PC Speaker with Soundcards that Don't Have Headers

Look here.

PC Speaker with Soundcards that Don't Have Headers

Isn't the PC speaker rather a TTL device (+5v DC on/off) than an audio source ?
If so, wouldn't be an optocoupler way more elegant ? 😕

Jo22 wrote:

PC Speaker with Soundcards that Don't Have Headers

Isn't the PC speaker rather a TTL device (+5v DC on/off) than an audio source ?
If so, wouldn't be an optocoupler way more elegant ? 😕

Not only more elegant.
I am really afraid when thinking about connecting the chipset directly to an outside cable - a little ESD spark can kill it if I connect it as suggested in the thread mentioned by @Matze79.
So I consider an optocoupler a necessity. At least for people like me who have one of the more rare chipsets that are extremely difficult to replace.

I guess I'll try to find out how these sound cards with speaker input convert/filter that signal into clipping-free analog later tonight.

Hmm, good point. Your words made me think and so I looked up some old amateur radio schematics.
That's when I remembered the drawing for the Quick Brown Fox (QBF) interface.
It's meant to interface the Commodore C64's very sensitive cassette interface to a radio transceiver.

I guess this scenario is at least comparable to the PC speaker, which in essence is a single channel of the PC's System Timer that is exposed to a header.
So maybe it would be possible to use an optocoupler that consists of an LED/photodiode (instead of an LED/phototransistor).
That way no external power would be necessary. The diode would get powered by the light emmissions of the LED.
It's not much, but the Line-In on the soundcard might be able to amplify the signal to an usable level.

Edit: I didn't mean to to criticize the other schematic mentioned above, of course.
As far as I understand, one R is a pullup, the C is for smoothing, and the other two Rs act as a divider and attenuator.
Also, speaker's GND is missing, since there's a common ground assumed (the target soundcard is installed inside the same PC).
Or do I miss something ? The odd cap alignment could also be some kind of filter (raising two specific frequencies ?)
Anyway, if there was another cap on the output (1 to 10uf ?), some rudimentary isolation would be provided, I believe.
Edit: As a side effect, such a cap may also work as a low-pass, depending on the value it has.

Edit: Something like that was also mentioned here: http://www.deinmeister.de/e_sbpcqlnk.htm
"One end of the resistor is also to be sold with the capacitor. The main reason for using the capacitor is
to prevent hardware damage as a result of using the plug the wrong way."

Edit: Speaking of these schematics, what about fast diodes (germanium, schottky etc) in addition to the coupling capacitor ?
Such a diode could act as a protection diode if placed in the line past that "pullup" resistor.
It would block incoming noise or backflowing power from the soundcard etc.
Since the speaker signal is DC anyway, no loss in quality should occur.
(I'm looking at the PC speaker as an output of a single data line.
Like "TXD" on a serial port or "Data" on Centronics ports.)

Jo22 wrote:

So maybe it would be possible to use an optocoupler that consists of an LED/photodiode (instead of an LED/phototransistor).
That way no external power would be necessary. The diode would get powered by the light emmissions of the LED.
It's not much, but the Line-In on the soundcard might be able to amplify the signal to an usable level.

Damn! An extremely elegant idea! 😀
This could even make a galvanic connection unnecessary (no ground loop).
Not to have to supply a receiver with power would be great too...

Jo22 wrote:

Edit: I didn't mean to to criticize the other schematic mentioned above, of course.

Me neither... the thread @Matze79 mentioned is inspiring. And his schematic would be fine for me, too, if I had a socketed chip on the mobo that is easy to replace.

I really like the idea of using photovoltaic optocouplers...
They can indeed drive a 100kohms load like the typical Line In inputs without the need for external power.

Their only disadvantage is their speed.
The fastest I found so far is the Infineon PVI 1050.
Still need to find one that can make more than maybe 3, 4 kilohertz.

Will get a sample and then play around with a signal generator and oscilloscope 😀

Good luck! 😀

I'll trying do some experiments over weekend, too.

There are plenty of optrons that are used in MIDI gear and there you have 31250baud speed, plenty enough for sound. I don't remember any part number on top of my head though.

I'm investigating this currently making a "how to record the pc speaker" video. Line-in is 1V peak to peak, so you normally wouldn't connect the 5v speaker to it, but I've found that a lot of line inputs have automatic gain control, so it ends up working out. While this isn't proper or scientific, I tapped one speaker lead and ran it into a Zoom recorder and it worked well, with the exception of picking up a ton of system EMI/RFI noise. (in the video, I'm exploring 6 other methods too)

Tiido wrote:

There are plenty of optrons that are used in MIDI gear and there you have 31250baud speed, plenty enough for sound. I don't remember any part number on top of my head though.

The Sharp PC900 is the classic solution. The modern option is the 6N138 which consists of an opto-triggered Darlington pair. Other transistor based opto-couplers will work with an external transistor completing the Darlington setup so long as rise and fall times remain in spec for MIDI and current gain is good.

The old Access games with "RealSound" suggested using a 4.7uF capacitor in series with the speaker signal and connecting it to line or aux in of a stereo. The "Mean Streets" manual had instructions for this.

1PC SPK  ----- (+) 4.7uF (-) --==--
2GND  ------------------------/

I forgot to mention that I did finally make a video about this:

https://youtu.be/wi2tW23qb9c

It's not 100% exhaustive, but I did cover a lot of different scenarios.