• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

Ricoh chipset PCI to PCMCIA adapter card - possible?

Byrd

Well-known member
Hi,

I’ve a TAM with a USB PCI card and no CS slot riser = no Ethernet. I’m using a USB SMC Ethernet adapter with OS 9 drivers instead.

I’d really like FireWire, and combo cards are hard to come across - so what got me thinking is a Ricoh chipset PCI to PCMCIA adapter card, which is easily found. There I could have two PCMCIA cards, like a dual FW/USB card and Ethernet. Pull out the card for a CF adapter, or wifi.

Any chance such a card would work out of the box, has anyone ever trialled these cards in a PCI Mac?

Thanks

JB

 

Trash80toHP_Mini

NIGHT STALKER
Does the standard TAM riser differ from the 5xxx/6360 riser? Wondering if the Fatback riser is different than the 6400/6500 two slot riser?

Next tangent: That oft missing Fatback thingie.  What size it it and how difficullt might it be to print a three slot version?

- CommSlot is a rudimentary PCI slot, so I'm wondering if a dumb CS riser could be connected to the standard bus on a custom PCI riser?

- The interrupt/ID setup is there for your PCMCIA slot card, a standard NIC or possibly FireWire or Fast HDD interfaces?

 
Last edited by a moderator:

trag

Well-known member
The Umax C600, based on the 6400 architecture, has a 3 slot riser, and I think it also has the Comm Slot, suggesting that a four PCI slot riser should be possible.  I last looked into this several years ago, so I might be wrong about it having a full comm slot, but I think it does.  Okay, ran off and checked Everymac real quick and it says the C600 had three PCI slots plus a Comm-2 slot.   But it also mentions that the Comm-2 slot is occupied by the internal modem.    My concern is that IIRC, the Comm-2 slot is a slot with PCI and USB wires in it and it uses whichever depending on what's installed in the slot.   It's conceivable that teh Umax version dispenses with the PCI component.   But then it wouldn't work right if someone tried to remove the modem and install a PCI based device, so I don't think Umax would have done that, probably.  

 

EvilCapitalist

Well-known member
I think someone here tried using the two slot PCI riser from a 64xx/65xx machine and it worked, but it introduces new problems:

To use the two slot riser the hard drive bracket / cooling fan has to be removed.  With the standard riser the installed PCI card is resting on the HD bracket so installing the two slot riser necessitates removing that if you're trying to use the lower PCI slot and then I still think there are clearance issues with the top slot blocking the CSII riser, making you worse off than when you started (single PCI slot, no HD bracket, no CSII)

- When the PCI riser has a card installed the weight of the two (PCI riser and PCI card) is supported by the fat back ports bracket and the HD bracket.  A two PCI slot riser wouldn't work with that (the offset is wrong) so it would be unsupported and putting a fair amount of stress on the PCI slot on the motherboard

- When the whole shebang is installed it's too tall to reinstall the fat back (I think) so you'd have to run with the cover off

 

Trash80toHP_Mini

NIGHT STALKER
So the Fatback is needed for a SINGLE PCI slot and the CSII riser/slot???  :p Not a TAM guy at all, I was under the impression that it allowed for TWO PCI cards to be installed? Silly me.

Definitely need to come up with a design for a printed Fatterback and a three PCI slot riser. Header interconnect between it and a custom CSII riser with cross bracing should be plenty of support for three PCI cards.

The mass storage bottleneck just KILLS performance of 6400/6500 architectures. Haven't tried my OS9 bootable SATA card in a 6500 board as yet, but that's my workaround for the TEMPOtrio's incompatibility with Gossamer board's I've collected.

C600 is 6400/6360 Alchemy architecture, so TAM compatibility with three slots is just as much in question as the PCMCIA adapter card in the OP. Pretty sure that's a CardBus adapter and its 10.2.8 minimum OS requirement doesn't sound very promising for OS9 compatibility.

 

EvilCapitalist

Well-known member
Here we go, pics-a-plenty:

Comm Slot II riser / PCI riser / PCI riser from a 5500 for comparison

20180709_202322.jpg

Detail of the CSII riser

20180709_202401.jpg

CSII back

20180709_202414.jpg

PCI riser front

20180709_202425.jpg

PCI riser back

20180709_202457.jpg

 

EvilCapitalist

Well-known member
You can see the offsets in both the CSII riser and the PCI riser and why the two slot card from the 64xx/65xx series wouldn't work (pics of those below for comparison), and on a side note, I didn't realize that there was a difference between the 6400 riser and the 6500 riser but there definitely is.

6400 riser

6400 riser.jpg

6500 riser

6500 riser.jpg

 

Trash80toHP_Mini

NIGHT STALKER
The Umax C600, based on the 6400 architecture, has a 3 slot riser, and I think it also has the Comm Slot, suggesting that a four PCI slot riser should be possible.
Comm Slot is either/or. The 5500/6500 DevNote has pinouts for each implementation:

1 - Pin assignments for a universal serial modem card - the vast majority of which are n.c.

2 - Pin assignments for the PCI bus communications slot connector - wasn't expecting to find this pinout at all!  [:D]

It should be easy enough to determine the lines which identify the second slot of the 6400/6500 riser. Cut those and patch them to the appropriate signals on COM Slot 2 to determine if three slots will work. Heck, you could test it with a pair of cheap single slot generic risers.

Sounds to me like UMAX didn't implement the PCI side of things on the CS2 connector, opting for a third full-on PCI Slot and dumbing the CS2 connector down universal serial modem card connections only?

Can't find an exploded diagram of the TAM. It looks to me like you need the fat back and riser to use CS2 at all, whether you have a PCI card/riser or not? Looks like it flops the NIC over atop the single PCI card?

 

EvilCapitalist

Well-known member
Can't find an exploded diagram of the TAM.
I don't think there is one.  The Service Source manual essentially just called it a closed system and left it at that, though someone was good enough to do a teardown of their TAM and post it on iFixit.

It looks to me like you need the fat back and riser to use CS2 at all, whether you have a PCI card/riser or not? Looks like it flops the NIC over atop the single PCI card?
You've got it.  The ports bracket on the right side (if you're looking at the back of the machine) supports whatever cards are installed so they don't just flop down and then that bracket is supported by the fat back itself.  You can't use either slot (PCI or CSII) without the risers as whatever cards you'd install would hit the sides of the case and won't go in all the way.

This is borrowed from that TAM teardown but it shows how the cards are right on top of each other.

topless tam.jpg

Here's a side view borrowed from the Mac512 website (which is presently dead):

TAMFatandMiniCovers.jpg

You can see how the ports bracket (the recessed part) holds the PCI/CSII cards and then the whole fat back supports that.  The fat back definitely ruins the lines of the machine but without it you're limited to SCSI for expansion, LocalTalk for networking, and you can't install any CPU upgrades.

 

Trash80toHP_Mini

NIGHT STALKER
Thanks, very interesting. Since the lines are already "ruined" where you really can't see them anyway, making a back that's about an inch more chubby really wouldn't matter. If it gave you room for a Gigabit NIC and a much faster ATA-whatever or even a SATA interface to blow Gossamer's most serious bottleneck wide open for your SSD it'd be well worth the effort. Print it in clear and it's a display window feature! [}:)]

Back to the OP: isn't there room inside the case to mount the CS2 NIC vertically? Might need to lose the RJ-45 and a little of the PCB there, but wiring it up to the VONETS WiFi bridge ought to work nicely. Don't have a pic of the ports section, but it ought to be easy enough to route the antenna out somewhere there?

Back to the tangent: will the stock back fit with the CS2 NIC mounted vertically? With the Fat Back installed, would a twin slot PCI riser fit when the baseline NIC is vertical? I take it there are two cover plates, one with and one without the PCI backplane plate cutout?

 

trag

Well-known member
Well I went to Kennedy Brandt's SuperMac Insider site and downloaded the User Manual for the C600:

http://www.kennedybrandt.com/supermac_insider/support/downloads/C600UserGuide.zip

Kennedy is a fomer Umax employee who really went all out supporting the customers.   After Umax dropped clone support he put up his site to still provide some documentation.

Anyway, the manual calls it a "Comm slot" but it also says that one can install a modem or a network card there. 

Anyone know if the Comm Slot 2 ethernet cards are USB based or PCI based?   If they are PCI based that would answer the question.   If they're not, the issue is still ambiguous.   The manual does not say the slot is identical to Apple's Comm Slot 2.   It does give a very specific diagram of cards with notiches to show what kind of card fits in the slot.  I am guess that the "doesn't fit" illustration was of a Comm Slot 1 card.

One of these days, I'll dig out the C600 and probe the riser card.  By that time someone will already be building 4 PCI slot risers for the 6X00 families.  :)

 
Last edited by a moderator:

Byrd

Well-known member
Thanks for posting info - on the PCMCIA card, and CS II slot riser :)

I can't help thinking that one day I'll make up my own adapter to fit the CS II riser in the TAM, but it's not essential.  I use an SMC USB Ethernet adapter at present, which came with OS 9 drivers.

Re: PCI --> PCMCIA card, when I next see a cheap one I'll snag it.  With Cardbus being PCI compliant, wouldn't it just work out the box - Powerbooks for example have no PCMCIA drivers as such to run?

 

Trash80toHP_Mini

NIGHT STALKER
Cool, I'm assuming you have the Fat Back for your USB card or you wouldn't have asked the original question?

Have you got a CS2 NIC to try installing it directly into the slot with the Fat Back providing vertical clearance? Now that Ants has WiFi'd everything else, you should try it with your TAM. With no need for a wire, it wouldn't matter thst the port is buried down in the bowels of that wonderful kluge. [:D]

Anyone know if the Comm Slot 2 ethernet cards are USB based or PCI based?   If they are PCI based that would answer the question.   If they're not, the issue is still ambiguous.   The manual does not say the slot is identical to Apple's Comm Slot 2.   It does give a very specific diagram of cards with notiches to show what kind of card fits in the slot.  I am guess that the "doesn't fit" illustration was of a Comm Slot 1 card.
Trag: It's undoubtedly the standard 6360/6400 CS2:

5500/6500 DevNote:

The PCI Bus Communications Slot 4

The main logic board has a separate slot for an optional communications card. The
communications slot on the main logic board is a PCI bus–based communications slot
(comm slot II) rather than a processor direct PDS–based communications slot (comm
slot I) like that found on the Power Macintosh 5200 and 6200 computers.
The electrical interface of the communications slot is a parallel bus, the SCC lines, and
lines for supporting modem audio. The PSX+ custom IC provides bus conversion from
the host PowerPC 603e bus to the PCI parallel bus. Cards that use the communications
slot are memory mapped into the I/O space of the main logic board via the parallel bus.
The communications slot supports SCC port A (modem port) for installing a universal
modem card that is compatible with both the communications slot in the Power
Macintosh 5200 and 6200 computers and the PCI bus communications slot.

PCI Bus Communications Slot Connector 4

The PCI bus communications slot connector is a 112-pin half-height microchannel
connector. A communications card mounts vertically in the connector and its I/O
connector is accessed through the communications port access hole on the right hand
side of the back panel. The size constraints of a communications card are 1.57 inches (40
mm) wide by 6 inches (152 mm) long.

A maximum of 2.5 watts of power is allocated to the communications slot. The
maximum possible current ratings for each power line are
Voltage Current
+5 V 500 mA
+12 V 100 mA
Trickle +5 V 5 mA
–5 V 20 mA


Table 4-12 Pin assignments for the PCI bus communications slot connector

Table 4-12 lists the pin assignments of the PCI bus communications slot.

1 /DCD
2 /DTR
3 /CTS
4 /RTS
5 RxD
6 TxD
7 IN_SENSE
8 SCC_ENAB
9 INT_MIC
10 MIC_SENSE
11 MIC_RET 12 EXT_AUD_L
13 Reserved
14 EXT_AUD_RET
15 GND
16 +12V
17 -5V
18 +12V
19 SYS_WAKEUP
20 Trickle +5
21 GND
22 GND
23 A1
24 A0
25 A3
26 A2
27 +3.3V
28 +3.3V
29 A5
30 A4
31 A7
32 A6
33 +5V
34 +5V
35 A8
36 C/BE(0)~
37 A10
38 A9
39 GND
40 GND
41 A12
42 A11
43 A14
44 A13
45 C/BE(1)~
46 A15
47 GND
48 Gnd
49 SERR~
50 PAR
51 PERR~
52 SBO~
53 LOCK~
54 SDONE
55 +3.3V
56 +3.3V
57 DEVSEL~
58 STOP~
59 IRDY~
60 TRDY~
61 +5V
62 +5V
63 C/BE(2)~
64 FRAME~
65 A17
66 A16
67 GND
68 GND
69 A19
70 A18
71 A21
72 A20
72 A23
74 A22
75 GND
76 GND
77 C/BE(3)~
78 IDSEL
79 A25
80 A24
81 A27
82 A26
83 +3.3V
84 +3.3V
85 A29
86 A28
87 A31
88 A30
89 +5V
90 +5V
91 REQ~
92 GNT~
93 +5V
94 +5V
95 INT~
96 Reserved
97 Reserved
98 RST~
99 GND
100 Reserved‘
101 CLK
102 Reserved
103 GND
104 Reserved
105 Reserved
106 Reserved
107 Reserved
108 Reserved
109 CommGnd
110 RefGnd
111 AudToSlot
112 AudFromSlot

Looks like there would be 2,5 watts from the CS2 budget for a Gigabit NIC installed in a two or three slot riser.

 
Last edited by a moderator:

Byrd

Well-known member
Yes, have Fatback only on my TAM + very basic USB 1.1 card (it's a princess of a computer, no other USB cards I've tried work).

 

Trash80toHP_Mini

NIGHT STALKER
Curiosity got the best of me again.  :blink:   Couldn't find info on these three signals:

52  SBO~. . . . . . . . . .  
54  SDONE . . . . . . . .  
58  STOP~. . . . . . . . .

They're all n.c. on the serial modem card's pinout so they're on the PCI side of things, but I couldn't find a clear equivalent in PCI signal listings. Info on the rest is from: http://www.interfacebus.com/Design_PCI_Pinout.html which I checked against the DevNote - PowerMac 5500-6500.pdf

 49  SERR~. . . . . . . . . [Open Drain] System Error is for reporting address parity errors, data parity errors on the

                                 Special Cycle command, or any other system error.

 50  PAR . . . . . . . . . .  [Tri-state] Parity is even parity across AD[31::00] and C/BE[3::0]#. Parity generation is required by all PCI agents.

                                 PAR is stable and valid one clock after each address phase.
 51  PERR~. . . . . . . . .  [Sustained Tri-State] Parity Error is only for the reporting of data parity errors. Two Parity lines are made available,

                                  one for the 32 bit bus width (bits 0 to 31) and an additional one for the 64 bit expansion (bits 32 to 63).

                                  Two error bits; I assume, 1 for the LSB 32 bits and one for the upper 32 bits.
 53  LOCK~. . . . . . . .  [Sustained Tri-State] Lock indicates an atomic operation to a bridge that may require multiple transactions to complete.

                                  When LOCK# is asserted, non-exclusive transactions may proceed to a bridge that is not currently locked.
 
 57  DEVSEL~. . . . . . . [Sustained Tri-State] Device Select, when actively driven, indicates the driving device has

                                  decoded its address as the target of the current access.
 
 59  IRDY~. . . . . . . . . [Sustained Tri-State] Initiator Ready indicates the initiating agent's (bus master's) ability to complete the

                                  current data phase of the transaction. IRDY# is used in conjunction with TRDY#.
 60  TRDY~. . . . . . . . . [Sustained Tri-State] Target Ready indicates the target agent's (selected device's) ability to complete the

                                 current data phase of the transaction. TRDY# is used in conjunction with IRDY#.
 64  FRAME~. . . . . . . . [Sustained Tri-State] Cycle Frame is driven by the current master to indicate the beginning and duration of an access.
 
 36  C/BE(0)~
 45  C/BE(1)~
 63  C/BE(2)~
 77  C/BE(3)~ . . . . . . . [Tri-state] Bus Command and Byte Enables are multiplexed on the same PCI pins.

                                  During the address phase of a transaction, C/BE[3::0]# define the bus command.

 78  IDSEL. . . . . . . . . . [input] Initialization Device Select is used as a chip select during configuration read and write transactions.
 91  REQ~. . . . . . . . . . [Tri-state] Request indicates to the arbiter that this agent desires use of the bus.

                                 This is a point-to-point signal. Every master has its own REQ# line.
 92  GNT~. . . . . . . . . . GNT# [Tri-state] Grant indicates to the agent that access to the bus has been granted.

                                 The GNT line is a point-to-point signal, each master has its own GNT# signal.
 95  INT~. . . . . . . . . . . [Open Drain] Interrupts on PCI are optional and are "level sensitive," asserted low (negative true).   --- INTA#, INTB#, INTC#, INTD#
 98  RST~. . . . . . . . . .  [input] Reset; used to bring registers and signals to a known state
101 CLK. . . . . . . . . . .  Clock; rising edge provides timing for all transactions

32-bit Address/Data Bus: Time Multiplexed on PCI Bus, The DevNote doesn't have signal definitions.

23 A1              24 A0     
25 A3              26 A2
29 A5              30 A4
31 A7              32 A6
35 A8              37 A10
38 A9              41 A12
42 A11            43 A14
44 A13            46 A15
65 A17            66 A16
69 A19            70 A18
71 A21            72 A20
72 A23            74 A22
79 A25            80 A24
81 A27            82 A26
85 A29            86 A28
87 A31            88 A30

These are on the modem pinout along with a lot of sound and microphone oriented junk.

   1  /DCD . . . . . . . . . . Modem
   2  /DTR . . . . . . . . . . Modem
   3  /CTS . . . . . . . . . . Modem
   4  /RTS. . . . . . . . . . . Modem
   5  RxD. . . . . . . . . . . . Modem
   6  TxD. . . . . . . . . . . . Modem
   7  IN_SENSE. . . . . . . Modem
   8  SCC_ENAB. . . . . .  Modem

I'm wondering if anything at all needs to be jumpered from CS2 for adding a third PCI slot? I've got a three slot riser that works in my Beige G3 board, looks like it's time to test it in one of the 6500 boards? Dunno, whatcha think, trag?

 

EvilCapitalist

Well-known member
I have to imagine it wouldn't be terribly expensive to have a new run of CSII risers done.  It appears to be just a dumb extender with no logic on the riser itself and there's certainly a market for them given how fast they tend to go if they pop up for sale.

 
Top