Re: ...thoughts on Retro-uC?

Den Thu, 13 Sep 2018 09:15:49 +0200 skrev Francesco Messineo
<francesco.messineo@gmail.com>:
> On Thu, Sep 13, 2018 at 8:50 AM <groepaz@gmx.net> wrote:
> >
> > Am Donnerstag, 13. September 2018, 07:53:46 CEST schrieb Francesco
> > Messineo:
> > > On Thu, Sep 13, 2018 at 7:43 AM Baltissen, GJPAA (Ruud)
> > >
> > > <ruud.baltissen@apg.nl> wrote:
> > > > > if a 6526 clone would be simple - we'd have one by now
> > > >
> > > > I see it in a different way: if there was a demand for the
> > > > 6526, we would have one by now.
> > > >
> > > > Take the PLA as example: there was demand for it and one or
> > > > more guys cam with the EPROM solution and others various CPLD
> > > > replacements. So IMHO as soon there is a real demand for it,
> > > > somebody will create a replacement.
> > > Well, I agree more with Ruud, but the example is not well chosen
> > > because the PLA is easy (if we exclude timings problems, but if
> > > you use an EPROM, you're anyway adding different timings), CPLD
> > > versions of the PLA can be exact replacements even timing-wise.
> > > 6526 isn't so eadiesy, it's a complex piece of statefull logic,
> > > but it's doable. There're probably still too many real parts to
> > > start worrying.
> >
> > i've seen far more dead CIAs than dead PLAs so... oh well :)
> 
> sure, same here, it's like 10 dead (or almost, actually) CIAs against
> 1 dead PLA, and the ratio in the good old '80s was even larger.
> The reason is the CIAs are sitting naked against the external world:
> joystick ports, user port, IEC bus (well, the IEC bus has some gates
> between the wild world and the CIA ports anyway).
> We used to experiment (most of us), connect the wrong thing on the
> user port, plug and unplug things with everything powered on...
> I have tens of CIAs with just one or two bits on a port broken and
> otherwise working well.

To some extent people are probably more careful with their devices
today than back in the 80's.

Also mains power with ground connector is more common now than back
then.

In the 80's in many places (Sweden, the Netherlands and so on) you'd
have a 1541 hooked up to an unearthed socket leaking mains power to
signal ground. In some cases you'd have a TV where the aerial were
grounded through the cable TV or multi-apartment antenna system. In
some other cases you'd have a then modern TV with insulated chassis and
a euro connector for power, with a switch mode power supply leaking
mains power to signal ground. With the antenna grounded you'd have
current flow through signal ground from the 1541's mains filter through
the C64 to the TV. With a then modern TV (or monitor, like the
1084/8833 ranges) you'd also have a current through the same signal
ground path but this time due to that the two capacitors with identical
spec in the mains filter not being totally identical.

Nowdays you'd have a grounded socket or at least a grounded extension
cord and in many cases a grounded power lead to the TV/monitor.

Also people using 1541-II, SD2IEC, 1541 Ultimate or Pi1541 instead of a
real 1541 (or 1570 or 1571) removes the leaky power filter on one side.

(The C64 is perfectly insulated from mains power, so it in itself won't
cause this kind of problem)

(Btw it's not the current itself, but the voltage when
connecting/disconnecting cables and having contact with a signal wire
but not signal ground for a moment).

What seems to still fry C64 and VIC 20 CR is people forcing the power
DIN connector in the serial port. Most PSU's only have 4 pins and some
of them have so wobbly plastics that you don't even need much force to
force the connector in the serial port even though the pin positions
doesen't match. (I've got a cut-off connector from a C64 PSU which I've
tried in the serial port of at least two VIC 20 CR motherboards, no
unreasonable force is needed to insert it in the wrong port).

Failed CIA's due to incorrect handling are problably something that
more often happens when someone finds av C64 at a flea market or in an
attic, and then they maybe just throw it away. Failed PLA's and SID's
would rather happen when a computer is in use, and that would probably
lead to people being more inclined to get their computer repaired. But
this is just an educated guess.

> PLAs on the other hand could only die because poor manufacturing
> process (the CBM ones) coupled to high power dissipation, but I still
> have to find a broken original programmed PLS100 or 82S100.

Isn't there problems with broken 82S100's in the 8296?

-- 
(\_/) Copy the bunny to your mails to help
(O.o) him achieve world domination.
(> <) Come join the dark side.
/_|_\ We have cookies.