J. P. Gilliver (John) wrote:
> In message <
efplem...@mid.individual.net>, VanguardLH <V...@nguard.LH>
> writes:
>> "J. P. Gilliver (John)" <
G6JP...@255soft.uk> on 2017/02/05 wrote:
>>
>>> I've recently come to suspect that my internal speaker isn't working
>>> (one software I use has "use internal speaker" as an option [it's
>>> news/email software, and it's for indicating news/mail has arrived], and
>>> that doesn't seem to work).
>>
>> Do a cold boot on the computer. It should beep on startup. That will
>> indicate whether or not the onboard speaker works or not. This assumes
>> the computer actually has an onboard speaker (typically piezoelectric).
>
> It's a netbook. No bootup beep. As explained in other posts in this
> thread, when I tried the beep.exe that someone kindly provided a link
> to, it did beep, via the existing sound system (and as a configurable
> input to its mixer).
>
> On a lot of true desktops, especially older ones, it's a real speaker,
> albeit a grotty one. (Though not all - I've seen piezo ones on a stiff
> bit of wire to the speaker header.)
The standards included a PC Beep pin on the audio chip.
Description on Page 4
http://www.ti.com/lit/ds/symlink/lm4550b.pdf
The Southbridge (which contains the 8253 timer), has
a pin called "spkr", but it is buffered by a switching
transistor to drive the 1x4 SPKR header. And the
computer case speaker or motherboard piezo, gets
connected to that header.
It's possible there is an intent, to allow connecting the
spkr signal from the Southbridge to the PC Beep input, but
the schematic I've got, shows the AC'97 audio chip with the
PC Beep input AC-grounded and not used.
The SuperI/O apparently has an auditory signal it
would like to output, on a hardware monitor error
condition, but again, it's not clear if there is a
standards-compliant way of incorporating that or not.
When the BIOS wants to tell you about overheat, it
seems to use the 8253 timer itself for that.
When Asus installed their Vocal Post feature (a separate
Winbond chip), as far as I know, that was AC-coupled into
one channel of LineOut. And there was no attempt to be
"PC 99" about that.
I think I would need a time machine, to get details
on this one... The history of PC Beep.
And there's really nothing wrong with the speaker - what
is feeding the speaker is the atrocious part. You're
driving a square wave into the speaker.
If you have a "1-bit audio channel", it is possible
to provide decent fidelity - as long as there is a
reconstruction filter on the output to clean it up.
1-bit audio receives mention here. It was used
in a CD standard. And you might remember years ago,
seeing some stereo equipment with mention of 1-bit DAC.
It's all a matter of getting some good oversampling,
having a low pass on the output and so on.
https://en.wikipedia.org/wiki/Audio_bit_depth
Paul