Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

[SOLVED] Diagnose Bad Component In PCI Audio Card

Status
Not open for further replies.

kraka

Junior Member level 3
Junior Member level 3
Joined
Oct 6, 2023
Messages
26
Helped
0
Reputation
0
Reaction score
1
Trophy points
3
Activity points
316
I have not electronics background but after two technicians refuse to repair my Audio Card, (maybe they did not have time, or they underestimate my card, althouth it is an excellent one), i decided to try, at least, fix it.
My Audio Card (PCI type) stopped having sound. After a closer inspection with a known technique (because i do not own a thermocamera) i saw that after injected 5 or more volts to pin no57 of side A of the Card, which has a label : AD(02) and a description : "I/O signal (May be driven by initiator or target, depending on operation)", and maybe means Address or Data bus ?, two components that are located very close to each other,
1)the ALTERA EPM 7064S1C44-10 BFB249919 EMU98 9946 and
2) : the P98AB VT244 , are getting hot very quickly.
They are getting hot even if the Card snap onto the motherboard, and Windows boots up.
P98AB VT244 is a component with 10 legs on the right and 10 on the left side. I do not know what is this, but i read somewhere that it is a Line Driver (what is this ?) or Amplifier. I search on web for a replacement but find nothing except this : 74LVT244 and SN54LVT244. I do not know at all, if i can replace P98AB VT244 with 74LVT244 or SN54LVT244.
The bad think is that i learned that the ALTERA is a programmable IC, and that i cannot just replace it with a new one. I must first program it with the "program" that is now inside the ALTERA on my Card, but i do not have this "program" (even if i had it, i do not know how to do it).
So my hope is the initial component that cause heat rising on the ALTERA to be the P98AB VT244. So if i replace it (the P98AB VT244) with a new one, both stop warming up and problem will be fixed.

But how to know which of those two is broken ? How to diagnose ?

I am attaching a picture of my card (Side B). I have drawn two red arrows, one bigger than the other. The big one points to the ALTERA EPM 7064S1C44-10 BFB249919 EMU98 9946 and the small one to the P98AB VT244.

Any other suggestion or idea, on how to try to make it have sound, are welcome.
Card is detected by Windows in "device manager", but cannot install any driver. I tried tones of drivers and all the OS's, from its native Win95-98 to Win7.
Before the problem occurred, it was installed in my WinXP PC with no problem. I am not sure 100% but maybe problem related with a faulty PSU that i have bougth at this time (have replaced later).
 

Attachments

  • SideB3.jpg
    SideB3.jpg
    1.2 MB · Views: 45

90 dB divided by 20 gives 4.5

4 --> results in "0000"
5 --> results in about "3"
--> "3 0000"
so in total I expect a ratio of 1:30,000

****
1000000000 ... has 9 zeroes
9 x 20dB ... results in 180dB

just over the thumb....

Klaus
Hi !
This is how i (magically) calculate this 1000000000 :

SNR = 10 * [log10 (Signal Power / Noise Power)]
9db = [log10 (Signal Power / Noise Power)]
(Signal Power / Noise Power) = 1000000000
 

But how to know which of those two is broken ? How to diagnose ?
74LVT244 costs pennies so there is no harm in buying one to try but my guess is the audio circuits will work fine without it fitted at all. As far as I can tell, all it does is interface the Altera IC to the black socket beside it, and if that's the case and that socket is for a joystick I'm sure you can live without it.

Brian.
 

74LVT244 costs pennies so there is no harm in buying one to try but my guess is the audio circuits will work fine without it fitted at all. As far as I can tell, all it does is interface the Altera IC to the black socket beside it, and if that's the case and that socket is for a joystick I'm sure you can live without it.

Brian.
First of all, I read your unfortunate experience with the professional video editing program. They happen to have given me, as gift, of a similar program, without them knowing that I already have one, and that is why now this program is useless to me. If you really want it, I can give it to you for nothing. I only, have to look for it because I don't remember exactly where it is (some years have passed since then) If I find it, I will gladly give it to you as a gift.

Secondly, I have already de-soldered the 74LVT244 from the Card (it is not exactly the 74LVT244, but the P98AB VT244 and since i can't find that, I believe (as i have been told) that the 74LVT244 can replace the P98AB VT244 without a problem). But, as you said, I can live without the joystick.
I've de-soldered the P98AB VT244 off the card, but the ALTERA keeps heating up right after Windows boot.

BUT i have made, perhaps, one important observation: One of the voltage regulators, the +5V (MC78M05), gives on its output, instead of 4.8 to 5.2 V (as described in its datasheet), 3.78 V when receives in its input, 12 V. This measurement was made after the MC78M05 had de-soldered fron the Card.(off board)
I was wondering if this failure cause a problem to the ALTERA. But , the ALTERA , although the MC78M05 is no present in the Card, still warming up.
 

This is how i (magically) calculate this 1000000000 :
The math is correct for power.
But we usually measure noise in V, V RMS.
You also use the "RMS". In technical terms "RMS" is a measure of V and I, but not for power.

I know in audio there is this terminology "Power RMS"
... but indeed it has nothing to do with RMS = RootMeanSquare.
... they use it as an equivaled of "undistorted sine"

Klaus
 

Right not the APS system, just the E-MU PCI card
Altera & 244 chips look familiar?

For a normalized audio of 1Vpk -92 dB RMS noise level for wideband spectrum if the noise was "white" means the selective narrowband SNR much higher ( by 20 log ratio of BW) but if it represents some sidetone or spurious tone then it could be noticeable between recordings.

But 70 SNR is good enuf for most recordings with compression.

I read that EPS drivers are possible with Vista installed then delete the config files and upgrade to Win7 x32 beyond that is silence.
YOU impressed me! Where did you find this card ? Because I have worked with it, to correct it, endless hours, I know it very well and I see that it is similar, if not exactly the same. What is this W5 IP1016 A ? I can't find it anywhere. Can you provide me with its datasheet? What worries me is that I can't find any mention of this card (the EMU-PC545). In this category of so-called "professional" EMU cards, there are references only to the following cards: APS, 0404, 1212, 1616, and 1820.
What is this EMU-PC545 ???
Is it a patent? Is the photo really from a card named EMU-PC545 or it is taken from an APS ?
The seller, also, does not sell the second card, the E-Drive, i.e. the interface placed on the front of the PC.
Of course, there are also these two possibilities:
1) de-soldering the PLD from the PC545 (is the W5 IP1016 A the PLD ???) and solder it to my APS.
2) using PC545 as E-Card and as an E-Drive the one that I already have.
But are they compatible ??
--- Updated ---

The math is correct for power.
But we usually measure noise in V, V RMS.
You also use the "RMS". In technical terms "RMS" is a measure of V and I, but not for power.

I know in audio there is this terminology "Power RMS"
... but indeed it has nothing to do with RMS = RootMeanSquare.
... they use it as an equivaled of "undistorted sine"

Klaus
Klaus,
Thank you for the clarification.
 

You would find it difficult to source an exact "P98AB VT244" because the P98AB part is a manufacturer and date (or batch) code, in this case almost certainly 1998. The generic part is the VT244, shortened from 74VT244.

The regulator may be faulty, although if it is the chances are the Altera itself was damaged then overloaded it. Consider that the regulator alone will be unstable if all you do is apply 12V to it, you also have to fit a capacitor from its input and output pins to ground and for some regulators also apply a light load before the correct voltage will be established. Without the capacitors the regulator may oscillate and on a regular meter will show less voltage than expected, exactly as you observed.

Worth noting that it is unusual to regulate the on-board 5V supply from the 12V pin on the PCI bus, normally the 5V and 3.3V lines come directly from the computer power supply via the PCI connector. This would account for the Altera still overheating, it almost certainly isn't fed from that regulator anyway.

Brian.
 

You would find it difficult to source an exact "P98AB VT244" because the P98AB part is a manufacturer and date (or batch) code, in this case almost certainly 1998. The generic part is the VT244, shortened from 74VT244.

The regulator may be faulty, although if it is the chances are the Altera itself was damaged then overloaded it. Consider that the regulator alone will be unstable if all you do is apply 12V to it, you also have to fit a capacitor from its input and output pins to ground and for some regulators also apply a light load before the correct voltage will be established. Without the capacitors the regulator may oscillate and on a regular meter will show less voltage than expected, exactly as you observed.

Worth noting that it is unusual to regulate the on-board 5V supply from the 12V pin on the PCI bus, normally the 5V and 3.3V lines come directly from the computer power supply via the PCI connector. This would account for the Altera still overheating, it almost certainly isn't fed from that regulator anyway.

Brian.
When i injected 12V into input of the +5V Regulator (alone), i also, have connected a 0.33μF Capacitor to the input and ground pins. I think, a second Capacitor, to be connected in the output and ground pins, is optional.

About fedding the ALTERA :
I do not know the terminology . Does the PCI bus is the bus from the components (like Regulators), and the PCI connectors are the Card's 58 golden pins?
Does what you are saying, means that if i will closely examine the 5V and 3.3V line's (that come directly from the Card's 58 pins), maybe i will find a faulty component, that is causing the ALTERA's overheating ?
 

I do not know the exact design of the card but generally, all 3.3V components are wired with their supply directly to the PCI 3.3V line and the same with all the 5V components to the 5V line. I guess, and this is based only on what I would consider if I was designing such a card, that the regulators mounted on the card are to give a 'clean' supply to sensitive audio stages. The main 3.3V and 5V lines are normally quite noisy from all the other devices drawing power from them so they wouldn't be suitable to use in stages that carry audio.

I think the chances are that your Altera IC itself is damaged, they can only withstand a tiny amount of overvoltage. The absolute maximum (beyond which damage is guaranteed to occur) it can withstand is 7V on any pin and its operating range is 4.75V to 5.25V. It only takes microseconds of over-voltage to kill them, it isn't death by overheating which would be gradual, it is destruction of the silicon insulating layers of the IC die itself which is instant and permanent.

You can check where the power feed to the IC comes from very easily, Pins 3, 15, 23 and 35 are the VCC pins, they are probably all linked together and connected to wherever the power comes from, probably the PCI connector. The dot on the top of the package marks pin 1 and they are numbered anti-clockwise when looking from above.

There is a possibility that the Altera device is not involved in audio at all but is the logic interface to the three black box headers. If all else fails, and you are only using the card for audio, it might work if the Altera device is removed completely. Obviously, the box headers would no longer work and there may be other consequences but if it comes to scrapping it, I would try removing it as a last resort.

Brian.
 

I do not know the exact design of the card but generally, all 3.3V components are wired with their supply directly to the PCI 3.3V line and the same with all the 5V components to the 5V line. I guess, and this is based only on what I would consider if I was designing such a card, that the regulators mounted on the card are to give a 'clean' supply to sensitive audio stages. The main 3.3V and 5V lines are normally quite noisy from all the other devices drawing power from them so they wouldn't be suitable to use in stages that carry audio.

I think the chances are that your Altera IC itself is damaged, they can only withstand a tiny amount of overvoltage. The absolute maximum (beyond which damage is guaranteed to occur) it can withstand is 7V on any pin and its operating range is 4.75V to 5.25V. It only takes microseconds of over-voltage to kill them, it isn't death by overheating which would be gradual, it is destruction of the silicon insulating layers of the IC die itself which is instant and permanent.

You can check where the power feed to the IC comes from very easily, Pins 3, 15, 23 and 35 are the VCC pins, they are probably all linked together and connected to wherever the power comes from, probably the PCI connector. The dot on the top of the package marks pin 1 and they are numbered anti-clockwise when looking from above.

There is a possibility that the Altera device is not involved in audio at all but is the logic interface to the three black box headers. If all else fails, and you are only using the card for audio, it might work if the Altera device is removed completely. Obviously, the box headers would no longer work and there may be other consequences but if it comes to scrapping it, I would try removing it as a last resort.

Brian.
This (that maybe i can, without having any big problem, remove the PLD), is very useful information and i would have done it, but unfortunately, the top black socket, is the one that is responsible for the connection between this Card (E-Card) and the front interface Card (E-Drive) which hosts the inputs jacks for Mics, Guitars etc. (the EMU APS consists of these two Cards)

For audio, i use a second PCI Card, mounted on the motherboard, an Audigy 4, which works when i boot on Windows 7. (i have built my PC with dual booting)

From the beginning of my PC's design i had decided to build it with two Cards, one (the EMU APS) for recordings on WinXP and one (the Audigy 4) for audio on Win7.
 

it isn't death by overheating which would be gradual, it is destruction of the silicon insulating layers
What is the different ? How this kind of destruction is happening ? If not overheating, then what, exactly, is it ?
This will not solve the problem that i am dealing in my first question-post, but It may prevent some future damage related to human error.
 

There are several ways to die!
Inside that IC there are 1,250 logic gates and all the memory needed to electronically link them together as well as the external pin driving and reading circuits. That means they are incredibly tiny. The silicon itself is probably about 6mm square, most of the package size is to make it big enough to handle in production and to connect all the pins around its edges. All those circuits (think of maybe half a million transistors) are formed of layers of silicon, silicon oxide and various doping chemicals in a 3D construction, the oxides are the insulation to keep signals apart. Being microns thick, even tiny voltages can puncture the insulation, think of it as arcing although in reality it more likely to be a tear in the crystal structure of the silicon. If the insulation punctures, an unwanted connection appears and stops it functioning properly. Crystal lattice destruction is almost instantaneous, the atoms are irreversibly torn out of position and can't be put back.

Another common reason for destruction is what we call 'latch up'. This happens when an excessive or wrong polarity voltage turns part of the internal circuit on and makes it clash with another internal part. For example, to make an output pin go to high or low voltage, it has two switches, one to ground (=VSS) and one to supply (=VDD or VCC). To make the pin go high, the switch to ground is turned off and then the switch to supply is turned on, to make the pin low, the switch to supply is turned off then the switch to ground is turned on. When high, some of the supply current can leave the pin to feed other external circuits, when it is low, current from outside is drawn to ground, We call it "sourcing" and "sinking" current. The usual latch up failure is caused when one or other switch, actually transistors, is turned on at the same time as its partner and current from the supply goes straight through them both to ground, normally killing both switch transistors in the process and creating a short circuit across the ICs supply pin and ground pin.

Sometimes, the damage is simply caused by overload, too much current being sourced or too much current being sunk because an external circuit is pulling or pushing too hard.

Brian.
 
What was the history prior to failure?
Before answering this question, we must first consider whether the failure is not just one, but more.
There is definitely a failure that is the initial one, but after that and in trying to fix this, there
may have been other failures, which are also responsible for the current state of the card. Therefore,
the cases are as follows:

1) There is only the initial failure.
Then the current state of the card is only due to this initial failure.

2) There is the initial failure and one or more failures that occurred in my attempt to fix the initial
failure.
Thus, the current status of the Card is due to the sum of all failures.

Initial failure.
The Card, when purchased, was working normally. with drivers that were customized for WinXP. Suddenly
lost sound. The only thing I remembered and can be related to this, is that before that happened, I
had bought and installed a PSU (Corsair 650W). From the first day of its operation I had noticed a
strange noise, very quiet, like meowing, repeated every 10-15 seconds. At first I didn't think that this
could cause any harm, so I didn't examine it more closely, even though it bothered me because I have very
sensitive hearing. But later I decided to take a closer look at what was happening and found that the PSU
fan was stopping and restarting every 10-15 seconds and that was causing this strange noise.
Then I uninstalled the PSU from my PC and took it to the store where I had bought it, for them to test it.
There they told me that they didn't notice any noise and no stop and start, and they returned the PSU to
me. After that, I bought a new PSU, one on 450W that works well (also had the right socket connection for my motherboard)
Here, I must say the following: Because the motherboard I use, does not have the socket it needs for the Corsair, i.e. its pins is fewer, I cut the socket of the Corsair connection cable, to make it smaller and match the socket of the motherboard. Some of PSU connector pins, were outside motherboard's socket but Corsair worked. It was supplying power to the PC without any problem (?) and that's why I left it on, at that moment. The only problem was this noise I described above.

Next, possible failures.
After the initial failure and after countless attempts to fix the problem by treating it as something
related to software and after I got to the point of eliminating any possibility that the problem was due
to software, I started thinking about the possibility that the problem was due to the Card itself (hardware problem). And I came to this conclusion after installing, not once but many times, all the Operating Systems I could find: from native to the Card, Win95-98 , to 2000, Vista, Me, XP (home and Professional), 7. I also
tried countless combinations of settings in BIOS, Registry, drivers, etc. After this, i took the Card to two
professionals who refused to take it on. So all I had to do was throw it in the trash. But because I
didn't want that, and I don't want it , i decided to try to fix it myself.
Here's what I've done so far:
The first thing I started from, was tantalum capacitors. I replaced all the tantalum capacitors with new
ones (TAJ225 16V 2.2μF), but that didn't solve the problem. Then I replaced the CS3310-KS Volume
Controller, and two differential output amplifiers (DRV134), also did not solve the problem.
Then i replaced the voltage regulators MC78M00 and MC900 (+8 -8 +5 -5), and also this did not solve the
problem. Also i replaced two resistances.
The next replacement i did, was a fatal one. I did not remember why i did it but i think that for some reason i assumed that there was, in that component, a short (to Ground or an inside one). This component is the black square socket on the top of side B. It is the connector to the E-Drive (which is the second part of the APS Card and hosts input jacks for Guitar,Mic etc). After i de-solder it and re-solder it, and when Windows booted, sparks and smoke burned a small area of this socket . There was a big damage where the middle level of the Card, the one that host the big copper area (probably ground) has exposed, and also two tiny copper buses has broken from its pins.
At that time i need to buy a microscope and a solder station. I could only bought the second and i built
the first based on one smaller i previously had.
With the help of these devices I repaired the damage, (but not all, as I found out later).
When I booted Windows spark and smoke did not come out anywhere in the socket, but after 4-5 seconds,
spark and smoke came out of one of the tantalum capacitors.
I tried to clean the area again, in order not to be any short, and indeed,there was an improvement because tantalum cap, stopped got burned. But... Something else start to getting very hot : One of the three (green) fuses right below the black socket.
I was ready to give up until I decided to take a closer look at the original area that had been burned (and restored(?), the one where the copper plate from the middle level, had been revealed. Although I had thoroughly cleaned this area and placed a special insulator, apparently somewhere there, was a contact between the exposed copper surface and a copper bus. And here is this area, because after I lifted this tiny copper channel and placed the insulation underneath, green fuse stopped getting hot :
AreaCauseShortCircuit.jpg




I forgot mention a repair I made to the pad and to two copper channels on one of the voltage regulators, because in my attempt to de-solder this regulator, I damaged the pad and cut both buses.

After all this and seeing that the Card was still recognized by Windows (in Device Manager), but it was not possible to install the drivers correctly, or even when it seemed that the drivers were installed correctly (since there was a message in the device manager that the card was installed and working correctly), in the device selection window in the playback and record of Control Panel the card was nowhere to be seen... After all these, I did all these volt injection measurements. Then I noticed that ALTERA and the close to it, VT244, get too hot when Windows boots. Maybe they were being got hot in earlier stages, but I did not notice it.

That is all.
Somebody can, now, tell :

"The really problem of the Card was... you"

OK but what could i do ?

Here are pictures of some of my "repairs" (?) :

E_CardConnectorAfterRepair.JPG


NewBurnedArea.jpg


VoltageRegAfterBigPadTransplant.jpg
 

Finally i managed to solve it.
Everything indicates that although maybe (and I say maybe, because I made too many attempts to repair this card for endless hours and I don't remember exactly all the steps I took and I just speculate, with more or less probability) I caused an overload in the PLD, this was not the fatal blow. The PLD must have been short-circuited before I even started all these repair attempts. And it was shorted by a faulty PSU. I just didn't look earlier for overheating. It was like stabbing an already dead person and thinking that you killed him, while he was dead even before you stabbed him...
It is repaired by "transplanting" from a "donor" card. Α forum member gave me this idea and he really helped. It was very difficult to find such a similar card. I found it in an old shop and although it was sold only for spare parts, gave me its PLD and replacing with it the short-circuited PLD of my own original card I made this card work again.
I want to say thanks to everybody who help me with their replies and especially [B]betwixt[/B].
Not for a second did he give me the impression that he has any arrogance, although it is obvious that he has incredible knowledge. That's how real scientists are. And not for a second did I feel insulted by his comments.
So long. This is the end of my ...little story.
Have a nice DAY.

 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top