this.

It is considered bad form to email and post to USENET using HTML. Thus, I don't do that.

The person reading the post *can* control the font used to display their text. It is the obvious choice to use a fixed-width font to view newsgroups in which people post ASCII diagrams.

The key is for both you and the person on the receiving end to use the same font, and there's no way for you to control that. Typically, the sender will use a non-proportional font (Courier) to ensure the widest compatibility, and leave it to the receiver to configure his reader accordingly.

Leading non-space characters do not help. They either both look wrong, or both are fine depending on the font I select.

It's a matter of fixed-pitch versus variable-pitch fonts. Unless you send in HTML, you can't control the font that the other person uses to view your diagrams.

In a fixed-pitch font such as Courier, all characters are the same width. In a variable-pitch font, some letters are much wider than others (e.g., M versus l).

Everything lines up correctly now, on both my newsreaders.

The second waveform is better.

I read in sci.electronics.design that Chris Carlen wrote (in ) about 'Test of ASCII waveform display', on Fri, 26 Aug 2005:

That's better than the other, but why not:

! _____ _____ ! B: _____! !_____! ! ! __ _____ ___ ! A: !_____! !_____!

Actually, these days you can probably use | (shift-\\) instead of !.

Nice!

Thanks for the input.

Good day!

Verison 2 definitely: I want explicit edges in the diagrams, so they can be numbered too.

Sometimes I use my own "shorthand", where I use | for the original timing signal edges (the sharp clock edges) and \\ / or X for the derived signals or results. This gives a more compact text:

A B CCC D E F GGG ___ ___ ___ ___ CLK __| |____| |____| |____| |__ ___ _____________ __________ ADDR ___XXXXX_____________XXXXX__________ ____________________ ENA ____________/_/_/ \\_\\__ _______________ DATA -------------

This shows the ADDR and ENA lines changing as a result of the positive clock edge, with a timing uncertainty suggested by a slanting edge (or when longer, multiple edges). Also, it is easier to show tri-state (---) and undefined (XXX). Letters are used to indicate events or timing zones. Normally I can stay within 26...

Regards, Arie de Muynck

I use edges myself, either "/" "\\" or "|", depending on how accurate I=20 want to be. Slashes are more accurate but "|" are easier to work with,=20 especially if you need to modify the diagram. I use ASCII timing=20 diagrams (and schematics) extensively in my HDL code for embedded=20 documentation.

If you want to be sure spacing isn't a problem, do it all on one line:

B: ______------______------ A: ___------______------___

It's not as easy to read and, if you stick to "printable" ASCII, your=20 transition types are limited, but the spacing will always be right.

I've found that the following characters render properly in many=20 monospaced fonts and make 1-line waveforms quite readable:

"1" =3D =AF =3D hex AF "Z" =3D =B7 =3D hex B7 Active clock edge =3D =B0 =3D hex B0 Don't care =3D =D7 =3D hex D7

Ordinary characters: "0" =3D _ driven but undefined =3D - edges =3D \\ and / multi-bit transition =3D X

As an example, here is a segment of one of my PCI-X diagrams:

DEVSEL type =B0 =B0 =B0 =B0 "A" =B0 "B" =B0 "C" =B0 = =B0 Trdy# =AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF= =AF=AF\\______\\______\\_______________ =B0 =B0 =B0 =B0 =B0 =B0 =B0 = =B0 AD =B7=B7< ad X at X=D7=D7=D7=D7=D7=D7X D0 X D1 X D2 >=

AD =B7=B7< ad X at X=D7=D7=D7=D7=D7=D7X D0 X D1 >=

AD =B7=B7< ad X at X=D7=D7=D7=D7=D7=D7X D0 X= D1 >

=B0 =B0 =B0 =B0 =B0 =B0 =B0 = =B0 C/BE# =B7=B7< cmd X sz >=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF= =AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF Frame# =B7=B7\\__________________________________________________ Irdy# =AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF= =AF=AF\\_____________________________ =B0 =B0 =B0 =B0 =B0 =B0 =B0 = =B0 Devsel#_i =AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF= =AF=AF=AF\\=AF_=AF_=AF_\\=AF_=AF_=AF_\\______________

If this doesn't look like a timing diagram, we'll know that it doesn't=20 work for email. :(

My text editor of choice (NEdit) can do rectangular copy/paste/drag, and =

I find I can do diagrams and pictures almost as quickly as with a=20 dedicated tool. I have also defined a highlight pattern for the editor=20 and I find the result to be very readable. See a.b.s.e for a screen=20 capture of the above example.

--=20 Tim Hubberstey, P.Eng. . . . . . Hardware/Software Consulting Engineer Marmot Engineering . . . . . . . VHDL, ASICs, FPGAs, embedded systems Vancouver, BC, Canada . . . . . . . . . . .

I hadn't been to abse for a week or two, but after just now doing what I thought was a full download of recent messages I couldn't find your screen capture. Could you give message reference please?