Cc_reset

[whitebird]

Hello,

The PCMCIA reset signal is mentionned as an output on gayle's datasheet, however PCMCIA's reset correction circuits transiently pull this signal to 5V which creates a hardware conflict (short). How this solution is made acceptable without damaging gayle?

[Predseda]

Maybe this?

http://amigakit.leamancomputing.com/...roducts_id=637

But I have bought it and it doesn't seem to help my Amiga.

[whitebird]

The solution from Amigakit implements a transistor for driving the CC_RESET signal from the _RST signal, but it basically acts the same way as other capacitor based solution by pulling the CC_RESET high.

So if the CC_RESET is a true output from Gayle, the fix will create a hardware conflict. Maybe CC_RESET is an input at startup which lets enough time for driving the CC_RESET signal "externally"?

[Daedalus]

Usually those sort of signals are marked as bidirectional on the schematic, but it's marked as an output only so that's probably not the case. Given that it has a pull-down resistor applied, it's likely that the output is high impedence in the low state, meaning that you can safely put 5V across it. If it is an open-collector-type output, it will have two states - high and off as opposed to high and low, so applying +5V won't short it out.

Open-collector outputs are used for interrupt lines for this very reason, so that one part issuing the interrupt doesn't drive another output.

[whitebird]

Open-collector outputs are used for interrupt lines for this very reason

Which allows the well known wired OR making an OR gate useless.

Open collector outputs are mentionned as "o c" within gayle's datasheet but maybe a mistake has been made and CC_RESET is a PNP open collector output which will require a pull-down resistance instead of a pull up required for NPN transistors...

[Daedalus]

Originally Posted by whitebird

making an OR gate useless.

Not sure what you mean here. There are no OR gates on the CC_RESET line.

Open collector outputs are mentionned as "o c" within gayle's datasheet but maybe a mistake has been made and CC_RESET is a PNP open collector output which will require a pull-down resistance instead of a pull up required for NPN transistors...

Yep, that's my thinking, since there's a pull-down on the schematic which wouldn't otherwise be needed.

[E-Penguin]

You can do it in software with CardPatch and CardReset on aminet. It works well enough on my A1200 at least.

[whitebird]

Not sure what you mean here. There are no OR gates on the CC_RESET line

You mention interrupt lines as an example of open collector usage, so I completed by stressing that open collector can be seen as a replacement for an OR gate.

[whitebird]

Doing it from software will not avoid removing and hot reinserting the card each time the Amiga is powered as the drivers will not be ready fast enough and as far as I know the presence of the card will prevent the Amiga from booting.

[FOL]

Originally Posted by whitebird

as far as I know the presence of the card will prevent the Amiga from booting.

Not in all cases, I find in the rare board rev / occassions it stops the machine from booting. Other boards / rev, dont care and continue booting.

[whitebird]

Mine is rev 1D4

[Daedalus]

Originally Posted by whitebird

You mention interrupt lines as an example of open collector usage, so I completed by stressing that open collector can be seen as a replacement for an OR gate.

Ah, ok, I didn't quite get that from what you said, but yes, that's their advantage.

I've found that the presence of the card on powering up will not prevent my Amigas (three A1200s: two 1D4s and a 1D1) from booting - it just stops the card from being initialised properly, but Amigas are known for having strange quirks and differences even on the same revision of board.

[FOL]

Originally Posted by Daedalus

Ah, ok, I didn't quite get that from what you said, but yes, that's their advantage.

I've found that the presence of the card on powering up will not prevent my Amigas (three A1200s: two 1D4s and a 1D1) from booting - it just stops the card from being initialised properly, but Amigas are known for having strange quirks and differences even on the same revision of board.

I agree, no 2 are a like. Hell even same rev boards are different.

[whitebird]

The best is to give it a try.

The cause that prevents the Amiga from booting is probably well known. Is it kickstart related?

[FOL]

Originally Posted by whitebird

The best is to give it a try.

The cause that prevents the Amiga from booting is probably well known. Is it kickstart related?

I don't beleive so, as I have seen it on both KS3.0 and KS3.1.

[whitebird]

Maybe it has not been fixed in kickstart 3.1.

If it does not rely on kickstart, then it is hardware dependent, but it is hard to believe there are such variations between Amigas within the same revision.

I'm asking myself which signal from the PCMCIA port can cause the Amiga to hang.

[FOL]

Originally Posted by whitebird

Maybe it has not been fixed in kickstart 3.1.

If it does not rely on kickstart, then it is hardware dependent, but it is hard to believe there are such variations between Amigas within the same revision.

I'm asking myself which signal from the PCMCIA port can cause the Amiga to hang.

I have have seen atleast 500+ boards over the years and trust me, same rev does not mean same rev.

[E-Penguin]

Originally Posted by whitebird

Doing it from software will not avoid removing and hot reinserting the card each time the Amiga is powered as the drivers will not be ready fast enough and as far as I know the presence of the card will prevent the Amiga from booting.

I've not seen the Amiga hanging, but without CardPatch and CardReset it certainly won't use the network card. CardReset is a software remove-insert to let the drivers initialise properly.

[whitebird]

CardReset is a software remove-insert to let the drivers initialise properly

.

Yes but it is assumed the Amiga has booted. CardReset avoids mechanical card removal after OS got initialized but cannot solve the reset bug by itself.