Re: But is UseNet "Social Media"??

On Mon, 2/2/2026 3:56 AM, Daniel70 wrote:
> On 2/02/2026 12:17 am, Carlos E.R. wrote:
>> On 2026-02-01 13:34, Mr. Man-wai Chang wrote:
>>> On 2/1/2026 6:17 PM, Daniel70 wrote:
>>>> On 1/02/2026 8:38 pm, Mr. Man-wai Chang wrote:
>>>>>
>>>>> The same goes to Fidonet BBS.
>>>>>
>>>> Never got into BBS's so couldn't possibly comment.
>>>
>>> Echomail is the technology used by Fidonet to run a distributed messaging network like Usenet. Back then, it ran on top of dial-up
>>> modems as slow as 1200 baud. Those were the days! :)
>>
>> For some strange reason, Fidonet continued working in Rusia after it
>> stopped in Europe. They took over the development of software, too.
>>
>> For a while at least, I could participate in Fidonet using internet instead of a modem.
> 
> Sorry! What?? I've never used Fidonet but, I'm guessing, the Ones and
> Zeros being produced by your Computer would have had to be converted to
> varying Tones to traverse the Phone Network someway .... which is what a
> Modem does.
> 
> I suppose you could have had a Modem Board with-in the Computer which
> then produced the varying tones required to traverse the Phone network.
> 
>> Then my uplink failed, his computer broke down and he gave up. Could
>> not find another suitable node.

The WinModem was a device based on having an audio chip
in a computer. It used voice frequency samples, for handling
modulated voice band transmissions. It saved money by not
having a DataPump and DSP code in hardware. It uses the
computer main processor for the DSP calculations.

What the WinModem needed, was a DAA, a Data Access Arrangement.
Using that thing, that provided some measure of "immunity" to
POTS. For example, if you discharged a thousand volts into the
RJ11 phone wire, none of that would make its way through
the DAA and blow up the computer. That is the part that always
must be provided, in some way, for a connection to POTS to be
sorta "safe". If the 11kV wire on the power pole in my backyard,
were to fall on the phone line, then if I had a modem I could be
kissing some part of the computer goodbye. Even with a DAA.

A fully provisioned modem, such as a Courier, it had a "datapump"
running at 85MHz. It had firmware. It talked modem-talk. That
was better than a WInModem, at least at the beginning.

When I got my laptop, one of the things I tested, was the
WinModem in the laptop, and it actually *beat* the Courier
on goodput in a 56K mode. And to do that, the core CPU on
the laptop donates about "100MHz worth of DSP activity"
to do the job over the Winmodem audio path. The DSP
sorts the tones in "buckets", the same sort of bucket
or bin scheme that ADSL uses in a sense.

In the early days, you could use a PLL to decode the tones.
Just going from memory, the CD4046 Phase Detector II output

   https://www.ti.com/lit/ds/symlink/cd4046b.pdf

   https://www.eleccircuit.com/cd4046-datasheet-phase-locked-loop/

uses flip flops and a state machine, and it should be a
-2Pi to +2Pi phase detector. The distinguishing thing
about such a phase detector, is it can do a wide range
clock, without a harmonic lockup problem. The other phase
detector, shown as a simple XOR gate, that can harmonic lock
so you could have a 1KHz input and the VCO could be running
at 3KHz in response. Whereas with phase detector II, a
a 1KHz input gives 1KHz output, a 5KHz input gives 5KHz output.
And we don't particularly care about the clock. We care about
the control voltage driving the VCO. It might measure 1 volt at
1KHz and 5 volts at 5KHz and exhibit a linear relationship
between frequency and voltage. Then, if your "POTS phone had
the tones", the VCO control voltage was a voltage domain
copy of what frequency was being carried at the moment.
It's F2V conversion.

I think you can imagine the hobbyist excitement at this point :-)

F2V conversion, may have worked fine with the ZX81 tape recorder
interface, but eventually the "sea of tones" on modem connections
was just too much for the "PLL crowd". That's when the DSP boys
took over. And DSP is how we got from 1200baud to 53300baud
(a little short of the marketing 56K).

   Paul