I have one of the Microbees booting, having made a quick cable to supply power and video. In the end, I needed to reseat the top board and remove the corroded battery, and that was it!
Back on the 11th,I bumped my Atari STᴱ, it slid off the shelf and fell less than half its own length. Unfortunately the old brittle plastic has taken on the same material characteristics of enemy cars in the Fast and Furious movies, and promptly exploded! Seriously! it went into DOZENS of bits. I was still finding bits today.
I’ve patched it all up with plastic glue but this is no longer a “Pristine” system, alas.
I also have another system theoretically coming this week. I’ll post when / if I get it. I have learned to never trust I’ll get something until it arrives.
So that moratorium on buying new Computers? Yeah that’s totally gone. Look what I got!
So I got an incredibly kind offer of a Microbee, with box, no less, and two more “spare parts” systems that were nearly complete.
The Microbee is another Z80 based system. Later models ran CP/M with disk drives but these are from earlier series systems. The system is built around two boards. The lower board has the CPU, video chip and I/O. The upper board has the RAM. Interestingly, in this case that consists of 16k or 32k of SRAM. Yes, SRAM, not DRAM. There’s also a monster Capacitor on board, so it can be turned off for “some time” and not lose the contents of the SRAM. An interesting design. EDIT: This is incorrect. That’s just for smoothing the power supply. There’s a battery (As I found when I opened the sealed unit) that backs up the SRAM.
There’s a key cable I’ll need to make for these that is a DIN 5 on one end and is cassette in and out on the other plus power plus Composite on one cable! Should be fairly easy to make up so I’ll probably be visiting my local Jaycar this weekend.
So while I wait for RAM for the Acorn Electron, I thought I’d see if anyone wanted to trade for some of my “spare” systems. I had an extra Atari STᴇ that I had already upgraded to the maximum 4MB of RAM, so I posted on one of my forums if anyone had a spare Amiga they wanted to swap for it.
One of the forum-goers was amenable and offered to swap an Amiga 600 for it. I readily agreed. A day later we had arranged the swap and I now own an Amiga 600.
I like this system for many reasons. Firstly, it’s pretty close to a “perfect” OCS/ECS Amiga, which gives it plenty of compatibility. Secondly, it’s VERY compact. While not having a numeric keypad can be seen as a limitation, the A600 is short enough to fit away where longer systems like the A500 couldn’t. (I’m currently working out if I can build some custom shelving for the longer systems like the 464 or the other STs). Finally, it’s quite a versatile system, capable of taking a HDD, extra “chip” RAM and even PCMCIA devices. Maybe I’ll finally test out that network card I have. 🙂
At the moment, it boots and waits for a disk that I can’t give it. I tried installing a hard disk, but had problems finding one small enough to actually work, without having to upgrade the Kickstart in it. While, down the track I am contemplating a Kickstart Switcher, I don’t have one right now. (I’m contemplating using a 27C160 EPROM, which is FOUR times the size of the normal ROM, and then putting on 3 different Kickstart images, plus a Diagnostics ROM all on the single chip)
I’ve ordered an adapter to allow me to install a Compact Flash card in instead of the HDD, and I have a suitably sized CF Card I can use.
I’ll replace the Floppy with a Gotek as soon as I order some more. That’ll have to wait for the tax return as that will involve ordering 5 drives, 5 OLEDs, 5 rotary encoders and 7 ultra small form factor USB sticks. That’s a big whack to get in one go.
I have also swapped out most of the yellowed keys for less yellowed keys, as the kind swapper had almost a full set of spare keys. I will, however, need to fake up a stabilising bar out of something suitable as none came with it.
Finally, while it supplies good voltages, the PSU currently trips the RCD occasionally, so I’ll want to replace that too. I plan to upgrade it with a MeanWell in the next few months.
Oh and there’s potentially a very exciting system coming in the next few weeks. If it arrives I will be extremely happy.
I plan to eventually add a page per system linking off from these two landing pages. The aim is to aid people who want to explore my collection. I have some extensive leave coming up later in the year so I guess that’s a project for then 🙂
It was Acorn’s attempt to produce a “price reduced” version of then hugely successful “BBC Micro” family of computers. Much of the logic from the BBC has been reduced to a single integrated chip, the ULA. Unfortunately for Acorn, the ULA was delayed, meaning they missed a significant holiday season, and ended up late to the market.
Nevertheless, it’s an interesting creature, and this particular one came with the tape drive, all the cables and even a bunch of books, magazines, and even an example of an assignment that the previous owner had submitted as part of his studies. (He asked if I wanted it and I said “Yes!”. I often find these pieces of ephemera fascinating insights into the real world uses these systems.)
Having got it home and gone through all the bits and pieces, I started investigating the system. First up was some pre-flight checks. The PSU is an external “wall wart” unit with UK prongs. Luckily I had an adapter I’d purchased to go with my BBC Master. I cleaned the prongs and then carefully measured the voltages and got a nice steady 19V AC out of it, which is what I was expecting. I then found a composite cable and plugged everything together.
Unfortunately, on bootup, I was presented with slightly corrupted blocks, rather than characters. I tried both the RGB signal and the straight composite.
A bit of Googling later showed no results, but I had enough experience with systems to know it’d be most likely one of three things. Clearly the CPU was running, otherwise the blocks would not be there at all. This means it’s most likely to be (in order of likeliness):
A dirty /damaged ULA socket. The sockets for the ULA are rather unusual. The ULA is on a bonded ceramic carrier with many contacts exposed around the edge. The socket consists of a series of spring loaded pins that reach up and press against the matching contact on the ULA. Unfortunately in harsh environments, both the pins and the ULA contacts can oxidise and need cleaning.
Bad RAM. RAM of systems of this age is notorious. This system uses the same 4164 RAM as the C64C, and I’ve had to replace quite a lot of that in my time.
Bad ULA. The ULA itself may have failed. I was hoping it wasn’t this one, as this is a relatively “unobtanium” chip. Thankfully it’s also a fairly rare failure mode.
I started tackling the potential issues, one by one.
Starting with the bad or dirty ULA, I popped open the system, carefully removed the clip that holds the ULA in place, and checked the pins and contacts. Both looked fine, but I carefully cleaned the contacts anyway, with an ink eraser, and a tiny amount of Deoxit.
Unsurprisingly, it didn’t fix things. Seeing how clean it was, I didn’t expect this to be the fix. Oddly the only dirty component in the system was the keyboard connector, which showed some strange white corrosion. I removed this with some IPA.
Next I looked at the RAM. I ordered some 4164 RAM chips, as I couldn’t find the 1 or 2 ones that I had left over from C64s. Initially I thought I was going to have to wait for RAM to come, but I thought about how the RAM worked and decided to just see if I could see any activity on the Data line on the RAM.
I dragged out my BitScope Micro, and after some false starts, got it hooked up to the Electron.
Checking the first RAM chip gave me what looked like regular data at about the right voltage. I moved onto the next RAM chip and the third RAM chip with identical output, However, when I got to the fourth (and final) RAM chip, I noticed an immediate difference. There was simply no data coming from that last chip.
I’ll wait for the RAM to come, get a socket and we’ll see if replacing that RAM chip brings the Electron to life. If that does fix it, I’ll have to get some shots of some games on the system. Down the track I want to see if I can get a disk drive / joystick port expansion for the system.
So I survived running a convention, and wanted to ease back into doing some retro computing stuff. One of the projects that came up as a possibility was a pair of Atari Paddle controllers. I already have paddles for the C64, but they use different variable resistors, and so are incompatible.
(As an aside, while thinking about this problem, I mentally reviewed doing the project with a digital encoder, a digital potentiometer and a Pro Mini, as I could build a universal adapter for both c64 and Atari, but decided that’s a project for another day. In the middle of contemplation I found myself waking up at 3am and finding myself thinking “I wonder if I could design a digital potentiometer? Turns out I could indeed design one.)
Anyway, I already had some nice spinner tops left over, so it was going to be a quick trip down to my local electronics shop to pick up some bits and pieces. There were some out of stock so I ended up doing some substitutions at the last minute, but here’s what I got:
Originally I’d planned to terminate everything to a solder DE9 female and hide all the “gubbins” in the shell for the connector. Unfortunately they were out, this the change of plan.
I measured and drilled the jiffy boxes to make sure they’d be vaguely comfortable in hand. (I then had to revise as I wanted a push button where the box had a support, so I had to move the buttons down a bit)
I mounted the potentiometers and buttons into the box, cutting off the shaft of the potentiometer to the correct length with a hacksaw. This is easiest to do once everything is mounted.
I cut the cable into two lengths and fed it through into the jiffy box, stripped the end and soldered it onto the potentiometer and the fire button, making sure to record the colour codes of what wire was connected to what component. I made sure both matched, to make construction easier.
I the cable tied the cable from the inside of the jiffy box, and added some heatshrink to the outside (I use two layers) to anchor the cable in place. 4 screws in each box secured the base, and then I used an allen key to tighten the spinner knobs to the potentiometer shafts.
Having finished the controllers, I moved onto the rest of the design.
I crimped the IDC cable to the IDC DE9 using my desk vice. I then stripped and crimped on 6 Dupont female crimps. These were so I could easily change any wires if I mucked something up. I had used 8 strand cable left over from another project so I just trimmed the connectors I didn’t need.
Next I got the veroboard (A scrap left over from another project) and soldered in all the sockets and pins. this was a straight through arrangement with 1 pin connected to 2 sockets. It just made it easier for the two connections that needed to be doubled up. Past me would have tried something ugly like crimping two wires onto one Dupont connector, but I have learned my lesson. I will, however, need to buy a LOT more female Dupont crimps in the not too distant future.
Next up was crimping on the Dupont male crimps to the wires from the controllers. At this point I put the female crimps in the 6 x 1 shell and attached it to the vero by the pin headers. I then worked out which wires needed to go to which connection and built up the 6 x 2 shell and male Dupont crimps until I had a suitable connector to drop in the socket at the other end of the strip of vero.
I took time at this point to make sure all the pins went to the correct place on the DE9. I have been foiled in the past by the order the pins come out on IDC connectors so I took it slowly, knowing that if I had to, I could always rearrange the Dupont connector order. They’re quite mechanically strong, but you can pop them out with the aid of a pin.
Once everything was checked I tested on the actual Atari 800xl itself, and it worked fine! After that it was a simple matter of closing up the junction box over the 3 cables and sealing it with a cable tie. A final change was to gently bend back the “ears” on the DE9 to make a better fit.
It took about 3½ hours all told, and that was not pushing myself at all.
While I wait for a bit more financial liquidity before I purchase some more bits and pieces, I’ve been contemplating what to do next. I’m in a strange place where pretty much all my systems work, one way or another. The Wizzard may need a replacement controller, but that doesn’t mean the main unit doesn’t actually work. It works fine. The Amiga may need me to finally get off my butt and laser cut the case, but it works fine. The Archimedes? Works. The Mac? Works.
Most of my systems are even mostly upgraded. I mean, I still need a bit hereand there and I need about 5 Gotek units so everything has an actual Floppy Disk Drive and that means $$$ so that is on hold for now.
My next immediate repair purchase will be in a few days I’ll send some money so that I can get a new PSU (And spare CRT Tube) for my SX 64. That will be a nice unit to have working again, especially as this one is in excellent condition externally.
I’ll probably build a set of paddles for the Atari 800. Unfortunately Atari and Commodore paddles are ever so slightly different, which means I’ll have two sets of paddles. That job is really an easy afternoon interlude, rather than a full blown project.
I want to build an ACSI2STM when I get a chance. I’ve laid out a PCB but it’ll need ordering AND testing. I’m cheating by cutting down a DB25 to DB19 to make it fit the Atari.
As for new systems, well I’m still keeping my eyes out for either a reissue or second hand Microbee at a reasonable price, but I’ve not located one yet. One can hope eternal. I have seen someone trying to flog off 3 for $1300 AUD in an untested state. That’s just a little too rich for my liking.
In the last post, I mentioned that the only thing left to do was to make up a cable and I’d be done with the Archimedes.
Well, that turned out to be quite the project, didn’t it?
So the point of the cable is twofold. Firstly, it physically connects everything together and secondly it changes the drive number on the second drive, as I’d been unable to find any obvious jumpers to change the drive letter.
I had some lengths of ribbon cable lying around so I stripped out a length, and pared it down to 34 conductors. I used a permanent marker to mark one side as “pin 1” for some of the length and the opposite side as “pin 1” for the rest. (Archimedes floppy pin 1 is opposite to Gotek pin 1, for some strange reason)
I crimped on 2 of the 3 IDC connectors, making sure to get the connectors / connector notch on the right side, as well as having them the “correct” side of the cable to make it easier to insert. (There’s a bit of a trick here, as the strain relief actually flips the cable so you have to factor that in.
I then carefully separated the strands every 4 wires. for the length that was to go to the Gotek. This guarantees ground cables will stay together with data lines. I ended up with 2 sets of 5 in the middle, but I knew that was going to happen. This was all done to make the floppy cable much more flexible.
Finally I trimmed about a 5mm section out of wire 1 on the Gotek side (wire 34 on the Archimedes side) as the Gotek doesn’t like a signal that gets sent down that wire and can burn the controller. Let’s not burn the controller. Honestly, I have read so many forum posts and the like getting to here, I don’t even remember where I saw that.
Now onto the other side of the cable. Following this guide (Part of a two part article which appears to be the only article out there on adding second drives), I carefully split the cable between wire 9 and 10, and 12 and 13 and added a twist and clamped on the final IDC connector. We’re done! (Or are we?)
I inspected my work, triple checked where I’d put the twist and mounted it in the Archimedes. The connector for the motherboard was the central one to make routing much, much easier. Everything looked neat and proper so I gingerly plugged one end into the Gotek and fired ‘er up. Well so far, so good. I could read and write to the Gotek OK.
Unplugging the Gotek and plugging in the FDD, nothing caught fire, which was a good sign. I went into the Configure app and added an additional drive. I inserted the Floppy and… No.
First up I got a Drive 10 Error and then “Drive Empty”. Maybe it’s that the other end wasn’t terminated in a device? Nope. Plug in the Gotek and neither device works.
I check my guides. I check my cabling. I check my logic. I check my software settings. Nothing obvious I can see.
Having posted my issue, someone else reports they weren’t able to get things working quite right yet either, using the guide I was working with so I start looking for another guide. I stumble onto this article (by searching for slightly different search terms) which has a reference to the actual model number of the internal FDD. And that gets me to a drive manual. Guess what? I don’t need to put the twist in the cable, as this model has a switch cunningly mounted on the side that sets the drive number!
It can’t be that easy, can it? I set the jumper on the floppy, (This entails removing the drive from the computer AND the mounting bracket) connect a straight through cable and… check! it’s working on floppy disk as Drive 1.
Oh gods! I pull the cable out, break off the strain relief bar, remove the end of the IDC connector with a pin, un-twist the cable, reclamp it and reattach it.
Oddly initially it works once and then starts doing “Drive Empty” errors, but then, as it’s now nearly 10pm, I screw everything back in and tighten everything, it just works perfectly. Huh! I have a two drive setup. Sweet!
At this point, for reasons I don’t quite understand, you need to access the Gotek before you access the FDD. Go figure?
So I have made quite a lot of progress since my last post.
I replaced the damaged capacitor on the mainboard. I cleaned the case a bit further to eliminate some corrosion I missed the first time. I replaced the dead cooling fan. (I need to modify this fan as it’s way too loud currently. Time for a resistor in series I think) I made a replacement backplate, so I no longer have a gaping hole in the back.
After all that I fired it up. Unfortunately the internal MFM hard disk is toast. Initially the floppy drive also failed.
As an aside, the process I went through to write out the floppy disk image I used is as follows:
Get a PC with both a “real” floppy drive controller, floppy drive and a suitably old OS. (I used Windows XP as it doesn’t care about driver signing)
Get a double density floppy disk. I have exactly three. Hope none of them ever die, as they’re getting stupidly expensive, and I am recycling what I have to bootstrap various systems.
Find a suitable disk image. I initially grabbed one I was unable to write out. Turns out it was more like an archive designed to be copied to a hard disk.
Install OmniFlop. It’s a great piece of software capable of writing out quite a few different formats.
Replace the floppy driver in your windows install with the one from OmniFlop.
Open the disk image from OmniFlop and write it to your floppy disk.
After all the work, my floppy disk didn’t work and I determined my floppy drive wasn’t even spinning, despite making all the “right” noises.
At this point I remembered I had a Gotek in with my Amiga “bits” (I have an Amiga 1200. It just doesn’t yet have a case). Time to pull it out and pre-use it for the Archie. I’ll replace it later. (I currently need an additional FOUR Gotek drives. One for the CPC 464, One for each of the Atari STᴇ systems, one for the Amiga). It already had a Rotary encoder and an OLED (Crap! I’ll need another 4 of those too!) as well as the speaker piezo.
I ran a long molex to floppy power lead and a long 34 pin IDC cable out of the front where the floppy drive is supposed to go and ran it to the Gotek. Almost immediately I had booted my first software, and what better to run than a classic?
One thing that didn’t work, and doesn’t work is the MIDI card that came with the system. It locks the system up with a cryptic error. (Error:AddressException:Address exception at &%0 (Error Number &80000003)) and I have not been able to identify what exactly the issue is. The StarDot forums suspect it’s either a bad chip, a bad ROM or just a plain incompatible Podule. Unfortunately no one seems to have the ROM anywhere so I can’t eliminate that.
Oh well. So much for my plan to run an MT32 emulator and have MIDI coming out of it.
I decided to clean and lube the floppy drive, and much to my surprise, it now works! I just cleaned with isopropanol and lubed with lithium grease. Nothing to fancy, but it did confirm my disk burning process was fine.
This put me in a dilemma. I want to use a Gotek, but I also want to use the floppy drive. The system does support two drives, but there’s only one floppy drive hole, and it’s non standard. After debating cutting a hole in the front bezel (Something I really didn’t want to do) I realised the system is designed with modules in the back. Why not mount the gotek in the podule bay?
My initial tack was going to be to design something and have it 3D printed. I even investigated alternative software to OpenScad but decided there was an alternative. My backplate is made out of some cheap Aluminium extrusion I found in a discard pile, so I had no emotional attachment to it. I had painted it to match the rest of the system, but that was it. I decided I could, if I was careful, cut suitable holes in the plate and mount the Gotek directly. I even had spare Gotek cases I could use and abuse. What did I have to lose? I even had some cunning plans for the encoder and OLED.
I carefully scribed the hole for the USB, the LED and the two buttons, and then carefully scribed them AGAIN on the correct side. (Measure twice, cut one and all that gumph). Some punching, drilling, filing and general cleanup and I had the holes transferred from the Gotek case to the plate. Now for the first tricky bit. I then used a fine saw and a rotary tool to cut the fascia off the Gotek case. It added too much thickness to the overall project. Then I carefully epoxied the bottom part of the gotek case to the plate and waited for it to dry. This allowed me to mount the PCB in place with proper standoffs. Once it was dry I did a test fit. A little more filing to make the holes slightly wider and it fitted quite nicely. Onto the second tricky bit.
I wanted the OLED and Rotary encoder to be a separate module I could plug into the back of the unit so I sat down with pen and paper and designed a simple setup. It consisted of a 8 way Dupont connector with a hole masked off as a polarity indicator. (I used a toothpick to mask off the hole). This then connected to 7 points on the Gotek board. On the other end I got some simple pin header, cut it down to 8 pins, pulled the matching pin to match the dupont connector, and soldered it to a piece of veroboard for mechanical strength. I then soldered on a length of cable, securing it to the veroboard with a cable tie for mechanical strength. (Once everything was tested, I then put a piece of glue filled heatshrink over the whole end to make quite a practical connector.)
At this point I decided “practical prototype today” trumped “final product that may not work in a week”, so I hunted through my parts bin for the first vaguely suitable container I could find. Turns out it was an old spice container that I’d already used for an electronics project several years ago. Out came the IR LEDs. In went the OLED and Rotary encoder.
Another wander around found a generic knock off child’s construction block that made it easy to secure the dupont connector to the backplate. I then just cut a small rectangle out of the top of the plate, glued the block to the plate and then glued the dupont connector to the plate and the block. It’s surprisingly rigid.
The first time I fired it up, the OLED didn’t work. Turns out I had inverted the wiring on the dupont connector I was using to connect on the inside to the Gotek. Thankfully it’s easy to remove and rewire them. I also had the OLED in upside down. It was only held in with electrical tape, so I was able to rotate it without any issues.
I’m very happy with the way it all looks.
The next step (for now) is to get some more 34 way IDC connectors and build the custom cable needed to support two drives. Once that’s in I’ll probably just enjoy the system for a while. Down the road I’ll get an IDE Podule, but not yet. I have already spend all my monies for this month. Next month I’ll be getting a cartridge for my Atari 800xl.
There was a “Pop!” and a bad smell. Looks like one of the chips failed. It’s all surface mount so there’s no chance to fix it, especially as the chip that failed is a large QFP with about a hundred legs.