wynkoop@esquire.UUCP (Brett Wynkoop) (04/28/88)
Greetings folks-
I know I have seen this someplace before, but I can not remember where
or what the hack is. First what is the hack for changing the adddress of
the rs232 pack or the modem pack to be /t3 or I guess /m2 so I can run two
of them. Second thing what is the irq fix. I mean which pin to which pin.
I do not have any data on the 232pack or I could figure that one out since
I know it must be tied to the 6809 irq line.
Thanks folks
-Brett
..........!rutgetrs!cmcl2!esquire!wynkoop
OS9 seven steps ahead of OS/2pete@wlbr.EATON.COM (Pete Lyall) (05/04/88)
> >Greetings folks- > I know I have seen this someplace before, but I can not remember where >or what the hack is. First what is the hack for changing the adddress of >the rs232 pack or the modem pack to be /t3 or I guess /m2 so I can run two >of them. Second thing what is the irq fix. I mean which pin to which pin. > I do not have any data on the 232pack or I could figure that one out since >I know it must be tied to the 6809 irq line. The simplest of the readdressing hacks is one that was spotted on DELPHI. I have seen quite a number, but this one appears to have been the most straigtforward. I dearly hope the author doesn't mind my reposting it here: ========================================================================== CONVERTING THE DELUXE RS-232 PAK TO MODEM PAK ADDRESSING With information being circulated on converting the Modem Pak to work using Deluxe RS-232 Pak addressing, an obvious question to ask would be "why would any sane person want to do something like that?" In my case, it's a simple answer - I don't like to swap cables. Being an active Packet Radio operator on the ham bands, joining the ranks of landline modem users gives me TWO RS-232 devices to play with. The terminal program I have been using on packet is DATAPACK III+, with a CoCoIII, Multi-Pack, and two Deluxe RS-232 Paks. DATAPACK III+ supports either RS-232 Pak or Modem Pak addresses; by converting one RS-232 Pak to Modem Pak addressing, I can change devices under software control, and with the RS-232 Paks being sold for around $30 each, the cost of having two of them in the Multi-Pak interface was minimal. A few disclaimers before you tear into your Pak: 1) Converting the addressing on the I/O chip will render the internal ROM Program useless. It was pretty close to useless anyway, so no loss... 2) My RS-232 Pak has the numbers 8709313 REV A on the bottom side of the PC Board. There may be other boards out there, so this mod may not apply to all Paks. 3) This information is being provided 'as is' - the mod works great in my system, but I cannot guarantee that it will work anywhere else. The basis for its working is technically sound, but variables such as other devices on an individual system, variations in PC board layout, etc., could present problems in a few applications. 4) This modification requires alteration of the PC board in the Pak; doing the mod will void any warranty on the Pak. The I/O chip in the Pak is memory mapped; that is, address decoding is done from the CPU buss rather than by the I/O Select signal on the cartridge connector. This means that (1) the address can be changed at will and (2) the device does not need to be selected by the Multi-Pak. The normal address range of the Pak is $FF68 through $FF6B. The Modem Pak address scheme requires that these addresses be changed to $FF6C through $FF6F. In binary, the original address scheme: 1111 1111 0110 10xx must be changed to be: 1111 1111 0110 11xx Fortuately, this is an easy task, as only one bit (which was required to be 0) must be changed. The original '0' bit was decoded by inverting the A2 address line; the 74LS04 at U6 accomplished this. Detecting A2 High rather than Low simply requires bypassing the 74LS04's inversion. To change the address, perform the following steps: 1) Open the RS-232 Pak. This requires removing the screw hidden under the part number on the label (I used a sharp knife to cut the label and expose the screw's phillips head). After removing the screw, carefully pry the top and bottom sections of the Pak apart. 2) Locate U3 (the ROM chip). Remove it, since it's not needed; this will reduce the loading on the buss (if it's soldered in don't bother). 3) Locate U6 (74LS04). On the top side of the board (the chip side), you will see two traces leaving pins 8 and 9. Carefully cut both traces with a sharp knife, and scrape the coating from enough of the remaining trace to allow soldering a jumper. 4) Jumper the two traces as shown in the illustration below: -------| x-- 8 7 7 -------| x-- 9 4 6 jumper here ^ ^ 10 L 5 | 11 S 4 cut here | 12 0 3 13 4 2 14 1 ^ Notch or dot indicates pin 1 5) Carefully reassemble the Pak. 6) Re-install the Pak and configure your terminal software for the Modem Pak 7) Have Fun! Uploaded to Delphi CoCo Sig by Paul Schmidt (PAULK9PS), Owensburg, Indiana Amateur Packet Mail : K9PS @ WA8YVR (Bloomington, IN) via any W0RLI BBS ========================================================================== As far as the infamous IRQ hack, its really pretty simple (although I'm not near aschematic at the moment...). All that is required is that the IRQ line coming out of the 6551 UART be connected via some path to the IRQ line on the 6809. What we have done in the past is something like this: In the RS-232 pak ================= Open up the pak. You'll find a small blue jumper wire near the edge pin connectors on the top side. One end of this wire connects to a solder pad marked either '1' or 'E1' (again - this is from memory). The other side of this jumper wire connects to another solder pad near pin 7 (top side) which passes through the board and then out on pin 8 (the CART line, or the *FIRQ line into the CC3). Disconnect this jumper either by desoldering or clipping (former preferable). Then, attach a long (at least 1-1.5 feet) wire to the solder pad marked '1'. Hang this out of the side of the RS-232 pak (I burned a small detent with the soldering iron), and attach a subminiature male phone plug to the end of it. We're 50% done. In the COCO3 ============ Open the coco3 (standard disclaimers about warrantee apply here...). Locate the pullup resistor that is tied to the CPU's pin 3 (IRQ line). It is in the small group of resistors that lie to the upper right of the CPU as viewed from the keyboard. I believe it is R2 and I believe it is the second resistor from the left, but it would be best if you were to either confirm this with the CC3 schematic if you have one, or simply by verifying it with an ohmmeter between pin 3 of the 6809 and the side of the resistor facing the keyboard. Should be a dead short to that end of the resistor. Once you have confirmed that you have the correct resistor, solder one end of a wire to the side closest to the keyboard (I suppose a secure spring-loaded clip would also suffice for those that may be solder leary). Attach a female subminiature phone jack to the other end. You may choose to either hang this line out of the back of the CC3, or mount it on the rear of the case (I did the latter). At this point you should verify that you have continuity from the RS-232 pak to the CPU. Plug the subminiature jack/plug together, and take an ohmmeter reading from the solder pad marked '1' to the 6809's pin 3. If you have continuity, you're home free As far as the OS9/LII ACIAPAK driver goes, there are absolutely NO changes required to support this hack. Optionally, you may wish to enlarge the driver's per-line buffer size to 256 in/256 out (I'll post a patch later) - this will improve throughput at higher baud rates. For those that have been experiencing dropped characters, communications hangs, and other serially related trauma, this should pretty much clear things up. It has been in use for years on CC2's, and has been installed by dozens of (now) happy CC3 owners. Please don't attempt it if you are concerned about your warrantee or are not fully confident in your soldering abilities. I assume no responsibility or liability for the installation or consequences of the installtion of this hack. That's up to the individual. Hope you find it useful! -- Pete Lyall (OS9 Users Group VP)| DELPHI: OS9UGVP | Eaton Corp.(818)-706-5693 Compuserve: 76703,4230 (OS9 Sysop) OS9 (home): (805)-985-0632 (24hr./1200 baud) Internet: pete@wlbr.eaton.com UUCP: {ihnp4,scgvax,jplgodo,voder}!wlbr!pete