Wow that is really bad! I'd replace every electrolytic capacitor even if they look physically fine, they could be high in ESR since they look like very cheaply made stuff.

Grab the Signal on Pin3 of LM386 and check if noise is present.
This is unamplified Signal.

The LM386 is a trashy amp.

Is the hum present if you try with headphones? If it is only present when connected to external mains powered amplifier or speakers, then it could be a ground loop (the ground connection on the 3.5mm plug could be bad).

The reason why there is more static during 5.25" drive initialization is that it uses 12V for the motors while 3.5" drives use only 5V for everything.

Also there seems to be two connectors on the card that could have some audio input/output so the hum could also be coupled to the system through floating inputs (apparent if it hums with headphones too).

P.S. I don't think replacing every capacitor would help in this case.

Jepael wrote:

Also there seems to be two connectors on the card that could have some audio input/output so the hum could also be coupled to the system through floating inputs (apparent if it hums with headphones too).
P.S. I don't think replacing every capacitor would help in this case.

Yeah, you can try using jumpers to ground those inputs (they should be LGGR or LGRG or something like that, so putting two jumpers on each connector should ground everything).

Blindly short circuiting them with jumpers may not be the most recommended thing. Figuring out what they are before doing so is.

Jepael wrote:

Blindly short circuiting them with jumpers may not be the most recommended thing. Figuring out what they are before doing so is.

Now that you say it, one of them could also be for output or something else... Unfortunately the photo of the back of the board is blurry, so I can't read the FCC ID.

Is that "SC802" thing a YM3014 clone? Might just be crappy quality.

Yeah it's a bit blurry, but if both connectors have pins 2 and 3 disconnected, and the bottom connector pin 4 is ground, my spidey sense tells me the top connector could be PC speaker input and bottom connector could be PC speaker output (with FM sounds).

Weird thing is if the SC802 is YM3014 clone (looks like it's the DAC in this system), it is not just a clone because it does not seem to need an external op-amp.

Thank you everyone for the suggestions. I'm not that good at soldering, but if I need to do it, this will be a good opportunity to learn ! 🤣

About the 4-pin connectors :
Looking at the traces, there seems to be a 4-pin input (likely for PC-Speaker or cd-rom) that goes through resistors, and a 4-pin output (like a line-out / pc-speaker out) that is connected directly to the output plug.

Using the PC Speaker output :
I tried plugging my PC Speaker to the output, and the adlib sounds play through the PC Speaker (only when there are no headphones/speakers connected to the card). Of note is the fact that the hum is still present and plays through the PC Speaker, though the speaker's bass response makes it more difficult to hear.

About using headphones :
The hum is present when using headphones, so we can exclude an external ground loop. Could there be something wrong with my computer/motherboard instead?

Floating inputs :
I tried grounding the PC Speaker in / CD IN pins, but the hum stays present.

Onboard chips :
I removed the label on the largest chip, revealing its true nature : it's more than a YM3812 (OPL2) clone. It has more pins. Attached is a picture of three examples... a true yamaha ym3812 (right), a reproduction of the ym3812 (center - KS8001) and our current mystery chip (left - SC801).

Grabbing the signal from the LM386 amp :
I understand what you want me to test, but I don't know where to make the second connection. To a ground I'm guessing?

Capacitors :
I have experience in changing capacitors in one of my computer monitors, and it's been 3 years and the little guy is still working. But if a capacitor is blown, to my limited knowledge, current stops going through it and thus, a blown capacitor would remove sound, not add it. I might be wrong though.

Attached is a better picture of the backside.
FCC ID is J5QMF-001 but I didn't manage to find pinouts or useful data apart from the company name : Princeton Technology Corp.

Curious facts :
I have two ISA slots on this motherboard.
Usually I only have one soundcard, leaving the other slot empty.

Since the interference in our current case is electrical, and not RF, I did a little experiment.
I chose a second soundcard to go in ISA slot#2.

Results :
Clone (slot#1) and 1990 AdLib (slot#2) electrical hum present.
Clone (slot#1) and Tolshine clone (slot#2) electrical hum present.
Clone (slot#1) and Sound Blaster CT1320 (slot#2) electrical hum almost disappears.
Clone (slot#1) and Sound Blaster CT1350 (slot#2) electrical hum almost disappears.

Any idea why that might be? Does this mean that my mobo/psu could be in the wrong?

Beegle wrote:

Capacitors :
I have experience in changing capacitors in one of my computer monitors, and it's been 3 years and the little guy is still working. But if a capacitor is blown, to my limited knowledge, current stops going through it and thus, a blown capacitor would remove sound, not add it. I might be wrong though.

Capacitors generally become a short-circuit when they fail, rather than open-circuit.

When electrolytic capacitors are getting old (say, before failing spectacularly to short circuit), they start to lose their capacitance (ability to hold charge, so they filter less ripple) so in that sense they slowly become more like open circuits.

After looking at the sound card pictures, I think I have a theory. The difference between the sound cards that hum and don't hum is how they use +12V. Humming cards use +12V directly to power the single supply LM386 output amp, while the non-humming Sound Blasters have a 78XX type voltage regulator (with at least 55dB ripple rejection) powering the output amp (with at least 40dB ripple rejection) so they do not directly use the +12V and can in total reject more than 95dB of power supply ripple. The +12V can also be used elsewhere on the Sound Blasters, but mainly it is used to generate analog +5V with simple zener regulator which is not so good at regulation so there will be some ripple but not much, and also if +12V is used to power operational amplifiers, it is not so bad because they generally are not so sensitive to power supply ripple (ripple rejection is at least 77dB).

So try a different power supply, and if it does not help, then try on another motherboard. I suspect the power supply more, as the hum seems to be at mains supply frequency, so it is possible there is something wrong in the power supply that lets the 12V voltage fluctuate according to mains AC voltage waveform.

The cards should be fine, but the humming cards are just very sensitive to power supply ripple because the LM386 is a single supply amplifier powered directly from +12V and it needs quiet power supply as it has typically only 55dB ripple rejection and they might have made the design more sensitive to noise by lowering the bypass capacitor value on LM386 pin 7, or the cap has started to dry out and lose its capacitance but when there is same issue on both cards this is unlikely.

(To put things in approximate decibel perspective, 8-bit audio range is 48dB, 9-bit would be 55dB, 12-bit would be 72dB, and OPL chip output range is 16-bit so it has 96dB range even though its output resolution is only 10-bit).

Jepael wrote:

The difference between the sound cards that hum and don't hum is how they use +12V. Humming cards use +12V directly to power the single supply LM386 output amp, while the non-humming Sound Blasters have a 78XX type voltage regulator.

I understand that.

But I wasn't stating that Sound Blasters don't hum. I was stating that the problematic card stops humming, when a Sound Blaster card is present in another ISA slot. And THIS, I find really surprising.

In the light of this, I'm switching my suspicious from the sound card, to either the motherboard or powersupply, although, the powersupply is very recent (2014) and I'm using an ATX to AT converter to allow for -5v to be present for certain sound cards. That leaves the potential culprit to be the motherboard, and yes, it is from the '90s and would probably benefit from a good re-cap.

SquallStrife wrote:

Capacitors generally become a short-circuit when they fail, rather than open-circuit.

I didn't know that, and yes, that would confirm the current theory that the motherboard is not filtering the PSU's input enough and feeds it directly to the sound cards. Some sound cards deal with the problem with a regulator, others more cheaply made, are not able to.

I will try to find another motherboard with ISA slots, and come back with results.

Maybe the power filtering capacitors on the second card are helping? Or, maybe, the slightly higher load on the ISA bus is drowning the noise? I'd try other ISA cards as well (Network cards, video cards, etc.) just to see where it takes you.

Yes Audio Signal on Pin 3 of LM386, Second Connection to Ground.
Check if there is Hum. If yes, bad.
If no Hum is present, the LM386 needs better decoupling.
The LM386 is known for weird Sound.. its better used for Doorbell 😀

matze79 wrote:

The LM386 is known for weird Sound.. its better used for Doorbell 😀

Or cheap DIY guitar amps 😊 - no, really, the LM386 has quite a distinct sound when overdriven. I wouldn't use it for anything where audio quality is even remotely important, though.

matze79 wrote:

If no Hum is present, the LM386 needs better decoupling.

Or better filtering! In that case, solder a capacitor (220µF+) between Pins 6 (V+) and 4 (Ground).

Culprit finally found : against my assumptions it was the power supply...
Strange how the things I took for granted (brand new PSU, modular, modern, certified 80 bronze, good ATX->AT converter with dummy load) clouded my opinion.

I tried the same sound card, on the same motherboard, in the same system, with the same converter. But instead used the old PSU from my Pentium 4.
Boom. Interference reduced by 95%.

Since the card uses a cheap amplifier, and has little to no filtering, I can still hear a tiny bit of humming and 'computer thinking' if I crank the volume all the way up.
But under normal use at acceptable volumes, while listening to music I hear nothing at all.

For reference, here was my overkill PSU (for an AT Pentium 2 build) :
http://www.newegg.ca/Product/Product.aspx?Ite … N82E16817341050

This said, I still plan on doing the Pin 3 to Ground test, and the capacitor addition between Pins 6 and 4 (ideally with a jumper to compare easily).

Thank you to all who contributed in this thread 😀

Hello, I have the exact same "FCC ID: J5QMF-001" card and recently posted photos on how some enterprising soul had modded it DIY-style with an amplifier stage. See:
Tandy Nano - 3 Voice Soundcard ISA - Lowprofile