On 7/10/2024 10:48 PM, philo wrote:
Found a P-IV mobo in my workshop and built a machine ideal for fooling around with older operating systems. Original case long gone so I'm trying to figure out the headers.
On bootup I get a message that front panel USB and sound not present and I have to hit F1 to continue, There is no BIOS option to ignore faults.
I Googled and found out that in the wiring harness, there would have been a jumper and all one need do, to avoid the error is to jump the pins.
Unfortunately I have not seen any tutorial which covers my specific machine and there indeed are many variations.
The audio and USB headers are apparently normal but there is one 16 (15) pin header that evidentiary is the one I need.ÿ The on-line help I've found just shows the model with a 20 (19) pin header.
The manual does not give specifics.
If anyone here knows, please let me know,
Obviously this is hardly important.
1 MIC Front panel microphone input signal (biased when supporting stereo microphone)
2 AUD_GND Ground used by analog audio circuits
3 MIC_BIAS Microphone power / additional MIC input for stereo microphone support
4 AUD_GND Ground used by analog audio circuits
5 FP_OUT_R Right channel audio signal to front panel (headphone drive capable)
6 FP_RETURN_R Right channel audio signal return from front panel (when headphones unplugged)
7 AUD_5V Filtered +5 V used by analog audio circuits
8 KEY No pin
9 FP_OUT_L Left channel audio signal to front panel (headphone drive capable)
10 FP_RETURN_L Left channel audio signal return from front panel (when headphones unplugged)
When there is no AC'97 sound header going to the computer case front,
the motherboard comes from the factory with two blue jumpers
from 5-6 and 9-10. You remove the blue jumpers and install
the ten pin FP cable onto the 2x5 header pins (making sure the pin 8
key aligns with the cable). Because there is a 5V supply on that
cable, you don't want to short out that pin.
We no longer find two blue jumpers today, because the motherboards
use HDAudio standard instead, and the audio chip has LineOut for
the back of the computer and HeadPhone for the front of the computer
and there is no need for thru-wiring.
But if you go back far enough, there were two jumpers. On a
machine like an HP or a Dell, the front panel cable would already
be fitted, and they would not put the two blue jumpers in a tiny
baggy for you.
It's hard to say, in your case, why there would be a sensing function
to see if the front panel cable has fallen off. The only pins they can
use for "tricks", are pin 2 and pin 4. While normally they would both
be hard grounds, an engineer could make one a logic input and make
the other a hard ground, and then arrange the front panel header/microphone
PCB to make continuity between pin 2 and pin 4 wires. This would allow
sensing whether the cable was in place or not. Retail motherboards
do not do such a thing. Never seen it. But that's the only facility I
can see on the header, that can be re-purposed. Anyone making a headphone/microphone
PCB for the front of the computer, should be joining pin 2 to pin 4 on the PCB>
The 5V pin on that cable, it may be used to power a 2 watt amplifier
on the far end of the cable. I don't know if the practice is
to fit a Polyfuse for that pin, on the motherboard. I can tell
you that the 1x4 SPKR for the beep function on a PC, that has NO fuse,
and some manuals warn that if the "red wire gets jammed in the door,
the red wire may melt or smoke". Similarly, the +12V on the fan headers
is not protected by a fuse, and overloading a fan header (past the
crude value listed in the manual), can burn out the fan power track.
This means the motherboard PCB designer has used a 25 thou track
on purpose for the fan headers, ensuring the track burns out if
you short it. Ruining the motherboard is cheaper than fitting a fuse.
The absolute cheapest piece of work I saw, some joker used a single
Polyfuse for the entire motherboard. All the 5V draw for auxiliary functions flowed through one fuse. The parallel port 5V, the PS/2 keyboard and mouse,
all the USB ports, all flowed through one Polyfuse. You're not supposed
to do that. Mainly, because the users will experience functional outages
on their board, and won't be able to figure out the pattern as to what
is doing it :-) Typical motherboards have at least four or five Polyfuses,
but there are some things (like SPKR), they just "let her smoke".
*******
The USB wiring for USB2 (from that era), is like this.
VCC VCC \
D+ D+ \___ There are enough contacts to wire the four
D- D- / pins on each USB2 black connector. A 2x5 is
GND GND / enough for a stack-of-two USB2 connectors.
SGND <=== This is a shield ground for the outside of the cable
There are enough grounds there, you could arrange one of them to function as a sensor.
Again, there is no need to sense whether the cable is present or not.
Nothing bad happens if a front panel USB2 cable assembly like that, falls off.
At one time, there were *different* pinouts for USB2. They were not
all the same. Not much later, they settled on one wiring diagram
(as far as I know). However, some older stuff, you may find some
differences. Sometimes a manual would admit what the wiring was,
but not always. A similar kind of thing happened to the USB3 2x10
header, which is used for fitting two USB3 onto the front of a PC.
The 10 pins are sufficient for the 9 contact USB3 plus 1 shield ground.
The 2x10 is sufficient for two USB3 on the front panel.
They have added other, higher quality connectors for front panel USB
since then. Since I never seem to own the proper cable for those,
I'm not the least interested in what they are :-) Some day there
will be a Thunderbird or a Thunderbolt or a USB4, and I still won't care :-) I'm funny that way.
Summary: I don't know the answer to your question, but perhaps a bit of
detective work with your multimeter set to "volts" with respect
to chassis ground, will help you figure out how they have rigged
a sensing function. For example, on the audio, with PC running,
if pin2 or pin4 are riding pretty close to +5V, then the pin
is likely a logic input instead of the analog ground the standard intends.
You will have to use your electrical craft to figure it out,
what they've done as a "trick".
The shield ground and the logic ground should all go to the same
place, and in some cases, the shield ground receives ESD discharges
in that wire. That would be a lousy place to be fitting a logic input
for a cable sensor. ESD discharges into a front panel ground, not only
can upset what is at the other end of the cable, just the inductive
coupling between tracks can couple the transient into adjacent signals.
This is one reason you have to think about whether wiring front panel
USB is a good idea or not (the fascia can be plastic and floating and
the cable becomes the discharge path).
Paul
--- MBSE BBS v1.0.8.4 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)