Cyrix200+ wrote on 2020-06-20, 15:26:

You could try and swap out the TEA2025 if you have a spare one. Mine died and it gave weird results.

I do not have a spare TEA2025 but they are not difficult to source..

Lets see if a new LM324 and bodge resistor fiz things (though now I realize it is not actually on the digital audio lane..)

appiah4 wrote on 2020-06-20, 15:52:

Cyrix200+ wrote on 2020-06-20, 15:26:

You could try and swap out the TEA2025 if you have a spare one. Mine died and it gave weird results.

I do not have a spare TEA2025 but they are not difficult to source..

Lets see if a new LM324 and bodge resistor fiz things (though now I realize it is not actually on the digital audio lane..)

The bodge resistor is only for the mic input comparator, not for digital playback. But in general, you want to start at the DAC and work your way forward to the amplifier:
* Does the DAC work? Check R38, U12, C59, C53, and make sure you have -12V on pin 3 and 5V on pin 13. You won't see voltage on pin 4 since it is a current output.
* Check for audio at pin 8 of U5. Is there -5V at pin 11? Check R18 too.
* Is the mute circuit stuck on? Check T1 to make sure it is off. Check R10.
* Does the antialiasing filter work? Check U5 pin 1 to see if there's a signal there. Check C51, C35, R11, and R12. Wrong values will set up the filter incorrectly.
* Check the mixer. Is R25, R22, R24, and R23 the correct value? Also check R28 and R31 just in case the FM chip is too loud.

It helps to have an oscilloscope but you may be able to do with an audio amplifier or even just a pair of powered speakers.

TubeTimeUS wrote on 2020-06-20, 17:14:

The bodge resistor is only for the mic input comparator, not for digital playback. But in general, you want to start at the DAC […]
Show full quote

The bodge resistor is only for the mic input comparator, not for digital playback. But in general, you want to start at the DAC and work your way forward to the amplifier:
* Does the DAC work? Check R38, U12, C59, C53, and make sure you have -12V on pin 3 and 5V on pin 13. You won't see voltage on pin 4 since it is a current output.
* Check for audio at pin 8 of U5. Is there -5V at pin 11? Check R18 too.
* Is the mute circuit stuck on? Check T1 to make sure it is off. Check R10.
* Does the antialiasing filter work? Check U5 pin 1 to see if there's a signal there. Check C51, C35, R11, and R12. Wrong values will set up the filter incorrectly.
* Check the mixer. Is R25, R22, R24, and R23 the correct value? Also check R28 and R31 just in case the FM chip is too loud.

It helps to have an oscilloscope but you may be able to do with an audio amplifier or even just a pair of powered speakers.

Thank you kindly for responding.

* The DAC appears to work. I can launch DOOM and hear the digital audio if I max the sound on the potentiometer and boost it quite a bit using an external amp. The sound is there, it's just very low. I can confirm R38, C59 and C53 are the correct rating. I will check for -12V on Pin 3 and 5V on Pin13 on U12 tomorrow.
* R18 is the correct rating. What is the easiest way to check for audio at Pin 8 of U5? Unfortunately I do not have an oscilloscope but I can check to see if there is some kind of voltage there?
* R10 is the correct rating. I suppose T1 could be faulty, but I am not sure how to do the mute circuit check. I'm guessing if C and E remain shorted while there is digital audio, it is actually stuck on and muting when it should not?
* C51, C35, R11 and R12 are the correct rating. Unfortunately I do not have an oscilloscope but I can check to see if there is some kind of voltage there?

appiah4 wrote on 2020-06-20, 17:46:

Thank you kindly for responding. […]
Show full quote

TubeTimeUS wrote on 2020-06-20, 17:14:

The bodge resistor is only for the mic input comparator, not for digital playback. But in general, you want to start at the DAC […]
Show full quote

The bodge resistor is only for the mic input comparator, not for digital playback. But in general, you want to start at the DAC and work your way forward to the amplifier:
* Does the DAC work? Check R38, U12, C59, C53, and make sure you have -12V on pin 3 and 5V on pin 13. You won't see voltage on pin 4 since it is a current output.
* Check for audio at pin 8 of U5. Is there -5V at pin 11? Check R18 too.
* Is the mute circuit stuck on? Check T1 to make sure it is off. Check R10.
* Does the antialiasing filter work? Check U5 pin 1 to see if there's a signal there. Check C51, C35, R11, and R12. Wrong values will set up the filter incorrectly.
* Check the mixer. Is R25, R22, R24, and R23 the correct value? Also check R28 and R31 just in case the FM chip is too loud.

It helps to have an oscilloscope but you may be able to do with an audio amplifier or even just a pair of powered speakers.

Thank you kindly for responding.

* The DAC appears to work. I can launch DOOM and hear the digital audio if I max the sound on the potentiometer and boost it quite a bit using an external amp. The sound is there, it's just very low. I can confirm R38, C59 and C53 are the correct rating. I will check for -12V on Pin 3 and 5V on Pin13 on U12 tomorrow.
* R18 is the correct rating. What is the easiest way to check for audio at Pin 8 of U5? Unfortunately I do not have an oscilloscope but I can check to see if there is some kind of voltage there?
* R10 is the correct rating. I suppose T1 could be faulty, but I am not sure how to do the mute circuit check. I'm guessing if C and E remain shorted while there is digital audio, it is actually stuck on and muting when it should not?
* C51, C35, R11 and R12 are the correct rating. Unfortunately I do not have an oscilloscope but I can check to see if there is some kind of voltage there?

Since you don't have an oscilloscope your best bet is probably a set of powered speakers. A somewhat better calibrated alternative is to play a sine wave (1000Hz) test WAV file in a loop and measure the RMS AC voltage with your multimeter.

TubeTimeUS wrote on 2020-06-20, 20:28:

Since you don't have an oscilloscope your best bet is probably a set of powered speakers. A somewhat better calibrated alternative is to play a sine wave (1000Hz) test WAV file in a loop and measure the RMS AC voltage with your multimeter.

Thank you I will try this as well.

I have some troubleshooting steps in mind and would appreciate help from anyone who can 😀

1. My chief suspect is the mixer circuit. The signal from the DAC is there, I can here it, it's just at the wrong volume. Somehow, the mixing levels are wrong. I look at the Mixer cicruit:

.. and the only way for this appears to be is that R22/25 from digital audio or R28/31 from OPL2 are wrong. I checked all four, and all four appear to be correct.

What isn't exactly correct is the Op-Amp I used, though. I used an ONS MC3403P instead of a JCR NJM3403AD. They seem to be very similar but not identical. However, there are genuine Sound Blaster cards with Motorola MC3403P out there, and the Motorola MC3403P is 100% the same as the ONS MC3403P I used.

Well.. How about CMS? I haven't tried this yet. So here is my question: What is a good game to try CMS sound on the Snark Barker? If CMS levels are on par with FM then I know the issue is not FM being too loud.. Which would be a weird twist.

2. If the issue is not the Mixer, then it is almost certainly before that, as I can't see how the post-mixer Amplifier (U4) could be causing this - they simply take the mixed audio signal and amplifies it, it can't just mess with a single audio source on the card. So, when I look at the Digital Audio Output circuit:

I note that one of the following may be faulty:

DAC (U12)
Op-Amp (U5)
Mute_En Transistor (T1)
Op-Amp (U11)

I will order spares for each and try them out, being more careful about the bodge resistor this time. I will also replace the Amplifier (U4) just in case.. (I guess it pays to socket everything? 🤣)

3. I will start checking voltages with a multimeter after replacing all these ICs and seeing if anything changes. In the meantime, could anyone tell me how T1 should normally behave? How can I determine if it is stuck at mute? I'm thinking if there is a Mute_En signal on 2 (Base) then the 1 and 2 (Emmitter and Collector) are short, meaning the audio signal is sent to Ground (muted)? So if 1/2 are always short, that means T1 is faulty?

4. I will start going over the board and checking the following voltages:

x -12V on Pin 3 and 5V on Pin 13 of DAC (U12)
x -5V at Pin 11 of Op-Amp (U5)

5. I will try to check for signal/audio on:

x Pin 1 of Op-Amp (U5)
x Pin 8 of Op-Amp (U5)

I will report as I go through this list. In the meantime, open to all comments and help 😁

appiah4 wrote on 2020-06-22, 06:19:

<snip>

Well.. How about CMS? I haven't tried this yet. So here is my question: What is a good game to try CMS sound on the Snark Barker? If CMS levels are on par with FM then I know the issue is not FM being too loud.. Which would be a weird twist.

<snip>

Paku Paku is easy and free: https://deathshadow.com/pakuPaku

Another option is Monkey Island (1).

Prince of Persia has CMS support but that is speed sensitive (I think it only sounds correctly on a 4.77MHz XT).

All the early Sierra SCI titles have CMS support (Colonel's Bequest, KQ4, SQ3, PG2, ...)

appiah4 wrote on 2020-06-22, 06:19:

3. I will start checking voltages with a multimeter after replacing all these ICs and seeing if anything changes. In the meanti […]
Show full quote

3. I will start checking voltages with a multimeter after replacing all these ICs and seeing if anything changes. In the meantime, could anyone tell me how T1 should normally behave? How can I determine if it is stuck at mute? I'm thinking if there is a Mute_En signal on 2 (Base) then the 1 and 2 (Emmitter and Collector) are short, meaning the audio signal is sent to Ground (muted)? So if 1/2 are always short, that means T1 is faulty?

4. I will start going over the board and checking the following voltages:

x -12V on Pin 3 and 5V on Pin 13 of DAC (U12)
x -5V at Pin 11 of Op-Amp (U5)

5. I will try to check for signal/audio on:

x Pin 1 of Op-Amp (U5)
x Pin 8 of Op-Amp (U5)

I will report as I go through this list. In the meantime, open to all comments and help 😁

I'd start with these measurements instead of replacing lots of parts. If you think T1 may be a problem, just temporarily remove it from the board. If you pull it out and the sound is now at the correct volume, you have found the culprit.

After a couple of weeks I had time to look into the issue with this card.. I am now trying it in mmy 430TX Pentium 200-MMX system instead of a 386/486, which makes boot cycling etc. much easier.

Against better judgement from others I actually went with this first. It appeared to be the low hanging fruit since I could get the ICs for real cheap. Well, I replaced DAC (U12), Op-Amp (U5) and Op-Amp (U11) to no success. I have yet to try removing Mute_En Transistor (T1) - I will look into this tonight.

I also tried the CMS output and the volume is roughly the same as the OPL2, if just a tad bit higher.

Now that I am listening to the output with earphones I also noticed that the OPL2 and CMS music has some intermittent crackling, particularly on the left channel.. It gets stranger by the day. 🤣

I will try the other troubleshooting options tonight..

Quick question, since this is a DSP 2.01 card, SET BLASTER should be A220 I7 D1 T3 right?

If you're going that route, also try U4 if you have one (EDIT: a spare). My broken TEA2025 sounded like this: Re: Tube Time: Announcing The Snark Barker, a 100% Compatible SB 1.0 Replica!

appiah4 wrote on 2020-06-29, 11:17:

After a couple of weeks I had time to look into the issue with this card.. I am now trying it in mmy 430TX Pentium 200-MMX syst […]
Show full quote

After a couple of weeks I had time to look into the issue with this card.. I am now trying it in mmy 430TX Pentium 200-MMX system instead of a 386/486, which makes boot cycling etc. much easier.

Against better judgement from others I actually went with this first. It appeared to be the low hanging fruit since I could get the ICs for real cheap. Well, I replaced DAC (U12), Op-Amp (U5) and Op-Amp (U11) to no success. I have yet to try removing Mute_En Transistor (T1) - I will look into this tonight.

I also tried the CMS output and the volume is roughly the same as the OPL2, if just a tad bit higher.

Now that I am listening to the output with earphones I also noticed that the OPL2 and CMS music has some intermittent crackling, particularly on the left channel.. It gets stranger by the day. 🤣

I will try the other troubleshooting options tonight..

Quick question, since this is a DSP 2.01 card, SET BLASTER should be A220 I7 D1 T3 right?

Definitely try a different TEA2025! I had a bunch of broken ones. The TEA is supposed to be soldered without socket, for thermal reasons, but since they fail quite often I socketed mine. You can put a tiny heatsink on it, if you like...

In case anyone cares:

The TEA2025 amplifier chips by UTC (Unisonic Technologies Co. Ltd. Taiwan) have a datasheet with a copyright date of 2014 and I don't see any EOL notice in there.
So maybe they're still making them.

I will try replacing the Amplifier (U4) tonight as well, I have one spare. In the meantime, I took voltage measurements:

Pin 3 of DAC (U12): -11.7V (-12V)
Pin 13 of DAC (U12): 4.9V (5V )
Pin 11 of Op-Amp (U5): -5.1 (5V)
Pin 3 of Voltage Regulator (P2): +9.0V (9V)

They all seem to be in order.

Now, time to check for signal/audio on:

Pin 1 of Op-Amp (U5)
Pin 8 of Op-Amp (U5)

I wonder if I could just try using a PC mainboard buzzer (not a real PC speaker) with ground on one and and these pins on the other to see if there is audio.. Couldn't really hurt, could it?

I couldn't wait until tonight, so I tried two things:

1. I swapped the TEA2025 out for the spare (socketing everything saves a lot of time!)
2. I first remoed, then replaced the T1 transistor

and the behaviour is 100% the same.

1. SB part does not work. There IS audio, but the volume is VERY, VERY low, and it randomly gets stuck or stops working.
2. There is a strange crackling/noise (kind of reminiscent of the SB16 DMA clicking) when playing OPL2 or CMS music that is a lot more pronounced on the left channel (but is also there on the right channel).
3. There is a strange/weird high pitched static on the left channel during idle that is absent on the right channel.

I will record some audio of these tonight.

Regarding the crackling/noise; The fact that it sounds different in the left/right channels got me thinking. AFAIK the only stereo sound source on this card is the CMS, so I went ahead and moved the SAA1099P chips around but the more noticeable noise remained on the left channel. Removing the SAA1099P chips completely did not eliminate the noise. So the SAA1099P chips are not the source. I'm guessing this issue must also have something to do with either the mixing or the amplification circuit. Maybe the low SB noise and left channel issues could be related? That's strange because I already replaced the relevant IC in both circuits (U11 and U4). I suppose the potentiometer could also be causing the noise/crackling?

Also, more troubleshooting information:

I run into a strange issue with Ultimate Doom, picking Sound Blaster as the music option results in a quit to DOS upon starting a new game with the following error:

"Z_CheckHeap: block size does not touch the next block"

Adlib appears to work fine. This IIRC is a map-related issue and has absolutely nothing to do with sound. Which begs the question: Why? My only explanation, the block size check is done in memory and the Snark Barker is doing something in memory that messes with it. Maybe something's wrong with the DMA circuitry?

appiah4 wrote on 2020-06-29, 12:46:

I couldn't wait until tonight, so I tried two things: […]
Show full quote

I couldn't wait until tonight, so I tried two things:

1. I swapped the TEA2025 out for the spare (socketing everything saves a lot of time!)
2. I first remoed, then replaced the T1 transistor

and the behaviour is 100% the same.

1. SB part does not work. There IS audio, but the volume is VERY, VERY low, and it randomly gets stuck or stops working.
2. There is a strange crackling/noise (kind of reminiscent of the SB16 DMA clicking) when playing OPL2 or CMS music that is a lot more pronounced on the left channel (but is also there on the right channel).
3. There is a strange/weird high pitched static on the left channel during idle that is absent on the right channel.

I will record some audio of these tonight.

Regarding the crackling/noise; The fact that it sounds different in the left/right channels got me thinking. AFAIK the only stereo sound source on this card is the CMS, so I went ahead and moved the SAA1099P chips around but the more noticeable noise remained on the left channel. Removing the SAA1099P chips completely did not eliminate the noise. So the SAA1099P chips are not the source. I'm guessing this issue must also have something to do with either the mixing or the amplification circuit. Maybe the low SB noise and left channel issues could be related? That's strange because I already replaced the relevant IC in both circuits (U11 and U4). I suppose the potentiometer could also be causing the noise/crackling?

Also, more troubleshooting information:

I run into a strange issue with Ultimate Doom, picking Sound Blaster as the music option results in a quit to DOS upon starting a new game with the following error:

"Z_CheckHeap: block size does not touch the next block"

Adlib appears to work fine. This IIRC is a map-related issue and has absolutely nothing to do with sound. Which begs the question: Why? My only explanation, the block size check is done in memory and the Snark Barker is doing something in memory that messes with it. Maybe something's wrong with the DMA circuitry?

Sounds like something's wrong with the mixer stage. You also ought to look for cold solder joints on the board, short circuits (solder bridges), and possibly even a board with etching issues (which could cause open and short circuits).

That Z_CheckHeap error sounds like heap corruption. If you take out the sound card and run it without, I assume it works fine? So yeah, it sounds like the card is trying to put data on the ISA bus at the wrong time. This would also explain why it gets stuck or stops working while playing back sound. Take a look at the DMA circuit: U28, U20, U21, U30. I'd also check C65, C66, and R43 to make sure they are OK and have the correct capacitance/resistance.

TubeTimeUS wrote on 2020-06-30, 00:03:

Sounds like something's wrong with the mixer stage. You also ought to look for cold solder joints on the board, short circuits (solder bridges), and possibly even a board with etching issues (which could cause open and short circuits).

That Z_CheckHeap error sounds like heap corruption. If you take out the sound card and run it without, I assume it works fine? So yeah, it sounds like the card is trying to put data on the ISA bus at the wrong time. This would also explain why it gets stuck or stops working while playing back sound. Take a look at the DMA circuit: U28, U20, U21, U30. I'd also check C65, C66, and R43 to make sure they are OK and have the correct capacitance/resistance.

Thank you for your support, I really appreciate it! I am continuously looking over the board, so far I can see no cold solder joints, bridges or broken etchings. I noticed a few solder joints that looked weak, and although they had proper contact I redid them, to no effect. I will keep looking.

In the meantime:

1. Z_CheckHeap error does not happen without the card installed even if I select Sound Blaster music at Address 220 - so the problem definitely seems to start when the card starts transferring data to/from Address 220. This is Even the case when I set the game up for music only. I do not believe OPL2 music from Address 220 should require the use of the DMA circuitry. Maybe there is an issue at the ISA Bus or Address decoding stages? Rarely the game loads the level up but it crashes immediately with graphics corruption on the screen and the music playback stops.
2. I have goneover the DMA Request Latch and Control circuitry, and verified every single trace, as well as every single IC and their socket solder joints. Everything is connected properly (all the way back to/from the Bus/DSP signal inputs and outputs). U28, U20, U21, U30 are the correct ICs and every single pin on these appear to make proper contact with their sockets (reseated each and tested with a multimeter). I do not have the means to test these ICs, but I can order replacements to see if that fixes anything – they are added to my next order shopping list (by the time I get this working I will have enough components for a second card 😁 ). R43 and C66 are the correct rating, however (C65 and C68) are not. The BOM states 270pF 50V for these, but I have 270pF 1KV installed at each. Could C65 and C68 being rated for too high a voltage be what causes the issue?

My next course of action will be to reseat the Bus logic chips and the DSP and check their sockets to make sure they are all good. I will then check all traces on the mixer and amplifier stages. I am otherwise stumped for ideas, regarding the SB digital audio not working issue.

As for the crackling/noise, it’s almost as if the left channel is picking up more significant interference when there is an audio signal. The left channel circuitry is located close to the microphone on the card, so I wonder if it could have something to do with that. I will look into this. (EDIT: I redid all solder joşnts on the left channel starting at the mixer to no difference. There is *some* crackling/noise on the right as well so.. I dont know this soundd a lot like an amplifer error, but how can two different amplifier ics from two different suppliers fail similarly? Maybe it is a power issue at the amplifier?)

Great news (and some not so great news) below:

I checked for signal/audio on Pins 1 and 8 of the U5 Op-Amp. My method was crude: I hooked up this buzzer:

to the relevant pins and the card's GRNDA using these:

Certainly not recommended to anyone, but for my case it worked. I launched Doom II (Doom Ultimate still crashes, but this works.. Maybe a speed sensitivity issue with Doom Ultimate using this card?) There was a very, very distant but slightly audible sound on Pin 1 when shooting the pistol. Then I tried Pin 8, and I got VERY audible gunshot sounds from Pin 8, which was curious because the only difference between the two was that one is pre-antialias and loud, the other is post-antialias and almost inaudible.. I thought to myself: "This must be an antialias filter issue!" and reseated U5, then tried the card with headphones. Now the SB sound actually works, it's very LOUD and audible, but also terribly distorted and metallic. I'm thinking, somehow the antialias filter is no longer working and that is solving the digital audio volume issue, but results in a terrible sound.

I will now look more closely at the AntiAliasing circuit and hopefully fix this once and for all.

There is still the whole crackling/noise issue (more prevalent on the left channel) but I will take this as a small win for now 😀

Benedikt wrote on 2020-06-29, 12:20:

In case anyone cares:

The TEA2025 amplifier chips by UTC (Unisonic Technologies Co. Ltd. Taiwan) have a datasheet with a copyright date of 2014 and I don't see any EOL notice in there.
So maybe they're still making them.

Hm, indeed. However none of the big stores have them stocked. UTSource has the UTC ICs in their portfolio, but that doesn't mean they are still being produced... Would be interesting if UTC still makes them. The original TEA2025 was by ST I think.

Can anyone please check these recordings and tell me if this sounds right for the Snark Barker's OPL3 and Digital Audio?

https://drive.google.com/drive/folders/1AWtnF … vqX?usp=sharing

You can clearly hear the clicking/crackling on the left channel (some on the right as well) and the distorted/metallic digital audio (which is a bit too loud? Maybe this is clipping?)

The PCM sounds terrible. Broken DAC? The clicking on the left channel somehow sounds a bit like an echo of some OPL sounds.

I would say the TEA2025 is fine. The OPL sound is almost clear, apart from the clicking. So I would say that one is working.