Sleepy said:
Hi - Im currently using the onboard Realtek ALC888 sound from my
Gigabyte mobo
but wanted the clearer sound of a dedicated card. I also use the front
panel connectors
regularly - I plug in a headset to play Left 4 Dead.
My old Audigy 4 was no good (no front panel connection) and I bought the
X-Fi which does have a front panel connection.
It works but the X-Fi doesnt detect when the headset is plugged and
unplugged.
The Realtek chip promptly detects whenever I plug the headset in and
when I disconnect it.
For functionality the Realtek is perfect - its just that the sound
quality is a little lacking in range
compared to the Audigy or X-Fi. Does anyone have any tips on getting an
X-Fi to work better with
switching connections?
A possible manual for the sound card. The usual Creative CHM help file.
(
http://support.creative.com/manuals...prodName=X-fi Xtreme Audio PCI Express Manual )
http://ccftp.creative.com/manualdn/Manuals/TSD/10539/0xBDAB0661/Xtreme Audio PCIe Manual.chm
The manual says the header is strictly Intel HDaudio.
"Connection compatibility
Intel HD Front Panel Audio standard only
Not compatible with AC97 or Intel-compatible HD Front Panel Audio
Pin Signal Name Description
1. PORT 1 Analog Port 1 - left channel (Microphone)
2. GND Ground
3. PORT 1R Analog Port 1 - right channel (Microphone)
4. PRESENCE# Active low signal that signals BIOS that an Intel HD Audio
dongle is connected to the analog header.
PRESENCE# = 0 when an Intel HD Audio dongle is connected
5. PORT 2R Analog Port 2 - right channel (Headphone)
6. SENSE1_RETURN Jack detection return for front panel (JACK1)
7. SENSE_SEND Jack detection sense line from the Intel HD Audio
CODEC jack detection resistor network
8. KEY Connector Key
9. PORT 2L Analog Port 2 - left channel (Headphone)
10. SENSE2_RETURN Jack detection return for front panel (JACK2)"
There are many possibilities, as to how strict their interpretation
might be.
PRESENCE# would be a status signal, that says a compatible front panel
wiring assemble is present. Hardly any computer equipment bothers with that.
Strictly speaking, they could ignore your case wiring, if that wasn't connected.
Grounding the pin, says an HDaudio wiring assembly is present.
The SENSE_SEND, SENSE1_RETURN, SENSE2_RETURN, are part of a jack
detection network. It relies on an isolated side-contact switch
on the front panel microphone and front panel headphone jacks. The
connector would have Tip, Ring, Sleeve, plus the two switch contacts.
Hardly any computer cases have the right jacks (maybe only a couple).
(PDF page 172 shows how the switches on four jacks are wired to SENSE.
The resistors are 2x the value of their neighbour. Front panel audio
has two jacks, meaning two rear jacks could complete this "tree".)
ftp://download.intel.com/standards/hdaudio/pdf/HDAudio_03.pdf
The way SENSE_SEND, SENSE1_RETURN, SENSE2_RETURN works, is when the
switch on a jack closes, it is connected to a resistor. The sixteen
possible DC voltages that result from the arbitrary closure of up
to four switches, is converted by an ADC on the chip, to a four bit
digital code. Using this technique, a chip needs only one (analog) pin, to
sense the state of four jack detection switches. So this is a way
to conserve pins on HDaudio chips.
Now, on regular motherboards, an alternate technique is used. It
would appear that many HDaudio chips have the optional impedance
sensing function implemented. When a user connects an AC'97 front
panel header (large percentage of computer cases), instead of
relying on SENSE_SEND, SENSE1_RETURN, SENSE2_RETURN, PRESENCE# etc.,
the chip makes a direct measurement, to see if an electrical load
is present on Tip, Ring, Sleeve of the audio connector.
So the Creative .chm help file seems to be saying, that they
rely on the additional cruft SENSE_SEND, SENSE1_RETURN, SENSE2_RETURN,
PRESENCE#. Now, if that was true, you might not have heard any audio
in your front headphones. So not all the symptoms add up. On the
one hand, there are two methods for detecting when a user has
connected or disconnected front panel connectors, and yet you've
managed to get some sound from your front panel. They could, for example,
make the ports completely dumb (i.e. just buffer the audio that
was going elsewhere, and connect it to the jacks). But then, they
wouldn't need to make that statement at the beginning of the
section I copied from the CHM help file.
A diagram of the wiring of HDaudio front panel jacks is available here,
Figure 6, PDF page 25. The EMI filter stuff in gray is optional. If
you wanted to make a strict HDaudio wiring harness, that would be
how to do it. The fun part, is finding the jack with isolated side
contact. (You can always fake it, by using a separate switch, and
operating the switch when you want the port to appear it is in use.)
http://www.formfactors.org/developer/specs/A2928604-005.pdf
You can get a simple jack, without isolated side contact switch, like
this one. Two of these, two switches, could be used to make a fake
HDaudio cable assembly. You'll need some female pins to attach to
the HDaudio header etc. (I was hoping frontx.com would have an
HDaudio wiring assembly, but no such luck.)
http://www.radioshack.com/product/index.jsp?productId=2104044
I really have no way of knowing what Creative is up to. Maybe there
is a real HDaudio chip on the sound card. Or maybe they're just
copying the standard and insisting it is all necessary. I can't
tell from here, what their intent is. If they had half a brain
cell left in their heads, they'd make their interface compatible
with 99% of computer cases out there. But I guess that is too easy.
(Computer cases have mainly AC'97 wiring.)
There are experiments you can try - but that would require
fiddling with SENSE_SEND, SENSE1_RETURN, SENSE2_RETURN, PRESENCE#.
I doubt any one in Creative tech support would help, but you could
try asking them what the real deal is.
Paul
