Computers

8-Bit

Atari 400

Years (1979-1983)

Interesting Fact: Announced in 1978 and released in November 1979, the Atari 400 and 800 were the first efforts by Atari to enter the nascent home computer market. Originally intended to ship with 4K and 8K ram respectively – hence, the names 400 and 800 – a reduction in the cost of chip production meant that the 400 could launch with 16K of ram while the 800 had 48K. (source) Designed by a team that included Jay Miner, who would go on to design the Amiga 1000, the 400 was geared towards children with its wipeable, membrane keyboard that would not only make it easy to clean up after sticky fingers, but also prevent accidental choking from unintended key removal. (source)

Condition When Acquired: Fully Functional

Current Condition: Fully Functional

Project Details: I have to say, had I received an Atari 400 instead of a VCS back in 1980, I may have stuck with Atari when it came to buying subsequent computers throughout the 80s. The 400 is a fantastic little machine that has quickly become one of my all-time favourite systems. 🙂

I purchased this unit from my local retro store, Blast From the Past, where it had been sitting in storage for many, many years. Not knowing if it would work, I first tested the power supply and after noting a healthy 10.2 VAC (unloaded), I was pleasantly surprised when the 400 fired up right away. That all of the keys on the membrane keyboard worked flawlessly, was even better.

The only drawback was the machine was dirty after being in storage for what I gather, may have been a few decades. In fact, it looked like at some point, it had been in a barn or a farm building as there was grain (or something like grain) scattered inside the cartridge slot and across the tank-like RF shielding. This made me wonder if it had seen a few freeze-thaw cycles during its life, which would make replacing the capacitors a consideration.

As I disassembled the unit, it was evident that when this was made, Atari still emphasized quality. The cast aluminum RF shielding was a sight to behold as was the PCB with its power supply daughter board and vertical slots for the CPU card and the 16K ram card. I quickly noted that all the capacitors were axial except for the large 16v 4700μF capacitor. I also noted that only one capacitor was located on the main PCB, while the remainder were housed on the power supply. As I had some caps on hand, I placed an order to Mouser for the rest.

While inspecting the board and components, my attention was drawn to the RF jack which was quite corroded. It still worked, but it had reached the end of its life. Fortunately, I had a replacement on hand. This corroded part lined up with a few of the corroded screws that I had to remove during disassembly, which led me to further believe this unit had been stored in a dry, but temperature-variable environment.

I also made note of the date stamps on the chips and on the board itself. Knowing the production run for these units went into the early 80s, I thought it was cool that I couldn’t find a date later than 1979. This indicated to me, that it was probably manufactured in or at least close to its 1979 release.

Anyway, the cleaning process went well and soon the case and board were looking ship-shape. Re-capping did not take long, nor did replacing the replacing the RF jack. As there was a fairly good RF output including audio, I saw no need to put in a composite mod. I also left the 7805 and 7812 voltage regulators as they were, except for removing and replacing the old thermal paste. I considered replacing the regulators, but as I did not have the correct 7812 on hand, and as it was working properly, I simply left it.

After testing, I reassembled the machine and tested it again for an extended period. Everything worked well and I’ve since had fun playing different games on this system – and so have my kids. Again, had I received this in 1980, I may well have been an ‘Atari computer’ guy; eventually ending up with an ST instead of an Amiga… then again, maybe I’m just a fan of Jay Miner’s work. 🤔

Parts & Products Used: 99% isopropyl alcohol; Chemical Guys Natural Shine; thermal paste; PCB RF jack (1); electrolytic capacitors (axial) – 16v 10 μF (2), 16v 22μF (1), 16v 470μF (2), 16v 2200μF (2); electrolytic capacitors (radial) – 16v 4700μF (1)

Commodore VIC 20

Years (1980 Japan & 1981 international – 1985)

Interesting Fact: The VIC 20 was the result of Commodore founder, Jack Tramiel’s decision to push the company into filling a gap in the early home computing market. (source) His famous quote, “The Japanese are coming, so we will become the Japanese”, was amplified as his fist pounded the table while he argued with Commodore’s senior leadership, that it was now or never if they wanted to be the first to create and introduce a colour computer for under $300 to the home market. (source) His determination paid off as the VIC20 became the first computer to sell over 1 million units, with production reaching 9000 units per day and $305 million in revenue. (source) According to Commodore’s Michael Tomczyk, the goal was to create a friendly image of a computer for the home buying public, and the name, VIC 20, comes from the proprietary Video Interface Chip combined with ‘20’ because it sounded friendlier than the proposed ‘22’ after its 22 columns. (source)

Condition When Acquired: Partially Functional

Current Condition: Fully Functional

Project Details: This unit was a spectacularly priced, pawn shop find that came in the original box that included the original documentation, 4 games, and the Datasette. 🙂 Though it’s the later, Cost Reduced (CR) version of VIC 20, it still harkens back to the original version of the VIC 20 that was released in North America in 1981.

Before powering on the unit, I voltage tested the OEM power supply after noting a large bulge at the bottom of the casing. With the supply registering outputs of 5 VDC of the required 5 VDC but only 5.1 VAC of the required 7 VAC, I knew it was on the way out. However, as I wasn’t going to hook up the Datasette or my 1541 5.25 floppy drive, the AC voltage would not be needed to perform a simple test of the system. So, I proceeded to hook up the unit and power it on… success! It worked! 🙂

This was fantastic news, though I noted that the proprietary 5 pin DIN to RF modulator for video and sound output was wonky. I could get the main screen to display, but there was definite audio static and ghosting graphics. That the keyboard worked and the game cartridges also worked was a huge bonus. All in all, it looked like I’d stumbled onto a machine that was in fantastic shape minus the RF modulator.

However, before going further, I got sidetracked with my brother-in-law’s Amiga 1200 and when I ended up having to order a new power supply from Keelog for that project, I also ordered one for the VIC20. Once that arrived and I had finished the A1200, I got started on restoring the VIC 20.

On opening the unit, I was surprised at how dirty it was inside. It had obviously been stored in a dry location, but the board was coated in a thick layer of dust – more than what I’ve normally found in other consoles of the same vintage. No matter, a good scrub with IPA removed the grime and cleaned the cartridge slot. I also spent considerable time cleaning the keyboard.

Given the condition, and in case the video issues did not originate with the RF modulator, I recapped the board. A quick test showed that everything worked, but the RF modulator was still sketchy. I even replaced the two caps inside the modulator, which helped stabilize the output, but when I played more graphically involved games like Garden Wars, the output deteriorated quickly until it was unplayable.

As I did not have a 5 pin DIN to composite replacement cable, looked to modify the existing cable. Knowing that the VIC20 supplied a great composite signal along with audio through the DIN cable to the RF modulator, I debated whether to simply remove the modulator and affix a couple RCA cables as others have done, or – and this was the idea that popped into my head just as I was about to cut off the modulator – preserve the RF modulator by replacing the RF jack with a 3.5 mm TRRS composite jack.

This seemed to me to be a cool way to keep the original aesthetic and authenticity of the system, but upgrade it to a much better output. To do this was simple and merely required removing some components from the RF box. First, I de-soldered and removed the RF jack. I then drilled the hole so it was slightly larger so it would fit the TRRS jack. Next, I removed the channel select switch and anything that would interfere with the placement of the TRRS jack. I then de-soldered and removed the two electrolytic capacitors along with the blue 22 ohm resistor that was connected to the audio wire. As the 6v power supplied to the modulator was not needed, I snipped the red wire from the cable and from the modulator. Next, I attached a wire through the hole where the anode of the electrolytic capacitor met the trace coming from the input video wire, then I attached a wire through the hole where the removed resistor met the trace coming from the input audio wire. I then soldered a ground wire to the box at the point where the incoming ground was also soldered. Finally, I soldered all the wires to the TRRS jack, making a bridge between the two audio points so the sound (mono) would come from both the red and white audio cables. I then installed the TRRS jack, deliberately leaving lots of wire slack inside the box to ensure an easy fitting of the jack, before closing everything up.

That was it. It was perhaps a bit more convoluted than simply soldering on a new RCA cable, but in my mind, it was far more elegant. The new TRRS jack did not look out of place on the old RF modulator, and as I had paired it with a 6” 3.5mm to female RCA adapter, it would allow for more flexibility going forward. A quick test show a massive improvement in video and audio quality, with no glitching whatsoever when Garden Wars was played. Success!

My next step was to reassemble the computer, but before I did, it dawned on me that I still had some DIP40 heat sinks and thermal tape leftover from my work on the Bally Astrocade. That’s when I decided to put these onto the two VIA chips, the all-important custom VIC chip, and the CPU. I also left the cover off of the box surrounding the VIC chip for better airflow.

I then thoroughly washed and shined the case before reassembling the computer. Once done, I proceeded to inspect and recap the Datasette along with testing and cleaning the white Commodore joystick that I had found on Ebay. As I don’t have any cassettes to test the Datasette, all I could do was confirm that the belts were intact and that the machine’s buttons seemed to do what they were supposed to when hooked up to the VIC20. I also cleaned the play heads with IPA.

Once the machine was back together, I tested it the joystick and it seemed to also worked well, though it was obvious that Commodore did not put a huge emphasis on the build quality. Anyway, I’m very pleased to have this fantastic piece of history as part of my collection. 🙂

Parts & Products Used: 99% isopropyl alcohol; Chemical Guys Natural Shine; electrolytic capacitors (radial) – 16v 10μF (2); 16v 47μF (1); 16v 100μF (1); 50v 1μF (1); electrolytic capacitors (axial) – 16v 2200μF (1); 22 gauge wire; 3.5mm TRRS jack; 6″ 3.5mm to RCA female adapter; Keelog Vic 20 / C64 power supply

32-Bit

Commodore Amiga 1200

Years (1992-1996)

Interesting Fact: The A1200 represents the third generation of Amiga computers, and like the A500, was targeted specifically at the home market. (source) Unfortunately, by the time of the A1200s release in October, 1992, Commodore was already bleeding money and would declare itself insolvent in April,1994. (source) For an excellent history of Commodore and the Amiga, I would recommend reading: “A History of the Amiga” by Jeremy Reimer and watching “Amiga Story” by Nostagia Nerd. Both of these multi-part sources are fantastic.

Condition When Acquired: Non-Functional / Unknown

Current Condition: Fully Functional

Project Details: This is not my computer; it’s my brother-in-law’s (BIL). Yet, since it’s in the family, I’m counting it as part of my collection. 😁

A life-long Amiga aficionado, my BIL bought this system new when he graduated high school in 1993. After a few years of use, life happens and people move, change jobs, etc., and his beloved Amiga 1200 went into deep storage sometime in the late 90s, though he never stopped thinking about it.

After watching me hack out a side hobby of repairing old consoles, he asked if I’d take a look at the 1200 and see if I could re-cap it for him. To be honest, it’s one thing when you’re risking your own consoles to the dangers of meddling, but it’s another when you’re doing it for someone else, let alone close family. Couple that with the rarity of the system, and I was a bit nervous to say the least. 😯

Nevertheless, I accepted the challenge because as I’ve mentioned before, I consider Amigas to be the best gaming machines of their era. I had an A500 (as did my wife and BIL), but the A1200 was the next level successor in terms of speed and processing power. Unfortunately, it arrived on the scene when Commodore was going down the tubes and cutting costs – something that I could plainly see while working on this project.

After bringing the A1200 home, the first thing I did – even before unpacking the computer – was test the OEM A600 power supply. Noting that the voltages were badly out of spec, I put an order into Console5 for both a power supply re-cap kit and an A1200 re-cap kit. While I waited for everything to arrive, I set about cleaning the unit.

As soon as I removed the A1200 from the box, I heard something rattling under the hood. Once I had it open, I noted two obvious problems related to its years of storage: 1) a tantalum capacitor had fallen off the hard drive (the cause of the rattle); and 2) a poorly soldered, 50v 2.2μF smoothing / decoupling capacitor near the Video DAC had also come loose, and on further inspection, had lifted a pad and severed a trace in the process.

I also inspected the capacitors, particularly the SMDs, and while I didn’t see any obvious leakage, I knew from reading, that the ones used by Commodore at this time were about as reliable as the caps used by Sega for the Game Gear. Both of which had high failure rates due to using the cheapest parts available.

Once the re-cap kits arrived, I proceeded to re-cap the A600 power supply and test the voltages. Unfortunately, the re-cap didn’t fix anything and instead of wasting more time on repairing it, I ordered a new power supply from Keelog – a Polish company that has had great reviews for its retro power supplies. While I was at it, I also ordered a new power supplies for my VIC 20 and one for a ColecoVision.

While waiting for the power supply (which only took 2 weeks to come), I set about working on the Amiga itself (FYI the board was REV 1D). Using my heat gun at the same settings that I’ve used for other boards, coupled with copious amounts of flux, I began to remove the SMD caps. The first few went as planned, but the left (C324) and right (C334) audio output caps proved stubborn and in the process the pads and traces lifted. Crap! 😯

The same also happened for C409, C235, and C821. Fortunately, the traces were only severed for C334 and C409, and using high heat epoxy, I could reattach the pads and glue down the traces. For C334, the pad would not reattach to the trace, so I looked at the schematic and discovered that it was easily fixed by running a jumper from the anode end of the cap to pin 7 of the LF347M chip. For C409, the issue was extending the solder blob from the anode pad to the via that connected to the power circuitry sandwiched inside the board along with the backside of the board. This was important because the +5v of power that supplies the CXA1145M video encoder is supplied through here.

Having traces and pads lift like this was a fairly new experience for me and on a double-sided board with sections of sandwiched circuitry, it meant that I spent a lot of time reading the schematic to make sure I wasn’t overlooking something. That this happened at all made me question my abilities (and sanity), but there’s always a risk when you perform seemingly minor surgery. To be fair though, I do think that when compared to other boards I’ve worked on, the traces and pads were shallow and lightly affixed during the manufacturing process, making them more susceptible to lifting while heating.

Now that my follies had been corrected, I set about fixing the manufacturer’s mistakes, one of which, proved to be the most difficult fix of the entire endeavour: the lifted pad and severed trace for the 50v 2.2μF ‘add on’ decoupling or smoothing capacitor. I say ‘add-on’ because it was obviously an afterthought at the factory and does not appear in the schematics. In fact, I searched for quite a while and saw lots of REV 1D boards online and no one had this capacitor added. As Console5 notes, “Commodore seemed to use whatever was available on the market that met minimum specs and fit.” This statement also fits with the overall board quality as well.

Anyway, the trace from pin 18 (VREF) on the Video DAC runs to a via and then to the anode of the capacitor in question on D215, before continuing to an unused pad on D215A. The capacitor’s cathode attached to the ferrite bead at FB402D. The problem was, the pad the anode was attached to on D215, was not strong enough to hold the size of the capacitor and during moving and storage, the anode arm of the capacitor had caused the pad lift and the trace to sheer. The cathode had also completely severed, but it made little difference to the overall solder weld at the ferrite bead.

I thought long and hard about how I could reattach the capacitor where it wouldn’t come loose again, along with how I could bridge the very tiny trace. As I studied D215, I noted that only 2 of the 8 pads were actually used (1 on each side), and so my plan was to run a bridge from the severed trace over the remaining three unused pads to D215A. I would then use one or two of the unused pads across from the bridge to secure the anode arm of the capacitor. From the third unused pad along the bridge, I would run a second bridge across to the anode.

Using 30-gauge copper wire to make the bridges, my plan was successful and the new capacitor was firmly set into place and the trace was restored (Note: I initially forgot to cover the capacitor’s cathode leg with tubing to prevent shorting, but I went back later to add this). With the new power supply now in hand, I connected the board and fired it up for the first time. As I didn’t have an RGB connection, I was using the composite output which made all of my repairs crucial to the process. Anticipating something happening, I was disappointed that nothing happened…

I then tested the voltages and everything appeared normal, except for those on the CXA1145M video encoder, where I noted only 2.2v instead of 5v. I then correctly deduced that my bridge to the via underneath C409 was not perfect, so I removed and then reflowed the blob bridge. I reattached the capacitor and again powered the system on…

Success! I was greeted by a very clear start-up screen. I then turned my attention to the hard drive where I reattached a new, 10v 47μF tantalum SMD capacitor that I had ordered in from Mouser. I plugged the drive in and powered the system on again, and Viola! after all these years, the HDD fired up perfectly and booted right into Workbench. Amazing! 😀

Now that I was making progress, I then disassembled the internal 3.5” floppy, cleaning it thoroughly and replacing the single 25v 4.7μF SMD capacitor with a replacement hole-through electrolytic that I had on hand. I then reassembled the drive, hooked it up, and gave it a test. Success! It worked too. 😀

The final puzzle for me was the keyboard. At first it did not work and I spent time cleaning the ribbon cable along with the flex circuit membrane. When this didn’t fix the problem, I checked the continuity of each pin on the board along with the voltage on U13 – the keyboard MPU. Everything checked out. That’s when I realized that the white clip that holds the ribbon cable in place on the board, has a front and a back that is hardly noticeable at first glance. When I flipped the clip around and powered the system on, the keyboard worked perfectly.

After reassembling the A1200, I then worked on two accessories: 1) the Amiga 1010 external floppy drive; and 2) the Amiga tank mouse. First, I disassembled the 1010 drive (it was a Newtronics version) and replaced the three hole-through capacitors (6.3v 47μF 16v 47μF, and a 16v 4.7μF). The third capacitor required most of the drive’s mechanism to be removed to get to, but everything went smoothly and after a thorough cleaning it was soon back together. Next, I took apart and cleaned the tank mouse, and in the process, replaced the single hole-through 10v 47μF capacitor. This was a quick task.

With everything back together, I tested the machine along with drives for quite some time, noting that everything worked as it should. Boy, did the memories flood back while I was doing this and I think I’m as excited about seeing the Amiga up and going as my BIL is. 😁 I’m still impressed that the HDD worked perfectly and everything that he had put on it back in the 90s was still there, including his old game saves. Of the four boxes of floppy discs, some worked, but others did not, as age has started to take its toll. Nonetheless, this was a very special project to work on!

Parts & Products Used: 99% isopropyl alcohol; Chemical Guys Natural Shine; Amiga PCB electrolytic capacitors (radial) – 10V 1000μF (2), 16v 470μF (2), 50v 2.2μF (1); electrolytic capacitors (SMD) – 6.3v 100 μF (4), 16v 47μF (2), 35v 10μF (3), 35v 22μF (5); WD Tibit 60GB HDD tantalum capacitors (SMD) – 10v 47μF (1); 3.5 Internal Floppy Drive (original was an SMD but I replaced it with a radial capacitor) – 25v 4.7μF(1); Amiga 1010 External 3.5″ Floppy Drive (Newtech) electrolytic capacitors (radial) – 6.3v 47μF (1), 16v 47μF (1), 16v 4.7μF (1); Power Board electrolytic capacitors (radial) – 16v 470μF (1), 16v 3300μF (1), 25v 220μF (3), 50v 1μF (2), 50v 4.7μF (1), 50v 47μF (1), 200v 200μF (1), 450v 1μF (1); 26 gauge wire; 30 gauge coated copper wire; Keelog Amiga power supply; ABLEWE RCA to HDMI Converter; JB Weld Heat Resistant Epoxy