Kermit

[gareth evans]

The loss of serial ports fitted to PCs mean that we have
also lost the use of the lowest-common-denominator
hardware-wise when it comes to transferring files
between incompatible systems.

But it also means that we have lost the confusion
when using Kermit of knowing to which end of the
link you are talking from the keyboard! :-)

[Johann Klammer]

You have to pay the FTDI tax.

[Robin Vowels]

On Friday, April 2, 2021 at 8:59:44 PM UTC+11, gareth evans wrote:
> The loss of serial ports fitted to PCs mean that we have
> also lost the use of the lowest-common-denominator
> hardware-wise when it comes to transferring files
> between incompatible systems.

.
Buy a USB to serial port converter.

[Andy Burns]

Robin Vowels wrote:

> On Friday, April 2, 2021 at 8:59:44 PM UTC+11, gareth evans wrote:
>
>> The loss of serial ports fitted to PCs mean that we have
>> also lost the use of the lowest-common-denominator
>> hardware-wise when it comes to transferring files
>> between incompatible systems.
>

> Buy a USB to serial port converter.

Check if there's a 10 pin 0.1" serial header on the motherboard (then
wonder which pinout it follows).

[Ahem A Rivet's Shot]

Wire randomly to 25 pin D type, plug into breakout box, watch
flashing lights, rewire appropriately (or leave the jumper in place).

--
Steve O'Hara-Smith | Directable Mirror Arrays
C:\>WIN | A better way to focus the sun
The computer obeys and wins. | licences available see
You lose and Bill collects. | http://www.sohara.org/

[Carlos E.R.]

On 02/04/2021 14.22, Robin Vowels wrote:
> On Friday, April 2, 2021 at 8:59:44 PM UTC+11, gareth evans wrote:
>> The loss of serial ports fitted to PCs mean that we have
>> also lost the use of the lowest-common-denominator
>> hardware-wise when it comes to transferring files
>> between incompatible systems.
> .
> Buy a USB to serial port converter.

It is not complete. The latency is not constant, for instance. No
dedicated direct IRQ.


My current computer, bought a year ago or so, has both a serial port and
a parallel port in the MB (I haven't tested them).


--
Cheers, Carlos.

[Ahem A Rivet's Shot]

On Fri, 2 Apr 2021 15:09:10 +0200
"Carlos E.R." <robin_...@es.invalid> wrote:

> On 02/04/2021 14.22, Robin Vowels wrote:

> > Buy a USB to serial port converter.
>
> It is not complete. The latency is not constant, for instance. No
> dedicated direct IRQ.

OTO some of them go up to 1.5mbps.

[Carlos E.R.]

On 02/04/2021 15.57, Ahem A Rivet's Shot wrote:
> On Fri, 2 Apr 2021 15:09:10 +0200
> "Carlos E.R." <robin_...@es.invalid> wrote:
>
>> On 02/04/2021 14.22, Robin Vowels wrote:
>
>>> Buy a USB to serial port converter.
>>
>> It is not complete. The latency is not constant, for instance. No
>> dedicated direct IRQ.
>
> OTO some of them go up to 1.5mbps.

Yes, that is possible, USB is fast, but it is a different issue.

When a byte arrives at the true serial port, an IRQ is activated, and
unless the CPU is on uninterruptible state, it responds "instantly",
meaning within a time limit.

USB is a bus. If the bus is busy attending some other device, the pseudo
port will not be attended till finish. It is a polling system. It is not
that the response is slow or fast, but that the time to respond is not
defined, it is actually variable.

This is important for time sync with GPS receivers, for instance. A GPS
receiver on 9600 bps on a real serial port can thus be more accurate
than with an USB converter working at 1M.


--
Cheers, Carlos.

[songbird]

Robin Vowels wrote:
> On Friday, April 2, 2021 at 8:59:44 PM UTC+11, gareth evans wrote:
>> The loss of serial ports fitted to PCs mean that we have
>> also lost the use of the lowest-common-denominator
>> hardware-wise when it comes to transferring files
>> between incompatible systems.
> .
> Buy a USB to serial port converter.

reminds me of the time when i needed a cable and
all i had were some paperclips, serial cable ends,
a chunk of phone line and some match light solder.

yes, it did work long enough for me to transfer
the contents of the machine, but it was pretty warm
by the time i got done.

the current motherboard does have a serial port
header on it, but i'd be more likely to use a
converter as i have extra USB boards i picked up
for little to make sure i would have them if i
ended up wearing them out by plugging/unplugging
things.

as it has turned out the USB ports survived
longer than the mobo/cpu in that incarnation so...


songbird

[Joe Pfeiffer]

If you have a need of true bounded response times you shouldn't (in
fact, can't) use a modern desktop OS. Do your time-critical processing
on a $2.00 SBC and have it talk to your desktop over USB.

[Charles Richmond]

Sounds like a job for a Raspberry Pi or Arduino board -- local memory
to hold the "bounded response time" results coupled with onboard USB.


--

Charles Richmond

--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

[Carlos E.R.]

Not if the goal is run the ntp daemon as master in the computer. The
same machine should run the daemon and be connected to the GPS clock via
RS232 serial port with IRQ handling. If those small machines can do it,
I have no idea, maybe.


--
Cheers, Carlos.

[Joe Pfeiffer]

Oh, hang on, NMEA GPS messages contain the time. Reasonable latency
(even if not strictly bounded) isn't a problem.

[Mike Spencer]

Nah, we still have that. I had a hellava time explaining how to use
ftp to a friend. He was a non-programmer but after over a decade
working in a Unix shop, doing variously NTSC video, educational
multimedia project dev, management etc. etc. and using telnet
routinely, he was still very slow to grasp which end he was talking
to when trying to fetch half a career's files home ove ftp.

I have a serial->USB adapter to allow my no-serial-port laptop to do
dialup. It proved to be finniky about exactly which bytes (such as a
newline) were sent when. Required tedious trial & error and help from
kind persons on the net to get it to work.

I'm a rank amateur hacker/programmer and I managed to build a setup
with a mag switch in our pet mouse's cage connected to button 1
terminals on a serial mouse and then read, log and time revs of the
mouse's exercise wheel via the serial port. I would have been
completely stymied if all that was available had been a USB port. (At
his best, the mouse ran about 4 miles in an evening with numerous
snack and water breaks!)

--
Mike Spencer Nova Scotia, Canada

[Scott Lurndal]

>"Carlos E.R." <robin_...@es.invalid> writes:
>
>> On 02/04/2021 15.57, Ahem A Rivet's Shot wrote:
>>> On Fri, 2 Apr 2021 15:09:10 +0200
>>> "Carlos E.R." <robin_...@es.invalid> wrote:
>>>
>>>> On 02/04/2021 14.22, Robin Vowels wrote:
>>>
>>>>> Buy a USB to serial port converter.
>>>>
>>>> It is not complete. The latency is not constant, for instance. No
>>>> dedicated direct IRQ.
>>> OTO some of them go up to 1.5mbps.
>>
>> Yes, that is possible, USB is fast, but it is a different issue.
>>
>> When a byte arrives at the true serial port, an IRQ is activated, and
>> unless the CPU is on uninterruptible state, it responds "instantly",
>> meaning within a time limit.
>>
>> USB is a bus. If the bus is busy attending some other device, the
>> pseudo port will not be attended till finish. It is a polling
>> system. It is not that the response is slow or fast, but that the time
>> to respond is not defined, it is actually variable.

Actually, USB is very definitly _not_ a bus, notwithstanding the 'B'
in USB. It's a high-speed point to point connection with low latency.
Far lower latency than a 115kbaud or even 1mbaud serial port.

It certainly does have dedicated "IRQ" (albeit via MSI-X or IO-APIC).

Some latency is added if you have an external USB hub, but there
shouldn't be any noticable contention delays for any but the
cheapest low-end hub chips.

[Carlos E.R.]

On 02/04/2021 22.15, Scott Lurndal wrote:
>> "Carlos E.R." <robin_...@es.invalid> writes:
>>> On 02/04/2021 15.57, Ahem A Rivet's Shot wrote:
>>>> On Fri, 2 Apr 2021 15:09:10 +0200 "Carlos E.R." <robin_...@es.invalid> wrote:
>>>>> On 02/04/2021 14.22, Robin Vowels wrote:
>>>>
>>>>>> Buy a USB to serial port converter.
>>>>>
>>>>> It is not complete. The latency is not constant, for instance. No
>>>>> dedicated direct IRQ.
>>>> OTO some of them go up to 1.5mbps.
>>>
>>> Yes, that is possible, USB is fast, but it is a different issue.
>>>
>>> When a byte arrives at the true serial port, an IRQ is activated, and
>>> unless the CPU is on uninterruptible state, it responds "instantly",
>>> meaning within a time limit.
>>>
>>> USB is a bus. If the bus is busy attending some other device, the
>>> pseudo port will not be attended till finish. It is a polling
>>> system. It is not that the response is slow or fast, but that the time
>>> to respond is not defined, it is actually variable.
>
> Actually, USB is very definitly _not_ a bus, notwithstanding the 'B'
> in USB.

It is definitely a bus. You connect an USB HUB to the computer and on it
you can have dozens of USB gadgets.

There is no guarantee that the computer uses an internal hub itself
instead of a dedicated chip for each socket.

Latency on an RS232 port does not depend on port speed, but on how fast
the CPU switches to the IRQ procedure.



--
Cheers, Carlos.

[J. Clarke]

On 02 Apr 2021 16:28:39 -0300, Mike Spencer

This is the sort of thing a Raspberry Pi was made for.

[Joe Pfeiffer]

"Carlos E.R." <robin_...@es.invalid> writes:

> On 02/04/2021 22.15, Scott Lurndal wrote:
>> Actually, USB is very definitly _not_ a bus, notwithstanding the 'B'
>> in USB.
>
> It is definitely a bus. You connect an USB HUB to the computer and on
> it you can have dozens of USB gadgets.

That's not the definition of a bus.

[Richard Thiebaud]

Then what is?

[Joe Pfeiffer]

The key feature defining a bus is multiple sources and destinatiosn on
the wires. Looking around a bit, here's a really good introduction to
the topic:
http://www.cse.unsw.edu.au/~cs2121/LectureNotes/15s1/week4_notes.pdf

[Mike Spencer]

I drove into the city and bought an Arduino (well, one that claimed to
be Arduino-libraries compatible) but then decided the serial mouse
hack would be easier than beating up the Arduino libraries HOWTO. I
wrote serial port code for the Osborne years ago so I knew more or
less where I was going.

[Carlos E.R.]

On 03/04/2021 05.27, Joe Pfeiffer wrote:
> Richard Thiebaud <thieba...@aol.com> writes:
>> On 4/2/21 6:17 PM, Joe Pfeiffer wrote:
>>> "Carlos E.R." <robin_...@es.invalid> writes:
>>>
>>>> On 02/04/2021 22.15, Scott Lurndal wrote:
>>>>> Actually, USB is very definitly _not_ a bus, notwithstanding the 'B'
>>>>> in USB.
>>>>
>>>> It is definitely a bus. You connect an USB HUB to the computer and on
>>>> it you can have dozens of USB gadgets.
>>> That's not the definition of a bus.
>>>
>> Then what is?
>>
> The key feature defining a bus is multiple sources and destinatiosn on
> the wires.

No. But have it your way.

--
Cheers, Carlos.

[gareth evans]

There's a chicken-and-egg situation there, for how does one transfer
the files to the RPi from a system that is incompatible with it? :-)

[Scott Lurndal]

The pie runs linux, so any linux system is compatible with it.