RE: [CAD_CAM_EDM_DRO] Re: Parallel Port Interface cards

Hi Phil,

>
> You raise a fairly perplexing issue.

True. It certainly perplexed me initially.

> Since the PC parallel port uses
> open-collector drivers for outputs and does not have built-in pull-up
> resistors on inputs or outputs, any attaching device really
> should have
> them.

Only if it's being driven by Open Collector drivers. TTL or LS TTL are
totem pole drivers so pull up resistors aren't needed.

Checking out my National Semiconductor TTL data book it's true that standard
TTL drive is 16mA but we also live in a LS TTL world and it's spec'd at 8mA.
A 1K pull up is still well in line with the drive capability of one LS TTL.
For optical inputs expecting 20 or 30 milliamps from a TTL driver is pushing
it IMO.

The input current specifications on TTL devices is more on the order of 20uA
for a 74LS04 with a maximum of 1mA worst case. Wire length and current for
driving it along with noise susceptibility are all important issues.
Especially in our height noise environment. Is a parallel port interface
card placed near the drivers or near the PC?

If a LSTTL driver is providing a logic signal into pin 13 on the parallel
port, then no pull up resistor is required. It's only for OC devices that a
pull up enters the picture. But say to keep noise down we wanted to
increase the current on the wire so we're driving a solid 5mA with a TTL
device.

Unless the parallel port input pin itself sinks 5mA a pull up at the source
doesn't push any more current down the wire. When the device pulls the
signal low, the current still flows through the device resistor and device
driver. Current down the wire into the parallel port still remains minimal.
If we wanted good solid current flow through the wire into the parallel port
then the pull up resistors should be right at the parallel port. I can
envision some sort of parallel port dongle here with... a male DB25. Stick
a cable in between it and the parallel port and once again there's that
wire.

> The question of male/female connectors is debatable. I also would have
> thought the breakout board side should have a male connector,
> but research
> suggests most breakout boards have female connectors, and my
> experience
> suggests that DB25 male/male cables are easier to find than
> male/female.

I quite agree. The first design for parallel port interface cards probably
started with someone looking in their junk box for a cable and that
determined what they used. It was wrong IMHO, but somehow now everyone is
doing this. Just because by default everyone is doing it doesn't make it
right.

In my case I had to go through 6 cables before I found one. All my DB25
cables are, in effect, extension cords with a male on one side and a female
on the other. If I need to run a bit further I stick in another cord. ;-)

Anyway, running a low current high impedance circuit from the optical
sensor is never a good idea so for my Lathe Spindle sensor I have to either
buy a COTS device or build a little driver that sits right by the sensor and
can source at least 5mA. That's a good thing but a PITA. Especially since I
can connect the my current version directly into the PC parallel port and
have it work perfectly.

Maybe there's a market for a small optical sensor kit? A slotted or
reflective sensor, comparator (LM311), some resistors and a tiny board. I
have all that in stock so I guess I'll be prototyping it today.

And perhaps if noise is an issue, a small pass through dongle that plugs
directly onto the back of the PC with pull up resistors would solve some
peoples parallel port noise problems.

John Dammeyer