should be the most simple, low level audio signal, comparable to any headphone port of phones or ipods.

So, no hands-on experience here, but a hint: You can attach it directly to a dedicated header on many sound cards, regulate the volume by the usual mixer and proceed from there.
No one ever did because a) the low level speaker is more of a diagnostic tool and b) why would you want to hear IBM noise, you do have a sound card.
Try "Silent Service II" for really amazing, high end IBM sound 😉

Attaching it to the aux or CD in should do much the same by any other name.

Raw PC speaker signal is 5 or 3.3Vpp squarewave, depending on particular hardware. Most recording devices need much less than that, so something likely needs to be done to prevent possible damage. There are some circuits on this forum somewhere that turn the PC speaker output into something resembling line out that you can then record safely.

Shponglefan wrote on 2022-03-30, 12:51:

Or using an external microphone to record the speaker itself?

definitely microphone, any other recording method just doesn't represent what a PC speaker sounds like imho
the small cone speaker attenuates the signal in a certain way that just makes it sound different.

this is my very makeshift recording setup, the speaker is my original PC speaker from my very first PC from 1989 that got salvaged out of the case at some point and lingered in a box for decades before it got put to use again, still sounds great 😀
audio interface has since changed, but the basic setup is still the same ^^

here's what that sounds like:

http://www.oldskool.org/guides/speakerrecording

Also, common ALS100 cards can play PC Speaker sound without the need for any special cable - they listen to the appropriate I/O on the bus.

imi wrote on 2022-03-30, 13:48:

definitely microphone, any other recording method just doesn't represent what a PC speaker sounds like imho the small cone speak […]
Show full quote

Shponglefan wrote on 2022-03-30, 12:51:

Or using an external microphone to record the speaker itself?

definitely microphone, any other recording method just doesn't represent what a PC speaker sounds like imho
the small cone speaker attenuates the signal in a certain way that just makes it sound different.

this is my very makeshift recording setup, the speaker is my original PC speaker from my very first PC from 1989 that got salvaged out of the case at some point and lingered in a box for decades before it got put to use again, still sounds great 😀
audio interface has since changed, but the basic setup is still the same ^^
setup01_08.jpg

Interesting!
The case speaker is supposed to be the most ordinary speaker, ever. It's a small, highly efficient broadband type used in $5 transistor radios when those were still a thing. So, all of them probably came from the same factory in Canton.
It's designed to translate electricity into sound waves and the microphone to do the reverse. In a perfect world, it wouldn't make any difference.
But sure, a direct recording played back on any other device will never quite sound the same as on that squeaky little wart.
So, by recording it with a microphone, you're basically cultivating the imperfections of the speaker.
Which is perfectly legit - anyone recording a guitar by putting a microphone up to the Fender amp does just the same.

OP, just keep in mind that there's a huge spread in price, quality and characteristics in microphones so don't expect to get it right from the start.

imi wrote on 2022-03-30, 13:48:

here's what that sounds like: MIintroPCspk.mp3 SKYCATintroPCspk.mp3 Indy3introPCspk.mp3 duke_sounds.mp3 […]
Show full quote

here's what that sounds like:
MIintroPCspk.mp3
SKYCATintroPCspk.mp3
Indy3introPCspk.mp3
duke_sounds.mp3

Oh boy, that's a whole lot of squeaky goodness, we've come a long way!
Still everyone's favorite sound device:
https://youtu.be/uSZ0qkpHucs?t=19

Thanks for the replies everyone!

Sounds like recording with a mic is the way to go. I'll try rigging up a setup for that and see how it goes.

imi wrote on 2022-03-30, 13:48:

definitely microphone, any other recording method just doesn't represent what a PC speaker sounds like imho the small cone speak […]
Show full quote

Shponglefan wrote on 2022-03-30, 12:51:

Or using an external microphone to record the speaker itself?

definitely microphone, any other recording method just doesn't represent what a PC speaker sounds like imho
the small cone speaker attenuates the signal in a certain way that just makes it sound different.

this is my very makeshift recording setup, the speaker is my original PC speaker from my very first PC from 1989 that got salvaged out of the case at some point and lingered in a box for decades before it got put to use again, still sounds great 😀
audio interface has since changed, but the basic setup is still the same ^^
setup01_08.jpg

here's what that sounds like:
MIintroPCspk.mp3
SKYCATintroPCspk.mp3
Indy3introPCspk.mp3
duke_sounds.mp3

That seems like a great setup. Very clean recordings with no background noise that I could hear.

When I was recording PC Speaker music, I just made a 3.5mm jack adapter for the PC Speaker header and used the mic-in on my Sound Blaster ZX-R on my main computer.

Works just fine.

The problem with just using a mic is that there are a bunch of different styles of speakers. Anything from onboard to the little piezo speakers to 2" or larger speakers that mount in the case.

And the mic is going to play into it as well.

I'd rather do the minimum amount of signal conversion possible

The PC Speaker is part of the PIT chip, a system timer chip and uses 5v DC.
Shorting it can destroy the motherboard.

To avoid that, we may add a 10uF capacitor in series to whatever we want to connect.
That way, no DC can flow. However, I'm not exactly sure if this will affect the sound output way too much.

Another idea is to connect an optoinsulator with a resistor to the speaker. Like an LED essentially.
Maybe in parallel to an existing speaker device.
That way, no harm should be caused.

Anyway, there might be other options, too. Like magnetocouplers (gmr?), which are considered successors to optocouplers?

Edit: I like the idea of using the microphone, btw.
It's the most faithful, perhaps. A silent room for recording would be cool. Egg boxes made of cardboard are very good at absorbing sound.
Just use a shack or room and glue them to the wall. 😉

Edit: Never mind. Grzyb's link (http://www.oldskool.org/guides/speakerrecording) already described the wiring process.
Even though I'm a bit surprised/puzzled that the author spent so much work at it. I mean, it's not wrong. It's good, almost scientific.👍

Simple minds like me would have just used an estimation of the values, a simple GW-BASIC program that does PLAY something and a multimeter to find out the peak voltage etc. 😅
If it was too high, a voltage divider could have been used. Or some diodes in series, installed before the cap, maybe, not sure.

Anyway, the recommendation of a transformer makes sense. I really appreciate that.
However, fancy AF transformers aren't available at every corner.
They can sometimes be found in antique transistor radios and cheap 27MHz walkie talkies, though.
Personally, I often use old AC tranformers thud, 220v/12v lr 220v/6v or something, for my radio hobby (bandwith of the voice channels ~3KHz).
The high voltage side is of high impedance, the low voltage side..
These transformers inside heavy old power bricks do contain an iron core, which isn't great. But about ~5 to ~10 KHz bandwith should be possible with them.

Tiido wrote on 2022-03-30, 13:05:

Raw PC speaker signal is 5 or 3.3Vpp squarewave, depending on particular hardware. Most recording devices need much less than that, so something likely needs to be done to prevent possible damage. There are some circuits on this forum somewhere that turn the PC speaker output into something resembling line out that you can then record safely.

Hello, I am newbie on this forum, and my English is not good. I am russian medical devices engineer. I use decimal comma, not dot.

Generic unbalanced dynamic microphone input should be able to record 50 mV (0,050 V) sine wave undistorted. It also known as „magnetic head input” or „magnetic cartridge input” — it is all different, but on this abstraction level it is all a same.

Generic line input (and output vise versa) is operate with sine up to 500 mV (0,5 V) and impedance of load between 1000~100 Ohms. Usually this is an output of correction amplifier (in case of magnetic cartridge) or preamp (in case of microphone).

Generic headphone output operates with impedance 70~15 Ohms and voltage levels 500~2000 mV (0,5~2 V). At today technology level, line and headphone outputs are same, but in 1980's it was two different outputs — remember this!

Now we need to record signal from square low impedance output using sine middle impedance input. It is very easy — we need an attenuator. In case of big relational difference of impedances, we need to save original loudspeaker for impedance negotiation and connect an two resistors in series between loudspeaker terminals. The voltage on resistor, connected to Ground terminal (Rminus), is smaller, than overall voltage. And relation between voltages is a fraction:
Vminus = (Rminus) / (Rminus + Rplus)

Ok, this is very simplyfied theory, but what we shall to do in practice?

We need: Soldering Iron (15~50 W power), few wires (simular to loudspeaker's wires), connector for your soundcard (usually TRS jack), and an Trimming Resistor. I recoomend use any Bourns 3296 series in any style, but resistanse shold be 100 Ohm (Bourns resistanse code is 101). It is possiible to use 3296W-1-101, 3296P-1-101, 3296X-1-101 or equivalent from any another manufacturer.

First time you need to initialise your Trimmer. If you have no Ohmmeter at home or office, you may ask for help in shop (at least, in Russia this trick is useful). You need { (0,5 V /5 V) * 100 Ohm = 10 Ohm} between terminals 1 and 2 of Trimmer. If you have no meters, you may check the number of turns for full whiper travel in datasheet of your trimmer (25 turns for Bourns 3296), rotate some more CCW, and 10% of full to CW (2,5 turns for 3296). And 100 Ohm with 5 Volts provide 0,050 Ampers of current or 5 V * 0,050 A = 0,25 Watts of power — it is less than fire burning limit.

Now solder wires from pin 1 of trimmer to the Speaker ground (usually Black wire, connected to minus of the Loudspeaker) and to TRS Jack ground (usually S). Then solder an wire between pin 2 of Trimmer and Tip of the TRS Jack. And solder last wire between pin 3 and the plus of loudspeaker.

NOW CHECK ALL CONNECTIONS, DOUBLE CHECK POLARITY OF LOUDSPEAKER AND ITS CONNECTOR. And you should understand, what connecting two PC via wires is electricaly unsafe — TRIPLE CHECK YOU GROUNDING FOR YOUR (not PC) SAFETY!!!

You must control signal level on trimmer using small screwdriver. Sound should be undistorted. In any distortion you should slowly rotate wiper screw of Trimmer CCW.

I have no chanses to make pictures until tomorrow evening. But I will draw them if you requested.

Best regards,
Mikhail "SMA" Shpak
unemployed Service Engineer
(foremly work in Bakulev Science Center, DRG (Moscow represetative office), and few medical equipmet dealers and distributors in Russia).

Jo22 wrote on 2022-04-01, 00:56:

Edit: I like the idea of using the microphone, btw. It's the most faithful, perhaps. A silent room for recording would be cool. […]
Show full quote

Edit: I like the idea of using the microphone, btw.
It's the most faithful, perhaps. A silent room for recording would be cool. Egg boxes made of cardboard are very good at absorbing sound.
Just use a shack or room and glue them to the wall. 😉

Edit: Never mind. Grzyb's link (http://www.oldskool.org/guides/speakerrecording) already described the wiring process.
Even though I'm a bit surprised/puzzled that the author spent so much work at it. I mean, it's not wrong. It's good, almost scientific.👍

Anyway, the recommendation of a transformer makes sense. I really appreciate that.
However, fancy AF transformers aren't available at every corner.
They can sometimes be found in antique transistor radios and cheap 27MHz walkie talkies, though.
Personally, I often use old AC tranformers thud, 220v/12v lr 220v/6v or something, for my radio hobby (bandwith of the voice channels ~3KHz).
The high voltage side is of high impedance, the low voltage side..
These transformers inside heavy old power bricks do contain an iron core, which isn't great. But about ~5 to ~10 KHz bandwith should be possible with them.

It should be bad idea: standart IBM PC Speaker is just an loudspeaker connected to CMOS Buffer. It have only two states: ON and OFF. And freqency is not fixed - it may be bigger, than 60 kHz. What transformer you want use?
https://wiki.osdev.org/PC_Speaker#Pulse_Width_Modulation

It is better to use optocouplers if you want use transformer. But, during switching voltage swing was intensified by inductance of Speaker. In this case building accurate converter is too complex work. For recording all taste of sound in this situation direct connect is better when good microphone - speaker drive negative voltage by itself. Simular effect is provided by electromagnetic relays - for protection from reactive currents during shutting down an diodes in reverse polarity is widely used.
https://resources.altium.com/p/using-flyback- … e-your-circuits

Jo22 wrote on 2022-04-01, 00:56:

Edit: Never mind. Grzyb's link (http://www.oldskool.org/guides/speakerrecording) already described the wiring process.
Even though I'm a bit surprised/puzzled that the author spent so much work at it. I mean, it's not wrong. It's good, almost scientific.👍

Good link, but direct connection should not use an capacistor. With connected loudspeaker you retrive an filter with uncalculatable parameters (loudspeaker is not an ideal inductor). Without speaker it should be an high-pass filter, but sound will be different without selfinductance of speaker and flyback reactive current.

SMA wrote on 2022-04-03, 13:29:

NOW CHECK ALL CONNECTIONS, DOUBLE CHECK POLARITY OF LOUDSPEAKER AND ITS CONNECTOR. And you should understand, what connecting two PC via wires is electricaly unsafe — TRIPLE CHECK YOU GROUNDING FOR YOUR (not PC) SAFETY!!!

And please understand - all direct connection practice are UNSAFE. It is good for episodic use only.

But if you are Professional, you have knowledge to make an insulated amplifier (in example, using special IC), and provide safe and accurate solution. But PRO will not ask questions...

This is kinda funny. Unless you are hooking the computers up to completely separate power sources, they are going to be sharing a common ground through the power cords.

Plus you are only getting a max of 12v through the power supply unless you have some sort of catastrophic failure or lightning strike.

There is so much protection built into computers that you shouldn't have to worry at all about any amount of power that could be considered dangerous.

cyclone3d wrote on 2022-03-30, 17:05:

The problem with just using a mic is that there are a bunch of different styles of speakers. Anything from onboard to the little piezo speakers to 2" or larger speakers that mount in the case.

And the mic is going to play into it as well.

I'd rather do the minimum amount of signal conversion possible

during the late 80s / early 90s pretty much any IBM clone machine had the same 2" cone speaker so I think it's fair to say that was the "standard" even if there was some system with a piezo speaker or different setup here or there
sure, there can also be differences between these, especially considering the one I'm using is at least 33 years old now, but for me that just adds to the charme, especially because it's the very same speaker I played all of these games on originally ^^

I also chose a neutral universal instrument mic for that reason, it wasn't even expensive: https://www.thomann.de/intl/superlux_hi_10.htm
as I have the direct comparison between listening to the speaker and the recording I can say it's pretty faithful ^^

the problem with the original square wave signal is that that's not what anyone heard when they were using a PC speaker, sound cards only supported the speaker signal later on, but then again that was more to listen to system sounds than anything, cause if you had a sound card you were primarily using that ^^
I did not have a soundcard until at least 1995 I think, so most games I played with PC speaker and it's very close to my heart ^^

it's a bit like pixels of old games, nobody ever saw the pixel graphics like they are displayed natively on a LCD screen today, they all were viewed through a CRT as a "filter" and many artists took that into account to make use of that, and I'm sure the same is true of the PC speaker.

so in that sense, "signal conversion" is necessary to get the "true" sound of what a PC speaker is supposed to sound like, as the signal itself does not represent that.

and recording the pure signal does not give you any benefit imho, as it's not easy to play it back again on a fitting speaker to make it sound what it was originally sounding like, as the pc speaker output is a very particular beast as mentioned in this thread ^^ the "true signal" so to speak, is bits.
recording pure native pixels from a game for example certainly has an advantage as you can always just go back and display it on a CRT as it was like any time.

Hi,

I did 2 Video where I recorded PC Speaker games:
https://www.youtube.com/watch?v=8M0miQ6vtaE
https://youtu.be/UK_0K1ibMAk

cyclone3d wrote on 2022-04-03, 16:27:

Plus you are only getting a max of 12v through the power supply unless you have some sort of catastrophic failure or lightning strike.

There is so much protection built into computers that you shouldn't have to worry at all about any amount of power that could be considered dangerous.

You are right. If you make great attenuator, but make mistake in connection, you just connect 5V оf UltraRareOldPC to ground of your recording device… In good result your UltraRareOldPC would have an burning multi-io or LPC chip… In bad result you will trash mainboard and PSU…

I have some ideas to make the Ideal Pro Soulution for this job… But what people make the PCB for VOGONs users if I develop something? Is VOGONs have users with elctronics DIY skill?

P.S. I forget good solution for this job — an COVOX! This is R-2R matrix was not known in Russia, but popular in US. LPT have 5V signalling and you just need proper connection…

Or, maybe we simply use the speaker output to drive a transistors? As a pull-up or pull-down ?

If we add a diode and a resistor directly between speaker output (SPK) and the transistor base, it should provide some sort of protection.
The transistor simply acts as a quick and semi-digital switch:
The signal that actually interfaces with the soundcard nolonger is the speaker output itself. It rather comes from closing a circuit by connecting it to ground or power.

I vaguely remember that my fellow hams used similar circuitry to connect an RS232 pin to control PTT of a radio transceiver.

Some also used an additional relay with an anti-parallel diode, which prevented backflow of the inversely polarized energy of the inducted coil.

Edit: That being said, the core problem might still persist.
DC doesn't belong to AF applications. Feeding a soundcard with a DC signal isn't a good manner.

Also: The MIC input of the soundcard has some DC bias (phantom voltage) to power condenser microphones. So better use Line-In.

On SB style sound cards, it's on an extra pin, but some devices like smartphones may have it differently wired.
In that case a capacitor is required to filter out the DC.

Anyway, that's a good moment to remember not to use any kind of mono connectors anymore.
Even if you don't use stereo. Same for the other way round.
If you need a mono signal, just use the left audio channel and the "left" pin on a stereo plug (and ground).
Let the "right" pin unconnected. That way, stereo/mono devices can be interchanged without worrying.

Edit: Typos fixed.

Yes, this is good idea, but your record will not differ from the DosBox simulation. It is same as direct connect recording without speaker connected. If you want record the real taste of PCSPK, you should record flyback reactive current too…

I think, better is using the SilLabs SI8910 insulated amplifier… and we need additional Nyquist filter in case of infinite spectrum of the Delta function.