Prev: Maximum SATA drive size on P4V800D-X
Next: Asus P5W GH Deluxe: latest SATA RAID Controller?
From: Paul on 28 Sep 2009 12:13
GMAN wrote:
> In article <h9or96$v13$1(a)news.eternal-september.org>, Paul <nospam(a)needed.com> wrote:
>> GMAN wrote:
>>> In article <addix.3z4oda(a)no.email.invalid>, addix
>> <addix.3z4oda(a)no.email.invalid> wrote:
>>>> Hello,
>>>>
>>>> I have read your posts, and I tried, but doest work with my case. First
>>>> of all the ADH front panel drives me crazy... because I never meet this
>>>> situation
>>>>
>>>> I have the the connectors:
>>>>
>>>
>>>> I just want to make the headset work... I dont care if its HD AUDIO or
>>>> AC 97. Any suggestion, how I could make that?
>>>>
>>>> Thank you!
>>>>
>>>>
>>> Set the bios for AC97 mode, then download the following and look at page 21
>>>
>>> http://www.formfactors.org/developer/specs/A2928604-005.pdf
>>>
>>> And study here too!!!!!!!!!
>>>
>>> http://www.intel.com/support/motherboards/desktop/sb/CS-015851.htm
>>>
>>> The way it works is if its wired correctly, when you plug in your headset or
>>> microphone, it should disable the rear PC speakers, then when you unplug, the
>>> rear speakers return to normal.
>>>
>>> mic in = Pin 1
>>> gnd = Pin 2
>>> mic vcc = Pin 3
>>> ear r = Pin 5
>>> line r = Pin 6
>>> ear l = Pin 9
>>> line l = Pin 10
>>>
>>> Also when done, reinstall the audio software for the SupremeFX II card either
>>> off of the support CD or get the latest from Asus.
>>>
>>> I have heard that someone that was having problems had to reinstall the
>>> drivers after making some changes cause windows wouldnt detect some wiring
>>> changes.
>>>
>> Based on your answer, I can see my interpretation of the wire names
>> was incorrect. Normally, the cable should have "ret-r" and "ret-l",
>> which are the AC'97 return signals from ear-r and ear-l. So perhaps
>> they've changed those two signal names to something less meaningful.
>>
>> I stand by my original answer, including wiring table. The wiring only
>> needs five wires in this case. I'll go though the connector cases,
>> but the OP can ignore what follows.
>>
>> *******
>>
>> There are three header connector configurations.
>>
>> The original AC'97 header. This predates the introduction of
>> HDAudio. (The OP's Supreme FX II is an HDaudio device.)
>>
>> http://www.formfactors.org/developer/specs/fpio_design_guideline.pdf (page 20)
>>
>> AUD_MIC AUD_GND
>> AUD_MIC_BIAS AUD_VCC (+5V filtered)
>> AUD_FPOUT_R AUD_RET_R
>> HP_ON
>> AUD_FPOUT_L AUD_RET_L
>>
>> That standard supports muting as follows. The CODEC sends a line level
>> signal to FPOUT_R and FPOUT_L. The signal enters the front panel connector. A
>> side contact switch, sends the audio signal back and it gets connected
>> to RET_R and RET_L. If no headphone plug is installed in the front jack,
>> then the signal flows back via RET_R and RET_L. RET_R and RET_L are hooked
>> directly to the computer rear green colored "Line_out" jack. The muting
>> function is thus, implemented with switches in the headphone hack on the
>> front of the computer. When you plug in headphones, the switches open,
>> killing the signal on the rear green Line_out jack. This allows the headphones
>> to mute the function of the front speakers, wired to the green Line_out
>> jack on the back of the computer.
>>
>> When Intel did their HDaudio spec, not only did they invent a new pinout
>> for the 2x5 connector for HDAudio, they also redefined the AC'97 header.
>> (Which is, strictly speaking, a stupid thing to do.) That basically
>> makes it less compatible, with all possible applications of AC'97 that
>> case manufacturers may have chosen to use in the past.
>>
>> First of all, one difference with the HDaudio codec chip, is it has more
>> channels
>> on it. The "ret-r" and "ret-l" muting concept is no longer required. Muting
>> can be electronic, as sensed by the codec chip, rather than relying on switches
>> in the headphone jack. Thus, the basic "analog" signals of interest, look
>> like this. These signals exist on both the HDAudio pinout, and also
>> the HDAudio redefinition of AC'97. So you cannot go wrong, by hooking
>> analog wiring to these pins.
>>
>> AUD_MIC AUD_GND
>> AUD_MIC_BIAS ---
>> AUD_FPOUT_R ---
>> ---
>> AUD+FPOUT_L ---
>>
>> This is the HDAudio pin definition. Intel renamed the MIC and Headphone (FPOUT)
>> signals, to "Port", to emphasize the fact that the connections are retaskable.
>> You can plug a headphone into a mic jack. Or a mic into a headphone jack. The
>> codec senses this, the driver software asks "did you plug in headphones?" in
>> a popup box. The user "retasks" the jack as required, by answering the
>> questions.
>>
>> PORT1L (MIC) GND (AUD_GND)
>> PORT1R (MIC_BIAS) PRESENCE#
>> PORT2R (HEADPHONE_R) SENSE1_RETURN
>> SENSE_SEND
>> PORT2L (HEADPHONE_L) SENSE2_RETURN
>>
>> The four remaining signals include PRESENCE#, a signal you ground, to indicate
>> you're providing SENSE info. The SENSE send and return signals, are part
>> of the mechanism the front (true HDAudio) panel jacks would provide.
>>
>> Very few jacks have the isolated side contacts to connect to SENSE send and
>> return. HDAudio chips rely on impedance sensing, to detect the presence of
>> a new connection. The SENSE send and return, are tied into a resistor tree,
>> which in turn feeds an ADC converter at the CODEC. The CODEC gets a four
>> bit code from the ADC, which tells it which jacks isolated switch has closed
>> or opened. This complex mechanism is used, to reduce the number of pins
>> on the side of the Codec chip.
>>
>> So, since HDaudio has redefined the RET_R and RET_L pins, you should
>> not be using them any more, for the redefined AC'97.
>>
>> For an HDAudio motherboard which mentions support for AC'97, these are
>> the signals you can depend on. This is the picture you should find in
>> a motherboard manual. The reason the other pins are dashed, is because
>> a user should not wire all the AC'97 wires from the front panel, to the
>> old places they used to go. Only five wires are needed. You don't want
>> to upset the SENSE function, by wiring audio RET_R and RET_L to them.
>>
>> AUD_MIC AUD_GND
>> AUD_MIC_BIAS ---
>> AUD_FPOUT_R ---
>> ---
>> AUD+FPOUT_L ---
>>
>> *******
>>
>> So that leaves three header definitions.
>>
>> Original AC'97 motherboard. Wire up seven wires total. Remove the two
>> blue jumpers <--> that ship from the factory, then wire it up. HPON
>> is not used. AUD_VCC is not used. That leaves seven wires. If you see
>> mention of jumpers in the user manual, and "ret_r" or the like,
>> then that tells you the motherboard is likely a pure, old style,
>> AC'97 setup.
>>
>> AUD_MIC X X AUD_GND
>> AUD_MIC_BIAS X X AUD_VCC (+5V filtered)
>> AUD_FPOUT_R X<-->X AUD_RET_R
>> HP_ON X
>> AUD_FPOUT_L X<-->X AUD_RET_L
>>
>> The OPs case is HDAudio. You could use a real HDAudio wire harness,
>> complete with isolated side contact switches, and wire all pins. Very few
>> computer cases support this. And even a computer case with "HDaudio"
>> stamped on the connector, may not in fact have anything wired to
>> SENSE/PRESENCE and so on. A real HDAudio would require nine wires,
>> but you're not likely to see that in a real computer case. HdAudio
>> motherboards do not have any blue jumpers factory installed, and
>> no jumpers are ever required.
>>
>> PORT1L (MIC) GND (AUD_GND)
>> PORT1R (MIC_BIAS) PRESENCE#
>> PORT2R (HEADPHONE_R) SENSE1_RETURN
>> SENSE_SEND
>> PORT2L (HEADPHONE_L) SENSE2_RETURN
>>
>> Most people will be using the AC'97 variant of HDAudio, which is this.
>> Wire up five wires. As I suggested to the OP in my answer. Since
>> an HDAudio codec has enough channels to not need a mute function,
>> there is no longer a need for return wires. (The Line_out on the
>> rear of the computer, has its own audio channels.) And they would
>> only conflict with the sense function in any case. Again, no need
>> to fit jumpers to this, if removing the front panel wiring later.
>>
>> AUD_MIC AUD_GND
>> AUD_MIC_BIAS ---
>> AUD_FPOUT_R ---
>> ---
>> AUD+FPOUT_L ---
>>
>> HTH,
>> Paul
>
>
> If you do the last wiring diagram you list with 5 wires, the rear speakers
> will not shut off when a headset is inserted in the front jack. If his case
> has the extra return wires why would you suggest he not wire the in the proper
> way?
>

They will, because impedance sensing is used to determine when things are
plugged and unplugged. The mute function is electronic and implemented in
software. Some HDaudio even have software support for cloning, as there
are some users who desire headphone drive and lineout drive at the same
time (with the same stream content). Because the function is implemented
in software, all of that is possible. When you have bidirectional widgets
on all ports, the only limitation is that imposed by the software. Since
there is jack retasking, a user can even plug a headphone into the mic
port, and a microphone into the headphone port. (The software design
intent seems to be, to duplicate how AC'97 works, presumably so
customers will not get confused. For example, I don't see an option
to drive outputs on all ports at the same time, in software.)

+---------+
Port1L_Mic -----| |
Port1R_Mic_Bias -----| HDAudio |
| Codec |
| |
Port2R_Headphone_R --| |--Port_For_Lineout_R --- Green jack
Port2L_Headphone_L --| |--Port_For_Lineout_L --- on the back
+---------+

The "real AC'97" case was done with switches. And blue jumpers were
provided, as a factory default for the header, to take the place of the
switches, and ensure the signals made it to line_out when a user had no
wiring on it. On AC'97 codecs, there were fewer ports, so one port was
stretched to perform two functions.

+---------+
Mic ---->| |
Mic_Bias ---->| AC'97 |
| Codec |
| |
Headphone_R --------+--<----| | +----- Lineout_R --- Green jack
Headphone_L -----+-----<----| | | +-- Lineout_L --- on the back
| | +---------+ | |
| | / | |
| +---------+ +----------+ |
| |
| / |
+----------+ +--------------+

Switch inside headphone jack (when wired up)
Blue jumpers inserted (when not wired up)
>
> If you notice, they kept the same pinouts for HD Audio as AC'97 to avoid
> header issues on cases
>
>
> So the pins i specified in my post still stands even if the OP is leaving HD
> Audio on in the bios, and what i stated still stands that he should reinstall
> the drivers after the changes.
>

While an AC'97 computer case with a fixed 2x5 assembly (not separate
wires) can be connected to an HDaudio header, it means the audio return
signals from the front jacks, are now wired to the SENSE resistor
tree. Since the port output impedance is somewhere between 1 and
32 ohms, that tends to override the sense resistors much higher
value. There likely isn't a lot of side effects on the audio signal
itself. I presume the reason the HDaudio codec doesn't go nuts, is
because the ADC on the resistor trees is turned off. I don't know
all the tiny details, as in, how does a BIOS setting for the header,
change the registers in the HDaudio codec. Is the header state
(HDaudio mode or AC'97 mode) passed in a BIOS table ? Or is a register
in the HDaudio written such that the resistor tree sensing is turned off,
and the chip is doing impedance sensing only ? The fact there is a
BIOS setting for this, implies some form of communications or control.
The only thing I see that might need changing, is turning off
the SENSE function, when the BIOS is set to [AC'97 header mode].

If you have loose wires for your front panel wiring, it is better to
just not connect the RET_R and RET_L wires, and not end up connecting
them to SENSE pins. But if you have no choice, I don't see the harm if you
do happen to connect them. It just doesn't do anything. So you
can connect them if you want (RET_R, RET_L on HDAudio). It means
your audio outputs to the headphone, are driving a tiny
resistive load, which they can do with no problem. They're
typically rated to drive a 32 ohm load.

The problem with HDaudio documentation, is it is written for
software people, and there is no way to relate what is done,
to registers. It is all hidden from view.

Paul
First  |  Prev  | 
Pages: 1 2
Prev: Maximum SATA drive size on P4V800D-X
Next: Asus P5W GH Deluxe: latest SATA RAID Controller?