LPT port output stuck high?

[Jeffery Jones]

I am trying to use an LPT port as output via a Windows driver. I have
a program that successfully controls output bits on other PCs. As
soon as I set the output, all 8 data lines go high and stay high, no
matter what I set the output to (00 or FF).

I tried manually selecting the bidirectional port control to output
with no success. I tried to select SPP in the BIOS instead of ECP.
MSD detects the port at 0x378 which is where I am programming.

What can i do?

(The port works OK with a parallel printer. Could there be a
Windows 2000 printer driver interfering with my direct I/O? I also
tried setting the printer to LPT2 but no success)

 

[John Miles]

I am trying to use an LPT port as output via a Windows driver. I have
a program that successfully controls output bits on other PCs. As
soon as I set the output, all 8 data lines go high and stay high, no
matter what I set the output to (00 or FF).

I tried manually selecting the bidirectional port control to output
with no success. I tried to select SPP in the BIOS instead of ECP.
MSD detects the port at 0x378 which is where I am programming.

What can i do?

(The port works OK with a parallel printer. Could there be a
Windows 2000 printer driver interfering with my direct I/O? I also
tried setting the printer to LPT2 but no success)

Compare the MSD result with the port's address in these three places:

1) The CMOS setup screen (if it's visible)

2) The port's properties in its Device Manager tab

3) The (little-endian) value in the first two hex bytes you get
when you do the following in a DOS box:
c:\>debug
-d0:408

If they aren't all the same, then the system may be confused as to where
the port actually resides. You may need to remove the port in the
Device Manager, reboot, disable it in CMOS, boot into Windows once to
make sure it's gone, then re-enable it in CMOS and let Windows re-detect
it.

And yes, if you can, make sure no installed printer drivers are
targeting the same port.

If your CMOS setup gives you the ability to assign the port to a
different address, you might try that, too.

It can be a real challenge to get direct LPT port access to behave
properly under Windows, but it is usually possible with patience and
inappropriate language.

-- jm

 

[Jeffery Jones]

If your CMOS setup gives you the ability to assign the port to a
different address, you might try that, too.

It can be a real challenge to get direct LPT port access to behave
properly under Windows, but it is usually possible with patience and
inappropriate language.

...Finally got it! I'll post here so future googlers can find
another thing to try. It turns out that the BIOS ignored my
suggestion to configure the port as standard parallel port and Windows
plug N play configured it as EPP somehow. And the EPP didn't respond
to SPP programming. So I found an EPP direct parallel port output
programming example and it worked!

(Took at least 3 inappropriate languages to get it to work!)


--------------------------------
// For this example, port_hi might be 0x778
// and lptPort might be 0x378
PortAccess.Output(port_hi + 2, 0x00); // switch back to mode 0
PortAccess.Output(port_hi + 2, 0x94); // epp mode, no interrupts

// Set port for direction write
PortAccess.Output(lptPort + 2, 000); // direction write

// Output some data
PortAccess.Output(lptPort + 4, outputBits);