RJK wrote:
> Oh thank you Paul ..HUUUUUUUUUUGE amounts of praise go out from here to you
> :-)
> ...you are indeed a PC Guru !
>
> I think you've confirmed what I've been pondering, over the Mic VCC bias:-
> ...(two thirds way down the page - some example Asrock pin-outs)
>
> http://www.driverheaven.net/audio-ge...nboy-case.html
> ...which suggests that Asrock usually wants that Mic. bias tail connected to
> Pin3.
> but, I don't like uncertainty where 5v? is flying around !
>
> ...much thanks for your response.
>
> regards, Richard
>
MIC_BIAS is a current limited source of power for electret microphones.
In the past, the MIC_BIAS might be a 5V supply with a 2K ohm resistor in
series. That means the electret can get no more than 2.5mA of current.
You could safely short MIC_BIAS without ever seeing a spark or having
damage.
On an AC'97 header, there was also a pin with a raw 5V supply on it.
That one would be more dangerous and I'd stay away from it. Now, it
may well have a fuse to protect it, or it could even be coming from
a small three terminal 5V regulator, in which case it has thermal
overload protection. On a really cheap motherboard, I'd be more nervous,
as they may have tried to save a few pennies, by not having any protection
at all.
If you go to page 20 here, you can see the old AC'97 header standard.
Pin 3 (MIC_BIAS) is safe. For pin 4, I would make no connection. Pin 4
is typically used, if there was, say, a 2 watt amplifier mounted on
the front panel audio interface PCB, and then pin 4 is a stronger
source of power.
http://web.archive.org/web/200404070..._guideline.pdf
Previous generation AC'97 header pinout
MIC2 X X AGND
MIC_PWR X X 5V_analog (intended to power headphone amp on front panel)
Line_out_R X X Return_R ( X---X jumper if not used )
NC X
Line_out_L X X Return_L ( X---X jumper if not used )
The next spec in the series, might be something like this. On
page 20, they decide to redefine the AC'97 header. Notice how
they redefined a few pins, which is not always such a
clever thing to do. Page 23 shows the HDaudio pinout, and
the redefinition on page 20 then makes slightly more sense.
(They're trying to retrofit both methods to the same header.)
http://www.formfactors.org/developer...928604-005.pdf
Now, formfactor documents are written by Intel, and some of
their ideas are copied verbatim into hardware, and others
receive slight modifications from defacto practice. This
is a typical Asus header definition, for a motherboard that
supports a BIOS setting for either AC'97 or HDaudio pinout.
Really, all the HDaudio is doing, is considering the sensing
pins, which are the four extra ones. Because the chips seem
to have impedance sensing (a second way to determine a plug is
inserted), those four pins aren't really that necessary.
AC'97 HDAUDIO
MIC2 X X AGND PORT1_L X X AGND
MIC_PWR X X NC PORT1_R X X Presence#*
Line_out_R X X NC PORT2_R X X Sense1_return*
NC X *Sense_Send X
Line_out_L X X NC PORT2_L X X Sense2_return*
* Part of the HDaudio jack sensing based
on a side-contact switch inside each jack.
Defacto practice is these are optional
The reason the "port1" and "port2" naming convention is used, is
because HDaudio is retaskable. It means you can plug a microphone
into the headphone port, or a headphone into the microphone port,
and the software takes care of the details. All ports have programmable
microphone bias, so an electret can be powered on any port. The power
source can be a tiny bit weaker than the old way of doing it.
Notice that the "Return_R" and "Return_L" is not used in the
new definition. That is because the HDaudio codec has more channels,
and drives the internal Lineout jack directly, without needing to
get a signal from the front audio. Muting is implemented in software,
so the user can control whether plugging in front headphones, causes
Lineout on the back of the computer to be turned off. With the old
way of doing things, you had to connect Line_out_R to Return_R and
Line_out_L to Return_L, to get the rear green Lineout to work.
The only time I have a slight concern with the audio wiring, is for
those computer cases that provide a 2x5 shell with all the wires inserted.
The loose wire method is better, because then it is easier for me to tell
people to just hook up the five necessary wires. If you use a 2x5 shell
and leave all the wires in place, and then connect to an HDaudio motherboard,
it means the sense wires can get hooked to the audio signals. Probably
not deadly, but not that nice either. (Maybe you'd hear more of a click
or pop when plugging in stuff.) The 2x5 shells can be rewired, by
lifting a tab on the shell, and sliding the wire out. But when you do that,
then the next person to use the case won't know what the hell is going on :-)
Just some trivia for future reference,
Paul