Own softwareScanned itemsRetroChallenge

RetroChallenge Summer 2012

In my second ever entry in the RetroChallenge, I'll try to learn more about my Basis 108, a German Apple+ clone. As time permits and I'll hit a hurdle, I will sprinkle in bits of Commodore and possibly other stuff.

Basis 108 tasks

  1. Locate the joystick port
  2. Buy an Apple joystick or adapt an IBM PC joystick.
  3. Investigate the built-in RS232 port.
  4. Use ADTPro to make more floppy disks, e.g. original boot disks and games.
  5. Try to locate a complete ROM set to get both Basis and Apple modes.

Commodore tasks

  1. Finish and test two home built MMC2IEC interfaces.
  2. Build a custom 1541-II and CD32 power supply out of a PCB from an Amiga PSU.
  3. Back up my 6 MB Corvus IEEE hard drive, either through a PET 4032 or CBM 610.
  4. Try to troubleshoot and repair some of my broken PETs.
  5. Do some VIC/C64 game/demo programming.

Process log - Latest progress

June 18, 13:30 (GMT+2)

Project defined and early preparations.

July 1, 22:15 (GMT+2)

I decided to start off with something "easy", i.e. finishing up the two home built MMC2IEC interfaces for Commodore computers. However it turned out to not be as easy as I had expected.

To make a long story short, these interfaces utilize an Atmel ATMEGA644 or ATMEGA1284. The bootloader code for sd2iec devices comes in Intel HEX format, and my chips are of the old DIL variety. I have access to two different USB powered programmers:

  • MCU Mall GQ-4X only recognizes 644 chips in whatever other package than DIL. This software can read the supplied HEX file but I can't program the chips.
  • Autoelectric MiniPro TL-866CS recognizes both 644 and 1284 chips in DIL-40, but it refuses to recognize Intel HEX files which have a record of the less common "03" variety, used to fuse execution address.

I can convert HEX to bin and load into the MiniPro, but it prevents me from setting execution address. I tried none the less. The MMC2IEC powers on one of its LED's, but never gets as far as searching on the SD card for full firmware to automatically flash.

Perhaps I can download the bootloader sources and relocate to $0000, assuming Atmel code is easy to relocate. Right now I don't see any option, as this particular interface doesn't have any ICSP interface for inline programming.

Update: I tried to program the 1284 with the full firmware instead of the bootloader. It would prevent me from having automatic firmware upgrades in the future, but in the best of situations it would make it work right now. Unfortunately this approach didn't make it work any better.

Tomorrow night I'll probably have a look at the Basis 108, my main target of the challenge.

July 2, 19:50 (GMT+2)

Stiff neck and busy with other things, so my RC event today had to be kept short.

Thus I made one last attempt with the MMC2IEC, trying the universal "read the error channel" approach to see if it is alive at all:

10 OPEN1,9,15
20 INPUT#1,E,E$,T,S

And yes it is alive! However when I try to read a directory, at best it hangs and results in a listing full of garbage. I tried a VIC wedge with my Megacart, and it gives the same result as the Basic program, except the communication hangs before it should return to the READY prompt. I suppose the problem somehow lies in how I programmed the ATMEGA, but I'll have to trouble-shoot it for another day.

Besides, that is a perfectly normal VIC-20 second revision (cost reduced), just that I swapped the worn out black keyboard for a good working white keyboard from a spare/broken C64C. It might look odd, but it works just as well as one would expect.

Tomorrow I'll be even more busy, so next planned update is not until Wednesday night.

July 4, 21:30 (GMT+2)

Not much done tonight, but at least I dismanted the German battleship otherwise known as Basis 108.

Just like expected, I found the empty socket where joysticks connect. So my next task will be to populate socket X9 with a homemade adapter to connect an analog PC joystick.

I found two designs in the Apple FAQ, regular joystick or joystick with autofire. I believe the simpler design without autofire will be enough for me. Capacitors are on their way, resistors probably already are in the drawer, connectors and cabling are also ready.

As a bonus, I include two pictures of my Basis 108. The thumbnails are clickable to get bigger resolution.

My Basis differs from most other machines one can find pictures of by that it has two half-height 5.25" floppy drives on top of eachother and the other full-height slot empty. Most other Basis 108's have two full-height original Apple drives installed. From the picture on the inside you also see how much space there is.

July 5, 21:40 (GMT+2)

Tonight I brought out the soldering station!

To start with, I had to measure that those blue, grey, brown resistors I found in the surplus box really were 680 ohm. The cables to my multimeter have been patched and kept falling apart, so my first soldering task was to put them back together.

Once that was done, it was time to find some cabling with enough wires, trim the ends and solder to a small spare board. Since most signals are meant to be passed through, I chose a board with continuous copper traces. My capacitors have not yet arrived, but at least I got to solder the two resistors in place.

One of the soldering tasks I like the least is to solder wires onto connectors. In this case, I had a female DB15 (unfortunately meant for board mount) and a DIP-16 socket in the other end. The socket was easier to solder, although I must admit I haven't yet checked continuity from the pin to the board.

I made a test mount inside the Basis. This construction will fit, although the DB15 part of the cable will hang out through one of the empty DB25 holes. Perhaps the mount can be improved by somehow fastening the DB15 or even go as far as use a different, regular type of connector. Since I don't have the caps in place yet, I haven't powered it on.

July 10, 22:30 (GMT+2)

After four days of doing nothing (or having other things to do), I soldered in the capacitors tonight.

It was a partial success. I tried two games and two different analog PC joysticks. One joystick reacts to up/down and sometimes fire. The other only reacts to up/down. I will have to run a test program to see exactly how well they work, and probably plug in either joystick to a PC to verify that it should work at all.

In the mean time, I have some other programming projects going on. More about those later on, although they'd be diversions from my main RC targets.

July 12, 23:45 (GMT+2)

Tonight I triple checked the DIP-16 pinout and then typed in this very simple Apple Basic test program:

10 PRINT"X: ";PDL(0);" Y: ";PDL(1);
15 PRINT" B1: ";PEEK(-16287);
17 PRINT" B2: ";PEEK(-16286)
20 GOTO 10

I tried both joysticks, and came up with these results:

  • The Logitech Wingman Light (to the left) actually works 100% when it comes to left-right and up-down. It centers fairly well to the middle of the value range. However neither of the buttons register.
  • The no-name black arcade joystick (to the right) is stuck to the extreme left. At best I can make it register a little bit of right movement. Up/down is quite OK, although it doesn't center very well. However the two buttons register in a pulsating pattern as long as autofire is disabled (or perhaps enabled, I don't understand what "ON" and "OFF" means as printed on the toggle button).

That means I actually have two analog PC joysticks that both have one or more issues, but the combined tests show the adapter circuit and reading the joysticks actually works! Now I need to find some cheap PC joystick on which both X and Y axises work as well as at least one or ideally both buttons.

I understand those PC gamepads, of which I've got three, won't work since they utilize digital mode on the PC, while I'm strictly looking for analog action.

July 14, 17:30 (GMT+2)

As a brief interlude, I brought out the PCB from an Amiga 500 power supply I bought off eBay for 1 a while ago.

The idea was to see if it (1) works and (2) can be adapted for other uses. In theory it should output +5V @ 2.5 - 4.5A, +12V @ 1.0A and -12V @ 0.1A.

In practise I get an alarming ticking noise, which I think is a sign of potential failure, so I'll probably dump the board and rather get some other cheap power supply that I can trust. I didn't even bother to measure the output voltages.

Update 19:00: I got curious, and measured the outputs anyway. They seem to be +5.25V, +13.80V and -13.80V without load. That is probably within spec for a supply that is rated at 5V + 12V at use. The ticking noise still makes me careful though.

Update 20:45: So I ordered a set of 1x 47uF + 5x 4700 uF caps from eBay. That is $4.38 including shipping, and perhaps can be worth the money if it makes me certain the board can work (inside a proper casing of course).

July 19, 00:10 (GMT+2)

Very little done on the challenge. I played a little with 6502 machine code the other day, and last night I fastened the DB15 joystick port to an unused slot on my Basis.

In the mean time, I have ordered parts to a brand new PC, but that is rather outside the scope of this challenge. Or maybe it isn't, as I'm planning to reuse an empty Commodore 8050 floppy drive case for the PC. It might not be built during July, but I might keep using this page for a build update even past the Retrochallenge.

July 25, 23:25 (GMT+2)

Still waiting for a few power supply caps. A spare PC joystick for the A2 adapter arrived today, will test it tomorrow.

When it comes to the new PC, most parts have arrived. I have temporarily fitted the board inside the case, and so far it looks good. I haven't decided which kind of power supply to use or how to secure it, but that will be solved in the next couple of days.

For those interested in seeing more pictures, you can find an album here with brief comments in Swedish.

When it comes to more Retrochallenge related events, I should investigate the RS232 on the Basis as well as having a go about backing up the Corvus drive. I'm also doing some programming, but nothing so far to publish on this log.

I'll see how much more action I'm getting out of the challenge, but will likely finish it up at the end of July or beginning of August even though that is past the challenge deadline.

July 26, 10:00 (GMT+2)

The new PC joystick works fine, except for fire buttons still operate in a pulsating on-off pattern with autofire in one position, and nothing at all in the other position. Perhaps I ought to adapt the circuit according to the slightly more complex autofire solution? Or get yet another PC joystick that doesn't have any autofire functionality at all, of course.

July 30, 21:45 (GMT+2)

New caps arrived today!

I replaced the five big ones, and discovered the board has another four smaller electrolytes, not to mention all the other components. Too bad I didn't order replacements for all of them at the same time!

However, I'm still getting the ticking sound, and it appears to origin from the big grey part with a transparent top... well, is that a transformer? In any case, all real Amiga people are moving away from the original PSU's and use better quality PC AT/ATX supplies, and it might be a sign it is on the end of its life, new caps or not. Thus I suppose I shouldn't bother more about it, perhaps use as a 5V + 12V source for less critical applications but then I need to place it in some protective box anyway.

One more day of the challenge, and I'm back at work which means I get home later in the evening and have other things to do, so I suppose I'll update this page whenever I have time (past the challenge deadline) just to wrap whatever is left to wrap up. The "AMD 8050 PC" build will of course happen during August, but perhaps I'll stick with updating the photo gallery on that one.

In any case, the Retrochallenge 2012 has been a fun event, and my attention lasted much longer than it did in the Winter Warmup 2010. Perhaps if I'm doing a bit of preparation and planning, I can enter again next time around with one or more coherent projects, even something involving "research".