what i went through dealing with my mobo with dead onboard sound!

[momaka]

Re: what i went through dealing with my mobo with dead onboard sound!

your wrong
...
putting a series resistor will only cause a flat level drop - it wont help or worsen the frequency response

Nope, quite the opposite (to both statements).

I'm sure you know what Damping Factor (DF) stands for. But if not / for other readers: DF = ratio of the load (speaker/headphone driver) impedance divided by that of the amplifier's.

Why DF matters?
A speaker drivers is an electro-mechanical system. Ever tapped on different speaker drivers before and noticed how each produces a different "thud" sound? Well, that's due all the various characteristics of the driver - i.e. cone material, cone weight, spider and surround tension / returning strength, and etc. The speaker cabinet will also affect this quite a bit (though with headphones, this is probably not that important.) The point is, when you tap on a driver, you displace it from its idle/resting position. The spider and surround tension return the driver back to its idle/resting position, but not instantaneously - that's physically impossible due to the weight of the driver. So the driver will "wobble" back and forth "a little" before returning to its idle/resting position. Essentially, it's analogous to resonance ("ringing") in R-L-C electrical circuits, and every speaker has its own distinct characteristics. Now, when this "wobble" back and forth happens, don't forget the voice coil is moving through the magnetic field in the speaker's pole piece, thus generating back-EMF. If you provide a low-resistance path for this back-EMF, you can greatly reduce the "wobble" self-resonance of the speaker driver / voice coil... and that's the key here! - you want an amplifier that has as low of an output impedance as possible, so that any self-resonance "characteristics" of the speaker driver (and its cabinet) are reduced to a minimum. The idea is not for the speaker driver and its cabinet to add too much of their own self-resonant characteristics... through it's impossible to reduce them down to nothing. And this is why no two speaker drivers or cabinets ever sound the same.

Now if you wonder what the above has to do with headphones... well headphones are pretty much the same thing as regular speakers, but just a smaller version. You might think the small mass of a headphone driver is insignificant, but it's not! Same physical principles apply as to regular speaker drivers, and the results are measurable.

With that said, here is one of the studies that I found a while back on this matter (I built a headphone amplifier for my senior project in uni, so this was part of the research that went behind it.)
https://www.badcaps.net/download/?id=9c3883f3

If you've read through the above study, you can now see how a series resistor with the speaker/headphone driver can actually impact its sound characteristic. Note that more output "distortion" doesn't necessarily mean you get less bass output. If you look at Figure 5 of that study, you can see that the bass output from 40 Hz to 160 Hz increased by more than 0.3 dB with the series resistance driver than without it. Sounds like a good thing, but it's not - it's the headphone's own driver that's adding its own self-resonance to the output sound. If you like more bass, then that's a good thing. But what if you use a different pair of headphones that has the opposite effect? Then you could loose bass output. So the series resistor could work in your favor or completely against it, depending on the headphones used... which is not a good thing. If you want "repeatable" results and less deviation between different headphones, then getting rid of the series resistance is the key.

So going back full circle...
You are 100% right in wanting the lowest possible ESR when it comes to output coupling caps, because that will work to reduce the series resistance on the output. But again, with virtually all motherboards using series 10-100 Ohm resistors, your drop of a few milliOhms using low ESR caps cannot produce measurable difference. If it did, then the old output coupling caps really must have gone very high ESR (more than 5 Ohms.) And even then, you'd only get a measurable difference if you drive a low impedance load directly (i.e. headphones.) An amplifier with an input impedance in the KOhms range will simply not care the slightest bit how bad / high of an ESR those output coupling caps have. However, leakage current becomes an even bigger factor here... and that's why it's important to use caps with low-leakage current in audio. So "audio grade" caps are not complete "audiophool-ery"... but they are quite over-priced for what they are and won't make a difference in every instance - that will depends on the circuit design too.

And yes, I'm speaking from experience here, too. I've upgraded the caps on a few motherboards from the original 100 uF GP series to 220 uF low-ESR (Nichicon HE / Chemicon KY.) When driving headphones directly - yes, it mattered, because the higher capacitance of the output coupling caps reduced the low cut-off frequency approximately in half. But when driving my headphone amplifier, it didn't make squat of a difference.

[stj]

Re: what i went through dealing with my mobo with dead onboard sound!

your wrong

i learned by experience.
low esr caps have a flatter frequency curve
shit caps lose bass and treble, good caps give a fuller range.
putting a series resistor will only cause a flat level drop - it wont help or worsen the frequency response

you can really tell if your playing something such as regae or other music with deep bass.

[momaka]

Re: what i went through dealing with my mobo with dead onboard sound!

lower esr = better response.

In theory... but in practice, this won't make a difference at all.

#1 reason is because many motherboards usually put 10 to 100 Ohm resistors is series with the output caps. So decreasing the ESR of the output coupling caps by a few milliohms won't even produce measurable results, regardless if you use 8-Ohm impedance drivers or 32-Ohm impedance headphones, or drive an amplifier with an impedance in the KOhms range.

When it comes to driving headphones directly, those 10-100 Ohm resistors are the biggest nuisance and are responsible for pretty much all of the distortion and "coloration" of the sound. But you can't remove them, because then you risk shorting the output of the IC when the jack is unplugged or if a defective 3.5 mm cord is used. Thus, even with a shorted cord, those 10-100 Ohm resistors are there to protect the IC outputs from overload.

Another point is that for output coupling caps (and especially also any input coupling caps), lower leakage current spec is actually way more important.

High leakage = higher noise and possibly DC offset.

In the case of amps with output coupling caps, the DC offset won't matter much... but still, higher leakage caps tend to be noisier.

And that's really the whole idea behind "specialty" audio caps - they are just GP caps with lower leakage.

Film caps would be nice indeed... but like you said, not enough space.
Also, if you do use film PP are the best choice and most linear frequency response.

[stj]

Re: what i went through dealing with my mobo with dead onboard sound!

lower esr = better response.

i know tants arent ideal, but there is not enough room to fit film caps.

[lti]

Re: what i went through dealing with my mobo with dead onboard sound!

Tantalum caps would be fine if they're for DC blocking (which is what coupling caps are for), but I've seen some cheap consumer audio equipment use polarized electrolytics in filter circuits with no DC voltage across them. There probably isn't enough current available in that case for a tantalum cap to explode, but it still isn't a good design.

I don't know about tantalum caps improving frequency response, though.

[momaka]

Re: what i went through dealing with my mobo with dead onboard sound!

have you looked at a datasheet or example schem for the sound chip?

No, I just used a multimeter to find the jack detection pin/switch on my motherboards. The A8V-MX I'm pretty sure I remember it had jack detections. The Gigabyte one was bought dead, so I have no clue.

That said, since you did mention the datasheet... yes I just looked at it, and YES, there is a jack detection circuit for the ALC889a. It's done via resistors getting shorted to ground by the audio jacks and the ALC889a detects which jack it is through pins 13 (Sense A) and 34 (Sense B).

But we don't need to know all that, because one just has to find which pin on the audio jack shorts the corresponding jack resistor to ground. And when that is found, bridge this pin to ground and the jack will always be enabled. SIMPLE.

[stj]

Re: what i went through dealing with my mobo with dead onboard sound!

have you looked at a datasheet or example schem for the sound chip?

[momaka]

Re: what i went through dealing with my mobo with dead onboard sound!

modern sound chips dont use switches, they read the resistance of the cable to determine if it's headphones (32ohms or less) or line or open

Perhaps that's the case on the newest gen motherboards.
But just for a sanity check, I pulled out a fairly modern AM3 motherboard with front HD audio (Gigabyte GA-78LMT-USB3)... and on the rear audio jacks, it uses a detection switch. Same with a much older ASUS A8V-MX that uses legacy AC'97 audio. I did the pinouts on both, so will post them here later today or tomorrow when I process the pictures (lol makes me sound like I used a film camera... but really, it's the annotation in Photoshop and whatnot... if I even do it this way.)

i always suspect the block of tiny electrolytics that always seen to be clustered around those chips.

They almost never go bad... and even when they do, the audio just gets either garbled or noisy, depending if they are doing filtering or just used for coupling audio signals to the jacks.

i replaced them with tantalums once to improve the frequency response.

That's not a good idea at all. Tantalums do NOT like reverse voltage.

Again, depending on function of the caps in the circuit, I would not recommend Tantalums on the audio output as coupling caps at all. Granted a set of standard 32-Ohm headphones will be pretty tame in terms of loads... but again, you can still end with a scenario where these caps see reverse voltage and then , there goes the audio chip output when the caps go short-circuit.

[stj]

Re: what i went through dealing with my mobo with dead onboard sound!

modern sound chips dont use switches, they read the resistance of the cable to determine if it's headphones (32ohms or less) or line or open

i always suspect the block of tiny electrolytics that always seen to be clustered around those chips.
i replaced them with tantalums once to improve the frequency response.

[momaka]

Re: what i went through dealing with my mobo with dead onboard sound!

Sry in advance if I dig up some older threads. Trying to see what I missed in the last half year or so since I was gone.

... i powered on the system and noticed there wasnt any sound from the rear onboard audio jacks. oh no! not again... i once bought the same board locally with dead onboard audio that i had to return to the shop. this time i had the same problem. on the realtek hd sound effect manager program, on the audio i/o analog back panel jack detection tab, it couldnt sense any of the jacks being plugged in. none of the 6 rear jacks worked (no audio totally) and it couldnt sense anything being plugged in.
.
.
.
so the realtek hd audio sounded so much better than the cmi8738 audio indeed! so now this leads me to believe the audio chip isnt focked. just the rear panel jack detection switches are farked. i believe the jack detection is needed to turn off the audio signal to unused audio connectors to prevent signal reflection and echoing to preserve the high signal integrity and snr of hd audio, so then this adds on a new layer of problems. if the rear panel jack detection switches are farked, this then makes the onboard sound useless from just a minor fault.

Sounds like the high humidity in Singapore has tarnished the contacts in those jacks when they sat idle for too long.

Try using electronics contact cleaner first.

Spray a generous amount in each 3.5 mm jack on the rear and then connect and disconnect your headphones a good number of times (10-20??)

so the next day, i tried it. i always remember to unplug the power cord first before tinkering unlike uranium-235. because the computer was under the table, i had to shine a usb torchlight to make sure the connector was lined up in the header properly and gigabyte decided to put the header right up against the rear panel audio jacks grrr! so i fired up the system, made sure the right inputs and outputs were unmuted, made sure front panel jack detection was already disabled then retasked the green front panel jack to line out and the pink front panel jack to line in. then i put on some skyrim music and the dtv set top box audio output into the mic input front panel pink jack and then the moment of truth! i plugged the speaker audio jack into the front panel green jack and i had my ears blasted with some good skyrim music. hooray it works! getting my ears blasted with some skyrim hd audio was a good thing also to boot, so no worries about my hearing. hahaha! oh and the mic jack also works in receiving the dtv set top box audio. x2

Nice, good to hear you got your sound working.

surfing around the internet, it seems im not the only one with this problem. it seems dead onboard sound and lan is a common fault on these gigabyte boards. grrr!

If they are, I imagine this could be the fault with the Southbridge, since LAN and audio usually go through that. Does the SB run hot?

On that note, one of my AsRock 939Dual-SATA2 motherboards that I use for GPU testing lost its sound a very long time ago. Reading your thread here, though, I really ought to see if I can get sound out of it if I try a front pannel connector. So thanks for the suggestion!
The LAN is also iffy... as in, it works, but half the times it will crash the OS with heavy traffic. USB will often drop out too with high usage. And SATA will do the same. PCI slots also randomly corrupt data. So I believe my 939Dual-SATA2 is suffering from a dying SB.

the next day, i decided to bust out the syba cmi8738 sound card i bought years ago with my first pc back in 2002 cause i was an audiophool back then that believed the myth that onboard sound sucks. well, onboard sound has come a long way since 2002 when i first built my first pc!

Indeed it has.
For the most part, I've been more than satisfied enough with most onboard audio chips - even the SoundMax in this P4 Dell Optiplex 170L, which is generally considered far from good. But with analog (passive RC) tone controls and a headphone amplifier after, it hardly matters.

For me, the more important feature is for the audio hardware to support StereoMix recording. Most do, but some don't. I use it a lot to record sounds from online sources that may not necessarily allow the content to be extracted or downloaded.

the cmi8738 also sounded soft and i had to crank up the volume. after searching this forum, i saw a post by bigbeark who said that if a sound card doesnt sound good and the volume isnt as loud or its losing its loudness, it means the audio caps on the sound card are failing.

Not necessarily.
Some audio cards just sound "softer" than others... and by "softer", I assume you mean more quiet? If that's the case, then this could be (likely is) normal. I have an old Yamaha YMF724 -based PCI sound card from my PII PC from 1998, and that too sounds "soft"... as in, when I go between folders in WinXP for example, the little folder "flopping" sound gets a little clipped in the beginning. Otherwise, that audio card can get really loud, since it has onboard TEA2025 IC, meant to drive passive desktop speakers directly... and I have used it before like that (to drive a pair of 8-Ohm bookshelf speakers directly - it sounded quite nice, actually.)

I also have one of those cheap USB to 3.5 mm sound cards from eBay / China. Pretty nifty to keep around for problematic systems with no sound. However, these USBto3.5 mm adapters have incredibly -weak- sound. Can't drive even a very efficient pair of headphones properly. But better than no sound at all. And certainly this could be rectified by using a headphone amplifier of some sort after it.

yea, before putting the syba cmi8738 in, i saw it had 4 jackcon junk caps but two united chemicon caps. not sure which audio channel the ucc caps were on.

... if any.
Could be that those caps are used for power filtering and not audio signals. I've seen cards use ceramics almost entirely (except for the output series caps.)

considering that the creative cards have at least 20 caps on them, a full recap is much too tiring work esp. since the 5-6mm audio caps are soo much smaller that the 8-10mm caps on mobos, i decided i just dont have the dexterity to do a full recap of a creative sound card.

The caps are smaller, yes... but much easier to work on that a motherboard, because most sound cards don't have thick copper planes. This makes the caps really easy to desolder.

so now i wanna ask if there is a way to trick the sound chip into thinking something is plugged into the jack bypassing the jack detect switches by hotwiring it or something?

Yes.
Just find out which pins on the audio jack(s) are used for the signaling using a multimeter set on diode test and a pair of headphones. With the headphones plugged into the jack and multimeter black (-) probe connected to ground on the motherboard, start probing the pins on the underside of the board with the red (+) probe until you hear clicking/static from the headphones. Then note which pin corresponded to the Left channel and which to the Right channel. Do the same for the other jacks so you get a complete pinout of all of the L/R channels on all of the 3.5 mm jacks. The pins that are not connected to ground and you did not find to make any static/clicking noise through your headphones are likely the ones used for the signaling. I can't remember what needs to be done (can check if you're still interested), but I did trick a silly HP Maverick socket 775 board into thinking there is a front audio header and front USB connected to stop it from holding me up with a warning on POST every time I powered on the board. I believe I used tiny 100-470 Ohm resistors for the purpose and soldered them directly between the pins on these front headers. But like I said, I'll have to look it up if you want more details, as I did this like maybe 8 years ago.

is it also possible to desolder the rear panel audio io jacks and replace it with a new one?

Yes.

While I haven't done an audio jack replacement, I have removed one out of a dead ASUS board... and all I can tell you is that it takes a pretty good soldering iron to do - not because there are thick ground planes or anything (there aren't), but because the easiest way to do it is to melt a huge blob of solder onto all of the pins and then pull the entire jack out. Just a word of caution if you do this: melted solder can and likely will flow through the holes once you pull the jack. So if you have added lots of leaded solder, beware that it may run through the holes and drip right onto your hand. Likewise, if the board is sitting tilted on the table, the molten solder could also easily run off the side and land on your hand or body. I've nearly done this the first few times. But afterwards, I got smarter and wore a thick insulated garden glove... and try to keep the board leveled when melting large pools of solder on it.

So anyways... yes, it's possible to replace the connector.
However, I don't advise. Get some electronics contact cleaner instead and spray it generously inside all of the 3.5 mm jacks. This should clean up the corroded audio detection switches.

using momaka's hammerblow punch fix on the northbridge, i was able to get the board to post but as soon as i take my hand off, the board hangs instantly

You sure it's not the CPU socket BGA?
If that's an Intel board and if it's ever had Intel push-pin CPU heatsink, I'd bet my money on the CPU socket... unless the NB and SB chipsets tend to run hot - they those could indeed be the culprit too.

Speaking of my punch-fix method... I can no longer get my ECS MCP61PM-AM with its GF6100 chipset to post anymore. It did its random KB LED flashes at boot up the last few times, but still booted. Then a few days ago, no more. The punch-fix has been ineffective anymore. Looks like I will have to take the board out and reflow the nVidia chipset. And it's not like it runs hot either - I have a small fan on the chipset HS and keep it pretty cool (much better than when it was all passive from the factory.)

im also wondering if blowing too much air at the mobo can cause esd dmg to the board and fry the audio chip partially or the jack detect switches?

With the high humidity you have in SG, I seriously doubt it.

P.S. i know i ramble on too much like momaka too. hahahaha!

LOL!
(I do, don't I?)
Yeah, I get carried away with telling too many details. But doesn't that make an otherwise boring fix more fun to read?

[Uranium-235]

Re: what i went through dealing with my mobo with dead onboard sound!

I think I have that motherboard.

I actually have two creative x fi titanium fatal1ty's but one was in my storage sooo, the caps are dead. Other one is in my system

At one time I had an aureal card with an amp but it was so long I don't remember what it was

[ChaosLegionnaire]

so i took out one of my spare gigabyte ep35-ds3r mobos from storage for use to do some cpu, ram and video card testing of stuff i bought from ebay and got for free from momaka. he bought 50 e8400 cpus for cheap from ebay some years ago and i decided to help him relieve him of some of his supply since he had waaaay too many!! what did we say about hoarding too much stuff and depriving others of them?! *cough* socialism *cough* lol!

i had to blow some dust off the board and heatsinks with the datavac as i didnt clean it up before putting it in storage. after finishing assembly of all the components, i powered on the system and noticed there wasnt any sound from the rear onboard audio jacks. oh no! not again... i once bought the same board locally with dead onboard audio that i had to return to the shop. this time i had the same problem. on the realtek hd sound effect manager program, on the audio i/o analog back panel jack detection tab, it couldnt sense any of the jacks being plugged in. none of the 6 rear jacks worked (no audio totally) and it couldnt sense anything being plugged in.

surfing around the internet, it seems im not the only one with this problem. it seems dead onboard sound and lan is a common fault on these gigabyte boards. grrr!

the next day, i decided to bust out the syba cmi8738 sound card i bought years ago with my first pc back in 2002 cause i was an audiophool back then that believed the myth that onboard sound sucks. well, onboard sound has come a long way since 2002 when i first built my first pc! however, i found that the 2002 cmi8738 didnt sound as good as the 2008 realtek alc 889a. the cmi8738 can only do 16bit/48khz with the dac while the realtek alc 889a can do 24bit/192khz.

the cmi8738 also sounded soft and i had to crank up the volume. after searching this forum, i saw a post by bigbeark who said that if a sound card doesnt sound good and the volume isnt as loud or its losing its loudness, it means the audio caps on the sound card are failing. yea, before putting the syba cmi8738 in, i saw it had 4 jackcon junk caps but two united chemicon caps. not sure which audio channel the ucc caps were on.

i decided to search through the internet for a replacement sound card. i had my eye set on the creative x-fi and audigy 2 cards for use on winxp and win98se respectively but after reading lots of driver nightmare horror stories on vogons and other sites as well as the fact that creative used junk caps on their sound cards. jamicon seems to be creative's fav cap on their sound cards and i've read online reviews of some people saying their creative card died due to bad caps. considering that the creative cards have at least 20 caps on them, a full recap is much too tiring work esp. since the 5-6mm audio caps are soo much smaller that the 8-10mm caps on mobos, i decided i just dont have the dexterity to do a full recap of a creative sound card.

next, i tried searching for creative competitor cards but the competition like the auzentech prelude, aureal vortex 2, diamond mx300, asus xonar series etc. are either unobtanium or too expensive on junkbay for my tastes. but while searching the forum, i accidentally came across an unusual thread of someone trying to hotwire their mobo onboard front panel header. so that gave me the idea of trying to connect a casing front panel audio jack header to the mobo front panel audio jack header. i also had to rtfm because i had read to watch out for what type of header u are plugging it into because the ac'97 header pinout is different from the intel hda azalia header pinout even though the connector is the same. the gigabyte mobo manual says their intel hda azalia header is backwards compatible to ac'97 but u must disable front panel jack detection in the audio connector settings.

so the next day, i tried it. i always remember to unplug the power cord first before tinkering unlike uranium-235. because the computer was under the table, i had to shine a usb torchlight to make sure the connector was lined up in the header properly and gigabyte decided to put the header right up against the rear panel audio jacks grrr! so i fired up the system, made sure the right inputs and outputs were unmuted, made sure front panel jack detection was already disabled then retasked the green front panel jack to line out and the pink front panel jack to line in. then i put on some skyrim music and the dtv set top box audio output into the mic input front panel pink jack and then the moment of truth! i plugged the speaker audio jack into the front panel green jack and i had my ears blasted with some good skyrim music. hooray it works! getting my ears blasted with some skyrim hd audio was a good thing also to boot, so no worries about my hearing. hahaha! oh and the mic jack also works in receiving the dtv set top box audio. x2

so the realtek hd audio sounded so much better than the cmi8738 audio indeed! so now this leads me to believe the audio chip isnt focked. just the rear panel jack detection switches are farked. i believe the jack detection is needed to turn off the audio signal to unused audio connectors to prevent signal reflection and echoing to preserve the high signal integrity and snr of hd audio, so then this adds on a new layer of problems. if the rear panel jack detection switches are farked, this then makes the onboard sound useless from just a minor fault.

so now i wanna ask if there is a way to trick the sound chip into thinking something is plugged into the jack bypassing the jack detect switches by hotwiring it or something? is it also possible to desolder the rear panel audio io jacks and replace it with a new one? i have another scrap ep35-ds3r that has either bad bga on the northbridge or lga775 socket. using momaka's hammerblow punch fix on the northbridge, i was able to get the board to post but as soon as i take my hand off, the board hangs instantly, so that board is an ideal candidate to pilfer the 6 jack hd audio rear io panel from. any suggestions on a more permanent repair for the hd audio rear panel jack set would be welcome.

im also wondering if blowing too much air at the mobo can cause esd dmg to the board and fry the audio chip partially or the jack detect switches? i know planes and helicopters have a grounding cable or the landing gear is slightly conductive to ground the plane on landing because the air friction of flying through the air causes electrostatic charge to build up on the plane.

P.S. i know i ramble on too much like momaka too. hahahaha! thanks for reading my grandfather's story post.

i also have another mobo with problematic onboard sound. the rear green line out jack only has sound on one channel, the right channel. there is only muffled nearly inaudible sound coming from the left channel but i think that is for another post which i will type later on in this thread when i get down to it using the same fix mentioned above with the front panel header bypassing the rear jacks.