Anyway I tried this debug fix to get my COM ports back
http://support.microsoft.com/kb/105940
and MSD shows all 4 now.....
But still no mouse found.
So is this a sign of a bad ISA card?
The implication in this article, is the PS/2 mouse would have to support
serial operation itself ("combo" mode mouse). The dongle in this picture,
seems to me, to be passive. (At least, at a price of $3.50, it would be
pretty hard to have an active circuit inside the dongle.)
http://www.hardwarebook.info/PS/2_to_Serial_Mouse
($3.50)
http://store.cwc-group.com/mdin6fdb9f.html
It's quite possible, your test results with the new PS/2 mouse are
bogus, because the new PS/2 mouse is not a "combo" type. It could be
that your older mouse was a combo type, and supported both. In fact, using
the dongle with the new PS/2 mouse, may have even ruined it (retest the new
mouse on a PS/2 computer to be sure).
I would test with an honest to goodness serial mouse.
(Example here - RadioShack 260-3043 serial mouse)
http://www.amazon.com/Radio-Shack-Serial-Port-Mouse/dp/tech-data/B001TJP6OQ/ref=de_a_smtd
I managed to find a serial mouse years ago, by going through the "delete barrel" at
a local computer store. I found one gem, in new packaging, in the barrel and only $5.
That's how I got mine, for testing.
Now, if you can't find a serial mouse for sale, the next step would be:
1) Open up the dead mouse.
2) Take note of wire color and solder pad each joins to.
3) Cut the cable, pull some fresh cable inside the body of the mouse.
Carefully strip the wires. Resolder to the appropriate pads inside.
4) Reassemble the mouse, which will now have a slightly shorter cable
as a result.
I've done this procedure, on a mouse with a broken wire, right where the
wire passes through the shell of the mouse. You'll need a 25 watt soldering
iron, and a bit of tin/lead resin core solder for the repair. A pair of
wire strippers, beats ruining your teeth, when stripping the insulation.
If you want, before doing the surgery, you can use a multimeter, to check
for continuity, from the old mouse connector, to the solder pads inside.
Perhaps you can detect a broken wire that way. Then, it's steps 1-4.
Paul
This is driving me crazy, in all the years I have had this PC I never had a
problem with the mouse.
So now what?
So you're convinced at this point, that the hardware works ?
Have you tried booting an alternate OS ?
If the hardware was working, it could be some kind of software problem.
Since I don't have any vintage hardware like that, I don't know
what a good alternative would be for testing. (The oldest PC I have
here, is a 440BX with a slocket in it.) I know that the
Linux OSes are being tempted to abandon older hardware, so they
might not all be equally good choices for testing. At least one
of them, has a pretty hefty minimum memory requirement (to run the
GUI).
http://www.geek.com/forums/topic/best-os-for-old-hardware
On the later Windows OSes, there is some option called /FastDetect, which may
disable serial mouse detection. I doubt an older OS would bother with
that, as a serial mouse would have been a more common type at that time.
Paul
I hate to spend money on that RadioShack serial mouse you told me about
because at $24 it is a rip off..... and my PS2 mouse was working for years
on this machine...... I am going to have to think about this some more......
Check your mouse cable. There could be a broken wire near where it
goes through the housing. While they do provide a strain relief,
nothing lasts forever.
You could test your serial port too. If you have two serial ports,
you can connect one port to the other, with a "null modem" (crossover)
cable. Open Hyperterm, type into one COM port, and the characters
should come through the other COM port. You can even use two different
computers, and one COM port on each computer, with a test like that.
If your computer uses Windows 7, and you have no Hyperterm, you can
download "putty" and use that instead. That's pretty nice (although
the documentation could use a bit of work - I had to read source
to get the right command line arguments to make it go).
http://en.wikipedia.org/wiki/Null_modem
You could even loop TX to RX on the port itself, but that's a bit
too geeky. (Also, you have to be careful about your echo setting,
as you can mistake a response as actually coming from echo. If both
echo and a loopback wire are at work, you'll see two characters for each
one typed.)
I keep a bag full of serial port adapter plugs and cables, and
even though I have quite a few, I don't have enough for every situation.
Virtually every time I go to do something, the right adapter is
missing from the bag. But in that bag, I do have enough parts, to
do the null modem thing.
Paul
Thanks for all your suggestions.... I have 2 serial mice on there way via
Ebay and
I will start testing my ports using your methods as soon as they arrive.
Sometimes problems are good because they teach me other stuff I would never
have thought of.
Thats why I like to keep this old 486 working because the new computers are
like the new cars
JUST TOO Complicated ! :-)
"Paul" <nos...@needed.com> wrote in message
news:ilsdqm$7rk$1...@news.eternal-september.org...
It's probably a matter of market opportunity. ISA disappeared, before
USB showed up, so nobody bothered to make a card ? There would probably
be enough bus bandwidth to do USB 1.1.
With FPGA technology, and a USB host IP block, you might be able to build your
own. It would likely cost a relative fortune, and you could buy a new computer
for the price instead.
There is a chipset, that goes from PCI to ISA. But I don't see any
statements about it being bidirectional, and intended for doing
the conversion from ISA to PCI.
http://www.costronic.com/W628_629.PDF
Paul
>does anyone make an ISA card that would handle a USB connection - just
>wondering
>why can't a 486 handle USB anyway, I have seen USB / SERIAL adaptors for
>newer machines...
You need ME or better to handle USB.
Win95c
Krypsis
IIRC in SE the hardware was there but without any real software
support--you needed a driver for anything you plugged into the USB
port. Once ME came out those drivers became scarse.
You know I never could use any USB flash drive on my Win98 system until
about a week ago
when I came across a Generic MASS Storage Device Driver online and it worked
like a charm.
C:\copy con >com1 >com2 >com3 >com4 >con
Everything I typed appeared back on the screen.....
Are you using a loopback connector ? How does the TXD signal
get to the RXD signal ?
At a low baud rate (with little risk of buffer overflow), you
can just connect pin 2 to pin 3 with a wire. I find it easier to
just buy a soldertail RS232 connector, and join the two pins with
one of those.
http://pinouts.ru/LoopbackCables/SerialPortLoopback9CheckIt_pinout.shtml
In some configurations, with a loopback wire between TX and RX,
the application software has "echo" enabled, while the serial port
itself receives a character via loopback. In such a case, for
every letter typed, you see the letter *twice*. So you have
to be careful when doing that test, to account for whether echo
was being used or not. You might just be seeing the echo character
provided by the computer, rather than something traveling through
the serial port.
*******
I plugged my serial mouse into COM3, fired up Hyperterm, used
1200 baud 8-N-1, and I can see garbage characters appear on the
screen if I move the mouse. That suggests the mouse is transmitting.
If I change the baud rate to 9600 on the computer end, Hyperterm
sees a stream of the letter "C" instead, implying the baud rate
isn't quite right. So the 1200 baud I was initially using (from
a previous experiment), might be close to the right value.
You could try that if you want. It's not much of a test, because
the mouse could be sending 8 bit codes, which may not make nice
7 bit printable characters on the Hypertrm screen. You need to
know more about mouse protocol, to formulate a better test.
Paul
That isn't telling you anything. The final >con rules the day.
"Take everything from the console and copy it back to the console."
That's not a serial test at all.
--
Andrew Smallshaw
and...@sdf.lonestar.org
>You could test your serial port too. If you have two serial ports,
>you can connect one port to the other, with a "null modem" (crossover)
>cable. Open Hyperterm, type into one COM port, and the characters
>should come through the other COM port.
I wrote the following DOS program to test COM ports:
http://www.users.on.net/~fzabkar/comstest.exe
You can test any input pin by looping it back to any other output pin.
Use a paperclip if you have nothing better.
- Franc Zabkar
--
Please remove one 'i' from my address when replying by email.
>You know I never could use any USB flash drive on my Win98 system until
>about a week ago
>when I came across a Generic MASS Storage Device Driver online and it worked
>like a charm.
See this thread for more ideas:
http://groups.google.com/group/microsoft.public.win98.gen_discussion/msg/a1cab28f495ef720
AIUI, the Maximus Decim driver set is kludged from Win ME. It supports
USB 2.0 host controllers.
>Ok I got my new serial mouse and still not being reconized..... just tell me
The "MODE COMx" command should tell you whether COM port x exists.
Otherwise, each COM port has a scratchpad register that is writeable
even in protected mode. For COM ports at 0x2F8, 0x3F8, 0x2E8, and
0x3E8, the respective scratchpad registers are 0x2FF, 0x3FF, 0x2EF,
and 0x3EF.
For example, the following debug session writes all ones, all zeros,
and alternate ones and zeros to the scratchpad register for COM2
(0x2F8-2FF).
C:\WIN98SE>debug
-d 40:0 f
0040:0000 F8 03 F8 02 E8 02 00 00-78 03 78 02 00 00 13 02
-i 2ff
3F
-o 2ff ff
-i 2ff
FF
-o 2ff 00
-i 2ff
00
-o 2ff aa
-i 2ff
AA
-o 2ff 55
-i 2ff
55
-d 40:0 f
-q
A more interesting example involves transferring data between two DOS
boxes in Win98SE via the scratchpad register. Normally one DOS box
"owns" the COM port, so any changes made to UART registers in that
session will not appear in the other DOS box. However, the scratchpad
register does not appear to be protected in this way.
For example, let's write 0x55 to the scratchpad register 0x2FF in DOS
box A. If we now open DOS box B, we see that the value of 0x2FF is in
fact 0x55. Now go back to DOS box A and change the value to 0xAA. If
we switch to DOS box 2, the value is now 0xAA. Similarly, any changes
made in box B appear in box A.
However, the remaining 7 registers are all protected. For example, the
Line Control register (0x2FB) configures the word format. The default
value is 0x03 which corresponds to 1 stop, 8 data bits, no parity. If
we change this value to 0x02 (7 data bits), we find that the value of
this register remains at 0x03 in the other DOS box.
One option would be to go haunt a few thrift shops (and pooter recycling
places) and snag a handful of used 9-pinner serial mice. The reason for
"several" is that you will probably run into at least one that doesn't
work.
You *might* get lucky and find an "full optical" serial mouse, M$ and
Logitech did make a few of them.
There's also the possibility you may run into a NOS (New Old Stock, IOW
still in the box with shrinkwrap) decent PS2 with a serial adapter.
Sometimes they put stuff like that in the locked glass displays.
The Local St. Vinnie's had a Logitech like that a few weeks ago.
--
"Shit this is it, all the pieces do fit.
We're like that crazy old man jumping
out of the alleyway with a baseball bat,
saying, "Remember me motherfucker?"
Jim �Dandy� Mangrum
Actually Win 95 OSR2 (the C version) did have USB support, but it was
pretty horrible for anything other than mice.
98 "Plain" was still sorta warty on USB 1.1 but did do 1.0 fairly well.