So despite starting several projects, I only got on successful project achieves this past week.
Firstly, I received some ROMs to use with one of my Apple II clones, as it wasn’t received with them installed. Unfortunately the power supply has chosen now to fail so I’ll need to strip it and see what I can work out. Oddly, people on the Apple forums want me to rebuild the power supply rather than just sticking in a suitable Mean Well PSU. Why? It’s not an original power supply anyway. A Mean Well will be cheaper, cleaner, safer and more efficient. I may even just gut the existing PSU and drop the Mean Well directly inside the cage.
EDIT:It’s not the power supply. Every time I install the ROMs, the PSU trips. I’m using 2816s which should work just fine. Clearly there is more to be investigated.
Secondly, after a trip to Jaycar, I had the bits to build a USB to Atari 800xl power adapter. You need a beefy USB Power source, but it works well. I just followed the guide here on Wolfgang Kierdorf’s channel. I made up both ends rather than dissecting an existing USB cable. It came out quite well, and means I don’t need the huge brick on my desk when using the 800 any more.
Finally, I’d found a Geode based single board computer while looking for parts. My housemate had salvaged it and hadn’t found a use for it. I thought it would make for an excellent DOS / Windows 3.1 system. I made up a power connector, but unfortunately it was DOA. No matter what I tried, I couldn’t get it to even send out a synch signal.
It’s not a huge loss. My desire for a 486 class PC is not high, and there are some nifty alternatives if I wanted them.
So what did your kids get you for father’s day? Mine gave me a cold!
I had two days off work, and to be perfectly honest, I probably should have taken at least one more day off as well.
Despite being “Death warmed up”, a friend put me onto something that had been dropped into the “recycling” pile at a local Maker Space. I rugged up, drove out and came back with a boxed Genlock! Awesome!
I did not move for the rest of the week 😀
However I was feeling well enough by Saturday to start a couple of projects, one of which wasn’t even RetroComputing related! Crazy, I know!
First up was a request from a friend of a friend to make a TV Stand for their flatscreen TV. It theoretically has standard VESA mount points, but with an 8Kg screen and a lot of potential leverage, I wanted to lean into the overengineering. I’ve made plenty of monitor stands for LCDs, which are often available practically for free if they don’t have a stand. I’ve got pretty good at it, but this was a whole new level.
I ended up leaning into the additional strengthening and I’m pretty happy with the result. I just hope the end recipient likes it.
Afterwards, it was time for me to work on another project. I’d ordered in a replacement PSU “gubbins” for the Amiga PSU I’d got that had a nasty habit of tripping the RCD. Rather than trying to diagnose, it’s a LOT easier just to gut n rebuild PSUs. Newer technology works more efficiently, is lighter and delivers much cleaner power.
It was easy work to extract the old “guts” from the brick and there was a LOT of space inside there for the new PSU. I didn’t trust the existing switch, so I went to Jaycar and bought a similar sized one with an integrated Neon illuminator. Makes it nice and easy to spot if it’s on or not.
The only problem I had is the PSU is so light I end up nearly throwing it across the room.
Today I spent a bit of time working on more of the Amiga 600. My desoldering station appears to have a problem, and I think I’ve identified where it is, but for now I have no desoldering capabilities other than hand pumps. I was able to desolder the RF converter, but I was unable to get the keyboard connector or the choke out. This means I have only been able to get most of the SMD capacitors out, not all of them.
I used a new technique, involving two soldering irons, that works quite well. You basically heat the cap from both sides at once, and the caps just pop cleanly off with no force required. It’s really gentle on the pads. I always apply extra flux, of course, but so much gentler than any technique I have used or seen, short of hot air. (I am fairly sure I first saw this over on GadgetUK164’s YouTube channel.)
A bit of braid afterwards and the board came up as new.
To do this I needed to get a second soldering iron up and working. A good friend had gifted a TS-100 soldering iron, but I’d had no end of troubles with it. I thought this was a good time to revisit it. I knew that one of the problems was the tip wasn’t fitted properly, but the included allen key was useless, and I couldn’t remove the screw holding the tip in.
I attacked it with my generic “giant box-o-screwdriver tips”, which has most of the common “bits” to open up stuff the manufacturer doesn’t necessarily want you you to. Thankfully, one of the imperial allen tips was a perfect fit, and I was able to refit the tip and replace the screw in short order.
I also updated the firmware. The first time I’d tried this it hadn’t worked, but this time it went on without a hitch. No idea what is different.
Finally I made a stand for it out of some coathanger and a bit of old Ikea bracket designed to hold bookshelves to walls.
For now, the A600 will go back in the project box until I can afford a hot air rework station.
I decided not to push things this weekend. I had some “job” related stuff this weekend so I wanted to take it easy.
I still managed to get in some small project work.
First project I worked on was I’m trying to set up a project to use an Arduino as a translator between Megadrive/Genesis controllers and “Atari” compatible systems like the C64, the Amiga and, of course, the Atari systems. I’ve made up some connectors and tried my hand at a first attempt, which unfortunately did not work. Thankfully some clever people on the Artifactory Slack were able to point me to a solution that I can use involving OptoIsolators, so I have some of those on order.
Secondly I set up the 3D Printer that I recently received from my wife. It’s an Anycubic Photon Zero, and setup was remarkably easy. I need to set up myself with some “extra” bits and pieces, such as a decent mask and some extra Isopropyl, but I did manage a basic test print. Next week, time and budget willing, I will actually print something emulation related. I’ve designed a little “carrier” to hold a rotary encoder and an OLED screen. Should be pretty sweet if I can get it working.
Next I made a modification to the cable I made for the Microbee. It now has an on/off switch, a power LED and some big diodes to drop the voltage down a few notches to keep the main system cooler.
Finally I went back to my old house and went through the shed. I knew there was an Amiga power brick in there, and figured I could use it with my A600. When I got in there I found a whole bunch of other “bits and pieces” including two SCSI CD Roms, 4 SCS HDDs (1 in SCA, 2 in 68 pin and 1 in 50 pin), two floppy drives, two ZIP disk and some assorted cards that seemed useful. Nice!
Oh, during the week I also started the long slog of decapping my Amiga 600. Unfortunately I hit a bit of a roadblock insomuch as I don’t seem to be able to remove the caps without hot air. I’ll give it another try next weekend, but I suspect it’ll have to wait for Christmas and a Hot air rework station.
First up, I am, once again in debt to the Retro Communities to which I belong. The people over at both Microbee Technologies and MSPP (Especially Brad) who provided advice, pointers and general assistance as I got the second of my Microbees working.
Now, the fault I was seeing was instability on boot. There seemed to be an issue where some times the ‘bee would boot into the wrong mode. Most times it would boot into the Menu. Sometimes into BASIC. Sometimes into WordBee. Sometimes it would just “hang” in a strange corrupted mode. The Menu was also only showing two items, not the half dozen it should be showing.
Once it was in BASIC, I could happily boot games and run them without a hitch. I was confused. Usually if a system was unstable, it would crash randomly. This meant it was something very specific that was only involved in choosing which ROM to boot from.
I started troubleshooting things, firstly by reading in the ROM images and comparing them to “known” images. All the ROMS came up clean. (There was a brief detour when I thought the Menu ROM was bad, but this turned out to be not the case. There’s two versions of that motherboard and I was looking at the wrong set of ROMs)
I then cleaned all the pins on all the ROMs. The socket the Menu ROM sat in was green with battery corrosion so I removed it and replaced it with a Machined Pin socket. This improved reliability somewhat, as now the system tended to boot into the Menu.
Around this point I removed the connectors completely between the base board and the core board to see if that helped. The system seemed no better.
I also discovered there was a diode in the area that was growing green inside its tiny glass tube. That came out and was replaced, but I was still unable to boot reliably.
At this point I put it away for a few days.
Brad over at MSPP started asking some leading questions about symptoms, and, based on my replies, was able to narrow the likely fault down to a handful of chips, most likely IC26, a 74LS74. He suggested that if the system was stable without the D ROM (and it was) then the problem was with the control logic, and IC26 was the one closest to the corrosion.
To test this, I dug out my logic probe and started checking logic states on the IC. I could see logic states coming in on the inputs (High, Low or Pulsing) but the outputs were just showing as floating. This was highly suspicious behavior, so I zipped down to my local Jaycar (I’m not sponsored. They’re just conveniently close) and purchased another Machined Pin socket and a replacement 74LS74.
Removing the 74LS74 was not easy. It really didn’t want to come out. While I could clear one row of pins, the second row just didn’t want to play at all. I tried all sorts of tricks. Extra solder, wicking solder away from the top. Extra flux. Resoldering. they just wouldn’t budge.
In the end I was worried I’d start damaging pads, so I cut the legs free of the chip and was able to remove the legs with a normal soldering iron. Even with this, I’d lifted one pad and damaged a second. Thankfully this was an easy repair, but still. I cleaned up the pads with braid and attached the new socket.
I dropped in the new 74LS74 and checked the logic again. I could see “sane” logic on the outputs, so that was an excellent sign. In went the D ROM and… it booted without a hitch. More importantly, my Menu was now showing all the options. Ooh!
I spent the next few minutes testing as many options as I could, including the Self Test. Everything just ran. The self test also confirmed that the keyboard was fine.
While I had everything open, I also put a proper heatsink on the 5v regulator, for longer life. Next time I go to Jaycar, I’ll probably build a little circuit with a diode inline and a switch, to drop the voltage from 12v to 10.8v, to take some load off the regulators.
My final step will be to make a nice clear acrylic case for it. Something so you can see all the internal “gubbins”.
So today my SRAM arrived, really well packaged and very promptly (So hats off to ScribblyGum Technology for that) and I thought, since I won’t be able to use my workbench for the next 3 days, I’d have a stab at replacing the SRAM tonight.
It was a battle to remove the old chips, even with a desoldering gun and tonne of my good flux. I had to clean the desoldering station halfway through just to get through the job.
Once the chips were finally out, I mounted in some machined pin sockets, and plugged back in the original RAM, to make sure my desoldering effort hadn’t made things worse. It hadn’t. If anything it seemed slightly better than before.
Now I knew things were not worse, it was time for a swapover test. I swapped in both of my “brand new” 2114 SRAM chips, powered it up and bingo!
That’s what I wanted to see!
A quick “Hello World” confirmed everything was hunky dory.
To see if it was one chip or both chips, I swapped the old chips back in, one-by-one and to my surprise, both were no good. They are MOS branded components, and they do have a reputation for being sub par compared to other companies for RAM and 74 series logic.
It’s awesome to have such a significant vintage system in my collection.
I’ve put the Microbee away for a bit. I’ve done a lot of work on it, and it’s a LOT more stable. Not completely stable and I’m yet to work out why.
(I’ve desoldered the connectors between the boards and completely replaced them. I’ve also replaced the ROM socket for the IC27, which was green and corroded)
In the meantime, I have had a delivery!
These two are the AVeraGe CART (AKA the AVG CART) which is a simply wonderful invention that extends the number of games I can play on my Atari 800 by a VAST number, and an IDE adapter (Complete with a teeny tiny IDE “Disk on Chip” HDD) for my Archimedes.
The first of these allows me to do something special. I can run Bad Apple on my Atari 800. This is a very special feeling.
The second of these adds a full blown hard drive to my Acorn Archimedes. This turns the system into a much more convenient environment for playing with that system. It also confirms that the problem I was seeing with the MIDI card was the MIDI card, not the Podule backplane, which is a huge relief.
So I’ve got a couple of days of leave, and now that my wife has plied me with Yum Cha, and I’ve done my first non retro repair around the house, I thought it was time to start working on another one of my retro-computers.
Sorry, not the PET. I’m still waiting for SRAM to arrive for that. Hope that fixes it, otherwise things get more interesting after that. (Hint: I will need to find an EPROM that I can actually plug into a system of that vintage.)
No I thought it was time to work on Microbee number 2.
One of the things that triggered this decision was the arrival of a pair of PCBs for something called a “Core Lifter“. This is an absolutely essential piece of kit for Microbee owners that plug in between the Main Board and the Core Board.
(Cor! Don’t I look fancy using all the “correct” terminology. This is mostly thanks to the people over at MTF and MSFF who are very patient with a doofus like me wandering in and asking strange questions like “What do you call that top board?”)
Now I had one, and had soldered on the various connectors, I was able to set the failing ‘bee up in such a way as to allow me access to all the core chips like the Z80 and the display chip. All ready. Let’s power it on and see if the fault is still the same, to wit: a black screen with no synch.
This is where things get strange.
See instead I got a stable “underscore” cursor. This means something is running.
I decided to get out the logic probe and see if it was clocking. I power cycled the system to see if I got consistency. I did not. This time, I got an error message!
OK so now I’m actually getting something booting through the ROMs. This is great! Pressing ESC even got me to the WORDBEE screen. This is more than booting. This is life. The vast majority of the system is actually functioning.
At this point I went out for a break. When I came back I power cycled it again and got… a BASIC prompt. Whoah. It was stable too. I wrote the obligatory “Hello World” BASIC and it was happily running.
I was a bit confused at this point. All I had done was add the Core Lifter. Was my problem just some dirty contacts between the two boards? DeOxit to the rescue!
Liberally spraying it into the connectors that connect the two boards, and easing them in and out again, followed by cleaning up the mess from applying DeOxit in the first place got me to a system where I was happy that the connectors were probably good. I also inspected all the connectors and checked the underside for dry joints. It all seemed fine.
It wouldn’t boot beyond the underscore.
I reseated all the ROMS carefully and…
OK that seems to be consistent. I have a working Microbee.
The story doesn’t end there, of course. On further inspection I noticed some corrosion around a whole bunch of connector pins. Looks like battery acid, which would not surprise me, considering Microbees had a battery mounted right next to the spot where the corrosion was. I reflowed all the pins on the connectors, desoldering the worst of them. a lot of crud came out which makes me think I was on the right track.
I’ve scratched back as much mask as I can with a glass fibre pen (A nasty piece of kit, perfectly adapted for this task. Any circuit mask that corrosion has crept under tends to come away from the underlying copper with the use of the pen. It does, however, leave evil glass fibers everywhere. These can get into your skin and are irritating. I always work over a piece of paper, clean with a brush at the end and chuck the entire piece of paper out at the end). I have followed up by wiping the exposed copper with white vinegar, followed by distilled water, followed by isopropyl alcohol. I have then covered the exposed metal with nail polish to stop further oxidation.
It’s not perfect. It needs a reboot occasionally on startup to get to the right prompt, but I wonder if that is related to the lack of battery. There’s also no sound. This will require more investigation. Finally, it lacks a case. I’m not quite sure what to do about that.
EDIT: After some more futzing around, it’s less reliable on bootup than I thought. It works most of the time, but not consistently. It sometimes boots to the Menu. It sometimes boots to BASIC. It sometimes boots to WORD-BEE. It sometimes just crashes.
EDIT 2: I’ve since replaced the connectors between the Main board and the Core board, replaced the socket for IC27 (as it was green on the inside), replaced the diode nearby (As it had green growing inside it and was testing rather oddly in my multimeter) and it’s still unreliable. I need to remove and test a couple of key ICs related to line select before I go any further, but 2 trips to Jaycar in one long weekend is enough. I’ll go play with my new toys instead.
Today I tried a little experiment and it almost worked. Unfortunately it didn’t work well enough.
I like the size and feel of gamepads these days, despite being raised on traditional joysticks. In the past I have hacked a couple of Genesis / MegaDrive controllers, purchased cheaply on AliExpress but I have always found them a little less responsive than I would have liked.
I wanted to see if I could improve the quality of the button connections, to make them more reliable. They currently use conductive pads on tinned plates directly on the board. For systems designed for more mechanical switches I sometimes find the contacts didn’t reliably register.
My solution? I found some really small switches and soldered them on over the spots where the contacts are supposed to land.
It was a pretty easy soldering job. I just worked out where the switches had to sit to correctly conduct on both sides, masked off with polyimide tape if I was worried about shorting, scratched back solder mask with a blade, added a dab of flux and soldered them on. Cleanup was Isopropyl as usual.
Unfortunately at the end I’d say the response wasn’t better, just different. For the fire buttons, it’s a great improvement. They feel lovely and tactile and you definitely know when you make contact. Unfortunately for the direction pads, they’re possibly worse. The directions don’t really work, especially in the diagonals. You can’t “lean in” to get that diagonal. I might take it apart again and desolder the pads for the directions.
The Microbee is happily working now. I swapped out an underpowered PSU for a more “meaty” one and now it’s booting reliably and is able to load games off my nearby Android phone. I just use the built in Android playback app to playback WAV recordings of Microbee software. I successfully played a range of games including Frogger.
To get Frogger to work, I had to build a joystick adapter, I built one according to the instructions in the Microbee Engineering Notebook. (Page 10)
This is a simple resistor array. across some inputs, so was a nice easy afternoon project. The original design suggested cutting the end off an Atari compatible joystick. Instead I just added a DE-9 socket so I could switch and change which joystick I wanted to use. It was a cheap and easy upgrade and I am happy with the way it looks!
Commodore PET 4032
This one scared me a lot. The age of the equipment. The integrated CRT. The power supply being integrated with the motherboard. So much to go wrong! I spend the week cleaning and cleaning and cleaning and cleaning the board. I initially brushed it with antistatic brushes to get most of the dust off it. (About a tablespoon of dust and a dust bunny of fluff the size of a golfball).
I then attacked it with isopropyl alcohol and a toothbrush. That got more of it off, but I ended up having to use distilled water on a brush, followed by isopropyl, followed by a cloth to wipe up what was left. It’s still got a thin residue of caked on dust on it.
After that I checked the rectifier diodes (OK) and then checked the voltages out of the transformer.
They looked OK so it was time to turn it on and…
Well that’s better than I expected to be brutally honest. There’s a cursor visible in there, and I was able to type in a “program” and see some of the garbled characters changing.
At this point I’m suspecting either a corrupt character ROM or bad Video RAM. Not sure which. More investigation will be required.
So I have now tested the Microbee with software and I can report I have successfully loaded several old tape games. It did highlight two issues: Two of the keys were not registering at all.
I carefully desoldered them but couldn’t work out how to get into them at first. They seemed quite different to the key replacement guide I’d found, but then I found a second one that was much more useful. this one covered the black keys in my unit.
Once I had the keys apart I was able to determine that the conductive pads were no longer conductive. They had a resistance in the MegaOhm scale. Scratching and sanding failed to expose any new conductive surface and I was about to start sacrificing another of the Microbees for keyswitches, when I decided I couldn’t make these switches worse so I tried some radical surgery.
I took out the “conductive” pad completely and replaced it with, wait for it, some rubber I cut out of a thong. I then stuck a carefully cut piece of aluminium foil to the top and tried it. Surprisingly it worked perfectly! One of the keys is a little stiff, but the other is indistinguishable from the other keys, except for the whole “Actually works” thing.
This is not a hack I am proud of but if it works, I will live with it 😀
I want to make a joystick for it next. I have a bunch of Joystick games and no way to play them. I am impressed with the fidelity of the graphics on this system. They may be black and white, but they are a very sharp black and white.