Re: re - Wonderboard PC interface

>
> From: "Harrison, Doug" <dharrison@...>
>
> Yes, we did get a little carried away with this one. I did manage to get
> plenty of feedback, so much that I printed it into a big stack of paper
> (more trees died) and took it home to read over the weekend. Thanks to all
> who contributed.
>
> I still have some more research to do before I commit to the final
> specification. Perhaps I should clarify my goals, which have not changed.
> The plan was to design a board with enough dummyproofing to use in
> industrial applications yet simple enough to be inexpensive and quick to
> build. I see the potential to use this board as an interface with several
> parallel port step motor programs. However, CNCPro is the one program I
> feel comfortable enough with to support for retrofits. That is the program
> I am designing the board for. This is not meant to be a negative comment on
> other software and I would be happy to work with developers of such programs
> if they wish to communicate their wants and needs. The info I have on most
> of them is inadequate to work with.
>

Here is the design I threw together over the weekend.
It is not final, it is a starting point only.
There is some serious proofreading and checking to be done, but the basic
interface is all there.
I'm out of town this week.
I DO have the design and the free design package with me (Easy Trax)
The problem, is on my laptop, it can only run in 640X480 Mode and
That is only shown as about a 4" X 6" rectangle.
I'll be seriously limited in any REAL work on in.

I can generate Postscript files of the board.
I have no real way to generate an image given the current situation.
If anyone can convert it, and post is someplace everyone can see it.
let me know.

>
> The preliminary specification looks like this:
>
> 1. Separate board - not in PC. ( I may change this again if I succeed in
> solving the DC-DC converter issue. Shucks! I might even go both ways.)

My board is external to allow lots of board space to add custom circuits.

> 2. All input and output lines optoisolated.

My design will use opto-isolators on the Board to PC interface.
The Board to device interface should probably have them where
damage is likely to happen as well.
I'll need about 20, about 11 outbound and 9 inbound to the PC.

> 3. All optos socketed.

I figured on socketing all the chips.

> 4. Selection of optos to allow interchangability between cheap and high
> speed optos for step/direction lines.

The 17 Cent parts BG Micro has looked plenty fast for anythink you will
ever be able to spit out of a PC.

> 5. Sockets for optional SS relays on control lines.

I have PC mount Simgle pole Double throw Relays I planned to use.
I don't have sockets for them, but leaving holes for them along with
a big array of space to install them and wire them where appropriate
seems reasonable.
What do you mean by "SS" relay?

> 6. On board 5V power.

I figure the power and voltage requirements could vary tremendously
depending on what a person wanted to install as far as external
drive and interfaces.
110 or 220V are very likely to be used by some people.

I figured on running AC straight to the box, and installing a
surplus regulated power supply in the box, to leave all the options
open.
BG Micro, all electronics amd electronics goldmine all have a
good selection of regulated power supplies for under 10 bucks,
many are universal and can take anything from 90 - 250 Volts.

> 7. DB25 and screw term capable input - your choice

Mine has a 2 row 26 pin connector to the computer.
I have both 25 pin Din and 26 pin header connectors that simply crimp
on ribon cable.
Cut the cable to length and crimp the ends on.

Of course everyone has the option to crimp whatever type of connector
they want on the other end, even if it only extends to a connector
mounted on the outside of the case.

For the IO ports, I suggest dual connectors designed onto the board
and people can install the type they want. Ie. a standard header, and
holes to mount whatever readily avalible type of screw connector
we select.
I DO have some, but they are big, they take crim on spade type
connectors. There are 3 lugs on each pin, they are held down
by a screw, you could always remove the screw, toss the spades and
connect the wire to the screw.

> 8. DB9 and screw term options for step/dir lines.

Good idea, both designed onto the board, install whatever you want.

> 9. Screw terms for limits, e-stop and control outs.

Same as above.

>
> Several of you noted that it should not be necessary to use optos to drive
> the step/dir amp inputs because most drivers are already optoisolated. Some
> of them aren't. But the main reason I stuck to my guns here is that there
> is a tendency (and I've seen it) to use the same 5V supply for the driver
> optos as is used for the limit switches. The drivers are protected but the
> parallel port is tied to the limit lines. Not good. The data out optos
> will be supplied by the PC 5V supply. Yes, I know this means more than just
> a single cable from the PC - another argument for making the board fit into
> a PCI/ISA slot.

My design would isolate all the PC lines at the point the enter from
and exit to the PC. That should protect the PC.
People could then isolate whatever IO From they board they considered
necessart to protect the "Wonderboard".

I would just as soon keep the board outside of the PC case.

The ISA board will now work with Laptops, Microchannel PCs and many
people would be afraid to install it inside the case anyway.

All the Status and Diagnostic LEDs I have designed in, are mounted
on the board itself. They would need to be run outside the PC and be
mounted in a project case of some sort.

Marc
BTW. we are still under 10 bucks in parts.
Granted, additions will add to that cost, but the basic board is under
10, plus a power supply, and whatever case people want.