Amiga Cleanup 2019

I wanted to play with my old A1200 but was worried about its condition and had to check...

Here I have opened my A1200 up and removed shield... caps looked okay but can be bad nontheless. Don't feel like recapping SMD though. I checked headroom on Alice with soft clay... not much, but I think an offset heat sink might fit. 68K CPU is tiny. There's another faster one on my Blizzard 1220/4.

The A1200 gets rather bit hot shielded. Might be the enclosure. I'm worried that the PSU brick voltages might be off... they can be a bit tricky to measure without a load however.

Opened up my Amiga 1200 PSU (cost reduced ver. afaik). Looked fine, except for the components floating in a 2mm deep sea of dust and hair from a cat long dead (T_T) I improvised a 5V line dummy load (0.56A, Amiga draws 2A though?). Measured 4.725V (low for TTL?) and 13.4V (ok for some FDD motors?). The A500 PSU is beefier but I haven't used mine in ages.

I didn't find anything which can dissipate 10 Watts (like an actual 2.5 Ohm power resistor, or bulb) so I put ten 0.5 Watt 100 Ohm resistors in parallel for a lighter load. They climbed to 70 degrees C pretty quickly. I insulated the crocodile clips with paper sleves to avoid accidental shorts.

But uh-oh! Spill or leaky cap in floppy drive >:G Went over it lightly with a brush and a SMD ceramic capacitor fell out >_< Meaning, I'll have to disassemble the whole drive, reheat some joints and probably recap some of the electrolytics... which I fortunately have in stock. Not SMD ceramics though. Oh. Looks like it can't be disassembled... so I can get to the PCB. It's too sturdy. Will attempt repair anyways. Some white vinegar + isoprop cleanup, then reattach the 400nF (says multimeter) ceramic cap. If toast... maybe Gotek heresy is the only solution.

This took 6 hours of cursing, and it was just two components. It seemed no amount of vinegar, isoprop, scraping, flux, solder tip changes, temperature adjustments could make those pads take solder. And the ecap had GND facing out so that meant massive ground plane. I ended up just twisting it off >_<

Also had to redo the component positioning four times as the joints broke. The disk slams down pretty hard and the shock could pop the solder joints. Ended up just tying GND legs together. Actually had no SMD 10uF 35V in stock, only a 25V, so I used a through-hole 50V instead (but maybe 2x12V would be enough?). Ceramic one was 390nF buy I put my bets on a 330 replacement.

Pixel-over of one of my shmup projects from '92. IIRC it was kinda playable with 3 levels and powerups. It heavily used BOBs but I think I could make a much better engine now than young me. Actually made one with some infinity scrolling but as usual I got bored after it worked.

She's alive. Alice seems to be the hottest. 50 degrees C open case with one offset heat sink due to poor headroom. Probably better temps than before though. I removed the shielding. HDD sits on standoffs made from cardboard so I better not rattle the machine around. I think I later added some smore heatsinks after measuring temps on various ICs with my gun thingy. I'm using my Video port -> SCART adapter here. Very nice image compared to composite (yellow RCA). TV actually flickerfixes Hires-Laced.

It was fortunate that I decided to put a DE-9 connector with "raw" switch output on the back of my PCE joypad project. I was able to make an adapter for the Amiga in minutes using a spare Tac-2 cord.

Finding tons of these little AMOS games I made in the very early '90s (it's hard to date files due to real time clock and fily copying issues). This appears to be a barebones and dull key-maze game that I compiled as a standalone along with a maze editor.

And that is indeed a raunchy HAM image of Cindy Crawford. Second place favourite after Diana from V. Compared to MS DOS garbage systems the Amiga & its OS was quite amazing.

Also, found some of my totally cool 1990 pixel art in 32 mostly wasted colours. Star shine ramp got more colour indices than the space slimes! I think it could be reduced down to 8 colours quite easily.

Managed to get AMiGA CrossDOS™ working with my USB FDD (on a Mac), and transfer files that way. Supposedly MF-2HD disks doesn't work as well as MF-2DD, but I still got CRC errors. A freshly formatted disk worked better. Squeezed in a 5cm label on my sheets, yielding 12 per A4.

Some tests in AMOS Pro. Here it seems long variable names does not cause much slowdown. Might be tokenized. Pretty much everything snaps to certain formatting. Makes me sad tokenization is not used for colour coding/legibility. REM statements do slow down things quite a bit!

I only have one proper disk box and this is not it. I don't know if cardboard outgassing/acids can somehow damage disks but I keep a bunch in a cardboard box. I made some separators. Also found some dev utiities I never understood as a kid. Boot & Disk tools, monitors, Seka-assembler, crunchers...

Repaired another Amiga Mouse, replacing the damaged old switches. The BA10339 IC has 4 comparators... I believe they were fairly common in old mice and used to clean up (square-wave-ify) the signals from the detector discs. And that's it. No serial IC in the Amiga mouse.

Unrelated, I thought about about making a (very silly) data-transfer device that uses the joystick port (rather than serial or parallel)... maybe using PC817 opto-couplers/isolators as switches, and an Arduino/MCU waggling the "joystick" to transfer bits to a BASIC program (then RAM disk) running on the Amiga. But rates would be very slow... likely several minutes for 10kB. Fire would act as data clock. I wanted to have isolation to make it safe and authentic. Amiga joysticks aren't powered by 5V, though it could probably be used to power an Arduino with pins in psudo-tristate mode. Amiga joystick switches just provide a path to GND (or not). With PC817s there'll be some minor resistance in the switches I suppose. I think the lines are pulled up to 5V via 4.7Ks inside the Amiga.

DE-9 connectors & cables are easy to get ahold off though, and the method could work for getting required files onto e.g. an A500 without crossdos or any working serial utilities. A DE-9 connector can sort of be soldered onto the edge of a prototype board. The pin rows are offset but here I've used Kapton tape to mask off pin 2-4.

The Amiga mice always felt like they had smoother movement than PC mice, rarely making nasty staircases. The DPI, rotary encoder wheels & comparators are... comparable, so I think it has something to do with lower resolution masking some of the jitter, and the lack of a PS/2 serial conversion/driver... maybe also polling rate, something special in Denise?

Floppy Disk Labels

I have a few hundred floppy disks and many aren't labeled properly. So, I decided to make some printable labels (using self-adhesive / sticky paper). The images are 500PPI (~6K×4K pixels) and A4, though they might fit on letter sizes. I've saved as GIF to keep file size down (<200kB). Fancy users can convert to CMYK and set up the K channel properly using profiles. Ideally they'd be printed using a single flat/100% colour but that's not how home laser printers work. Printer toner unfortunately tends to wear off a lot easier than real print.

I think having an assortment of styles and colours can assist in disk browsing. Because a label is aproximately 7x7cm it's not possible to fit three of them wide on a single A4 as the printable area is <21cm. The unused space could be used for jam-jar slabels or scraps. Smaller, non-wrap floppy disk labels are 5cm tall so those would fit.

Features: A little Sony-dot Amiga checkmark logo.
Ex Libris line: Disk owner if any.
Date: Project date or game release date. Date in Swedish is Datum.
Use% It can be handy to know how full a disk is. I use pencil here.
Disk name: Name of disk on computer.
Disk: Game disk A/B.
[ ] Archived: I back up some of my disks to HDD or elsewhere.
[ ] Copy: This disk is a backup copy.
[ ] Error: Disk has an error. It's possible to fix some disk errors by gently cleaning the disk surface.
[ ] Bootblock: Disk is bootable.
ID# A number for an external catalog
Lines: I left the center free of lines as I like to scribble larger logos there (to help when browsing). The lines can be used as colums if heed is taken to the small division mark.
[___]   [___]   [___] top/spine slots: These are for grouping game disks. I colour one of these using a marker, and fill in the other two using black.

Nashua disks had a code system on the spine (top edge) and initially I wanted to do something like it for my project disks (art, music, code, images, etc), but discovered that because project disks often fill up with mixed content... strict categorization is not really doable.

But, pirated games usually come on several disks which need to be kept together. Any kind of arbitrary colour mark across their spine can help with that. There's not much point in trying to create a system for genres. A single mark in one of three positions is quite sufficient. I pick a colour I feel represents the game, like orange for Dune II.

K240 label sketch.

I started thinking about simple single-player games, and which one sgives the most bang for the development effort. I think it's Asteroids, because the simple physics apply to all objects and still makes flying and drifting with the ship fun. It could scale into something like Star Control 1 during development too.

Many of the other games are not very fun, or have hidden complexities, such as needing flood-fill code, pathfinding, ai, or edge cases to deal with. Time consuming level design is an issue in puzzle games (Chip's Challenge) or platforming (JSW) games. Blob Attack and Cheese were some of my games that I managed to finish. Demon Attack is doable on systems with limited graphics capabilities. Wizard of Wor is a fun take on a maze game. Boulder Dash is a bit barren as a game, but something like The Pit looks more interesting.


Art by Arne Niklas Jansson

AndroidArts.com