Repairing an Epson HX-20 – Tape Drive & Printer

With the new battery in place it was time to test the main peripherals – being the cassette and the printer. The cassette drive is actually a plug in module – it could be replaced with a ROM cartridge, and as originally shipped with a simple blanking plate. The quality of the engineering is superb – it feels very solidly made and slots into place with a very positive action. I have a cassette supplied which contains the original logging program for the data logger complete with instructions.

As described above, the tape transport is entirely under computer control including winding, and there is even a digital tape counter accessible from BASIC. The ‘WIND’ command rewinds the tape, and I was pleased to find that sure enough it rewound to the beginning. The ‘LOAD’ command (as expected!) searches the tape for something to load so I typed ‘LOAD’… and nothing happened. No sign of life from the drive at all, and the computer reported a ‘IO error’.

It is very common for old cassette machines to have problems with their drivebelts as the rubber stiffens and perishes over time. However the belt was likely intact to some extent because the winding mechanism seemed to work. It clearly did need the drive stripping down and servicing, and this proved to be quite a task.

Looking round the ‘net it is possible to find some information on how to do this. Given the small scale it is unsurprisingly quite a fiddly job. I’ll make a list of useful HX-20 sites but I’d start with this excellent resource from Martin Hepperle, and specifically the ‘Tips and Tricks for the HX-20‘ PDF:

https://www.mh-aerotools.de/hp/hx-20/index.htm

There’s loads of useful stuff in here, including some advice on replacing the drive belt. There is also a huge stack of other manuals – including a massively detailed technical reference manual – here:

http://www.vintagecomputer.net/fjkraan/comp/hx20/doc/

The technical manual gives a very nice exploded diagram of the tape drive, and in conjunction with Martin’s guide tells you everything you need to know. However it is extremely fiddly. The drive itself is a modular unit which slides out (and I understand it originally shipped as an optional extra). There are a few screws to remove the outer shell, and then several more which hold the PCB to the back of the drive mechanism. Some of them have brass standoffs, and all of them are really small so great care is needed not to lose any.

One of the hardest things to do is separate the mechanism from the top shell – it’s not clear how to do it, and takes a lot of wiggling around with the door open to do it. The technical manual talks about removing the front of the tape door but I couldn’t see how to do it and after a lot of fiddling around it eventually came off. So I can’t really offer any advice other than to persevere!

One thing I did learn – do not under any circumstances remove the screw which forms part of the door release latch assembly. I did this thinking it would make things easier, and in doing so the tiny spring which reset the eject switch popped out. I spent a very long time trying to get it back in again… which was incredibly difficult to do. After countless unsuccessful attempts I was eventually able to do it although the spring ended up rather buckled in the attempt.

Once you have the mechanism out you need to unsolder a few wires from the motor, remove the tachometer cap and then finally you can unfold the mechanism and reveal the belt. I had got so absorbed in the process that I didn’t take many photos unfortunately but there are details in the guide linked to above.

To get the belt out, you have to remove another metal bar which sits across the mechanism. The one on mine was intact but had clearly hardened up and was very slack and clearly would not do a good job. However it did not explain why the drive was not working at all, as even a missing belt should still allow the mechanism and motor to start. The belt itself is 50mm in diameter (so about 160mm circumference) with a square cross section. I have a bag of belts of varying sizes so tried a few until I found one which fitted. They do need to be thin enough to pass underneath the tachometer cap so a bit of experimentation may be needed. The cross section should be anything between 0.8mm and 1.2mm square – mine was 1mm.

I then roughly reassembled the module and slotted it back in to see what else was wrong. It was clearly obvious that the real problem was that the tape head was not engaging, and so the computer was behaving as if no tape were loaded. The mechanism is circled on this picture, although here it is in the engaged position.

This mechanism is driven by a separate motor with a worm drive which moves the head in an out of position in a continuous loop. A sensor detects when it is in place which stops the motor. On initial inspection I could not see anything wrong. With long nose pliers I could turn the motor shaft and the mechanism worked smoothly. So the problem either lay with the motor itself or the logic driving it. To test this further I desoldered the wires from the motor (circled on the picture below) and used an external battery pack to see if I could get any life out it:

Initially it appeared completely dead, which was not good news. However, after a bit of fiddling around and without seemingly having done anything, it suddenly burst into life! I spent quite a few minutes watching the motor run smoothly and engaging and disengaging… quite hypnotic. I’m still not quite sure what made the difference, but I think the most likely thing is that the motor had seized through lack of use and a combination of moving the shaft manually and the external power source was enough to kick it into life.

It was then simply a matter of further testing. I reassembled the unit without the case, took a deep breath and…

Success!

It was then a matter of reassembling properly and testing with an actual tape. My unit was was used with a data logger and came with the logging program on tape. It’s not been touched for many years, but I put the tape in typed the ‘LOAD’ command and to my great satisfaction it found and loaded the program:

The final piece of the puzzle was the printer. It was clearly working although the ribbon in it had completely disintegrated. Amazingly you can still buy brand new replacements as a similar mechanism is still used in some cash registers:

You can easily find these – either originals or cheap knock-offs – online very cheaply. On installing it, I was pleased to find that the printer did work but I had a lot of trouble with it apparently jamming. The paper would advance very slowly with constant beeping from the computer. This turned out to be simply due to the plastic gears binding due to a lack of lubrication, and a quick spray with a silicone lubricant (designed for plumbing, but great also for plastic gearing) left everything working smoothly. You can ‘screen dump’ direct to the printer or output to it in BASIC.

The quality is surprisingly good, and I have a box full of original branded Epson paper rolls although these also are widely available.

So after quite an interesting journey I now have a fully working HX-20. There is plenty of other stuff to explore – although the HX-20 scene is quite limited, it is still possible to find software from various sources which can be loaded through an external tape port. I would also really like to see what the data logger can do, will have to think up a suitable retro project for it. Also need to clean up the case a bit. It remains one of a very small group of computers which have printer, display and storage in a single unit. In fact the only other one I can think of is the much later Canon NoteJet / BubbleNote series of laptops

This whole thing has given me a taste for these quirky early portable computers, and I know I would have absolutely loved something like this back in the 80s. Watch this space for more adventures…

Amstrad PPC640 Restoration

Whilst I am interested in old tech and computers of all types, I’ve come to realise lately that my favourite brand from the 80s is Amstrad. Not perhaps everyone’s choice, but I was really into Amstrad at the time (owning a PCW 9512 which I would still love to replace). I also used various Amstrad PCs which were something of a phenomenon at the time for their low price. Although I never actually owned any other Amstrad products I followed them eagerly and well remember looking at them in Dixons etc on the High Street. They had a reputation for making packaging up previously expensive or complicated things (eg hifis) and making them cheap and easy to use. Whilst at the time people grumbled about the quality, looking back it now I think that their stuff was actually pretty well designed and well made, and the savings were often made by clever design decisions.

One of the more interesting Amstrad products of this era is the company’s first take on a portable computer – the Amstrad PPC. Laptop / notebook / portable computers in the mid to late 1980s were expensive and exotic things costing thousands of pounds, and the idea of portable PC computing was completely out of reach for most. The PPC followed the usual Amstrad philosophy and delivered a portable computer at substantially reduced price.

I’ve recently picked one of these up, which is the top of the range PPC640D. The design is really interesting, with a small folding LCD screen, and then a full sized desktop style keyboard which folds over the whole of the rest of the machine. There are two 720k 3.5″ floppy drives on the right hand side, and an internal 2400 baud modem – which was a really impressive addition at the time. The overall package feels surprisingly solid and sturdy.

I bought this without any idea if it worked or not and there have been a few issues. The first concerning getting the machine to start up properly. It came without a power supply, but fortunately (and unlike many portable computers) runs off a standard 12V power supply with a barrel connector (centre positive). You do need a fairly beefy one because of the current requirements. You can also – uniquely I think for a portable PC – run it off standard ‘C’ batteries (10 of them!) although I can’t imagine you’d get much running time.

The power switch (just to the right of the screen) is also a bit odd in that it switches between two ‘on’ states – either ‘battery on’ or ‘external on’. With an external power supply connected I was pleased to see the lights come on and ‘Please Wait’ on the screen. However, after a few seconds there were three beeps and just a flashing cursor. The disk drives ran for a few seconds, but beyond that no response.

There isn’t a huge amount out there written about these machines but I did find this reference to a similar issue. It’s all about a small bank of DIP switches which are hidden away on the side of the machine. The relatively basic nature of the hardware means that the display is not auto-detected in any way and the switches are used to set the display options. If you get it wrong then there is nothing on the screen. There is a useful guide to the DIP switches here. When I got it, all the switches were down ie on, which led to a fault condition in which the internal display was set to a CGA (ie colour) mode but the screen type was set to MDA (ie monochrome). The solution was to either set the screen mode to MDA (switch 2 off) or the screen type to CGA (switch 4 off for 40 columns or switch 5 off for 80 columns). I opted for the latter, and the switches now look like this:

I can’t really work out why they were ever set to all on, except perhaps someone well meaning thinking it looked neater!

The second issue was about the casing itself. Although it felt sturdy enough I could hear bits of plastic rattling around inside. More significantly, there was clearly something wrong with screen hinge, and although it did work it was not very stable and you could see one of the pins on one side was loose.

Getting inside was pretty easy – 6 large screws on the base, with two of them inside the battery compartment. When I got it open I was surprised to find there was a lot more damage than I expected, with several of the screw posts broken off, and various other bits of trim rattling around:

Fortunately, the breaks were pretty clean and it was easy to see where the broken pieces went back, so it was just a question of mixing up plenty of epoxy resin and sticking the various bits back where they came from.

Getting screen out was a bit more tricky, and requires removing one of the motherboards, the modem and several connectors:

This picture shows where you need to get to. The board at the top of the picture usually sits on top of the other one, with a large connector block as well as a few power cables. The modem is a separate board on the right hand side (with the grey box).

Of particular note are the three ribbon cables top left, two of which are the keyboard and the other the screen. These push in to connectors on the top board:

First of all there are a few screws (indicated by arrows) to remove. There are so many different connectors it is tricky to get everything out, but the simplest thing to do is unplug the blue and white connector blocks first, then ease the board upwards and then it should be possible to flip it over so that you reach the ribbon cables, which just pull out. You do need to very careful with these as they are rather fragile.

Having done this, the screen can then be removed. It’s held in place by plastic pegs which locate in holes on the chassis, and with a bit of careful wiggling around it will come out. The problem was immediately apparent:

One of the plastic pegs had snapped off. This proved tricky to resolve – the peg is designed to move slightly in and out and is moulded to be flexible, but trying to fix it in place spoils this. I didn’t take enough pictures, but in the end I glued a thin piece of plastic inside the screen behind the peg, and then glued the peg to that to give it more support. Others have suggested drilling a small hole in the peg and using a metal pin for support, or else even more drastic solutions.

My approach didn’t quite work out as planned. Not all of the glue stuck properly, and the peg is still a bit floppy. However, it’s definitely better than it was before and the screen stays in place. I could have another go at it but it’s good enough for what I want it to do.

Having put everything back together again, it only remains to get it started. The machine did come with a pile of disks including one of the original Amstrad ones. This didn’t work, but with the help of a Greaseweazle (another article for another day) I was able to write a new image out to it:

Success! It boots up nicely to DOS 3.3. There’s quite a lot of interesting things to explore with this now, including testing the internal modem. However for now I’m happy to have it restored to working order again, and I’m debating taking it out somewhere and trying to do some real work on it. Would draw a striking contrast with the usual selection of Macbooks seen in my local coffee shop!

Upgrading a Commodore 64 power supply

One of the first things I bought when I came back to retro computing was a Commodore 64. I’d never had one of these before, having been firmly in the Spectrum camp when I was growing up. For some reason I’d never really had the curiosity to explore this before but I decided the time was right and bought a decent bundle off eBay, which came with a handsome bundle of games which were fun to explore. It was also interesting experiencing tape loading complete with music which didn’t happen on a Spectrum.

One thing I quickly discovered on reading around was that the original power supply was something of an accident waiting to happen, with some quite sensational articles with dire warnings of impending doom. Unlike a lot of early computers it is a reasonably complex power supply with a 9V AC and a 5V DC rail which feeds a DIN plug, and so it’s not absolutely straightforward to replace.

The original design of these is extremely odd (it seems to me) in that the they are fairly simple devices with a transformer and voltage regulator, but then the whole lot is encased in epoxy resin making it a heavy and completely solid block. I really can’t imagine why they did this, other than perhaps to make absolutely sure that no-one actually tried to service it. I’ve not seen any other computer manufacturer of the day do anything like this. The problem referred to above is due to a breakdown of the DC voltage regulator, which is buried in the epoxy resulting in a significant over-voltage and generally a destruction of the C64 itself. The epoxy is also poor at radiating heat which I’m sure doesn’t help the situation

There are various opinions out there on what to do about this – most people recommend avoiding the stock PSU, although I tested the voltages on mine and it was fine. There are quite a number of third party units which are very nice, but rather expensive for what they are:

After a lot of debate, I took the view that the only issue with the stock PSU was the regulator, and after finding a few other examples of people who had modified it I thought I would do the same. There are various different types of these, but mine is the so-called ‘wedge’. Others are just big rectangular bricks, but all of them have the same epoxy construction.

The first thing to do is get into it – easier said than done. The plastic base is glued on, and part of it is buried in the epoxy itself. So it does take a lot of pushing and prying (and some cutting with a sharp knife) but eventually it comes off. All you can see is what is below – most of the wiring is buried in the epoxy but there are taps straight of the transformer for the AC supply, and then via a regulator (circled in red) for the 5V DC.

So there are two jobs to do – the first is disconnect the existing regulator, and the second is to provide some kind of modern alternative. The first is a fairly simple matter of desoldering the pads and cutting off and removing the regulator legs. You could probably manage without this but I really wanted to make sure there was no danger of issues down the line.

There are various modules out there for regulating / rectifying voltages (Buck converters etc) but the guide I linked to above suggested using a ‘UBEC’ (Universal Battery Eliminator Circuit). This is something used in remote control vehicles for getting high voltage battery pack outputs converted down into the lower voltages needed. So it’s ideal for my needs, which is to convert the 9V AC from the transformed into the 5V that I need.

These are cheap and easy to find on eBay and all the other usual sources:

These are not particularly sophisticated devices, and there is apparently a fair bit of electrical noise on the output but these older computers are not sensitive enough to be troubled by it. You do have to configure the UBEC with jumpers to get the 5V you need. It’s also small enough to fit neatly inside. You need to connect the input to the UBEC from the main smoothing capacitor, and the output to the output pin from the voltage regulator (the right hand one on the the photo below). The negative output can simply be left unconnected.

If you’ve done it right, then you should be able to measure about 5V coming out and off to the C64. It’s then just a matter of fitting everything back together again.

Overall it’s a fairly easy job and I am now confident that I can use the power supply without issue. A nice spinoff is that it runs much cooler and still looks authentic. It’s been running happily for long periods with no problems at all.

Atari ST monitor cable – DIY / NOS

Since last posting things have moved on considerably and I have picked up quite a lot of new and interesting items. I’m going to go into these in more detail as we go but I’m going to start with a bit of curiosity.

One of my recent pickups has been a nice Atari 1040STE outfit. I’ve had a longstanding ‘relationship’ with the ST. It was the first 16 bit machine that I became aware of back in the 80s, and although later on I became more interested in the Amiga I never actually had either of them although I did borrow an ST for a time.

Looking back on it now, whilst I know the Amiga is superior technically I really like the industrial design of the ST which is I think is much sleeker and more interesting than the Amiga and especially the peripherals (mouse, monitors etc). So when it came to revisiting it all, I started with the ST again.

I didn’t mind too much which variant I got but after a bit of looking around I found a boxed 1040STE with various bits including several boxes of disks and a ‘Powerplay Cruiser‘ joystick. The STE is an enhanced (hence ‘E’) version of the original ST with a blitter, stereo sound and a few other refinements although from what I gather very few games supported this. It’s always a gamble when you buy tech like this but I was pleased to find that it was in good cosmetic condition and (surprisingly perhaps) in full working order including the disk drive.

Whilst it does have a TV modulator (and the picture was surprisingly good on my Trinitron TV) I wanted to connect it to a proper monitor, and one of my other pickups in recent times was a Philips CM8833-II monitor. So the first job was to get a suitable cable.

There are plenty of ebay traders selling ready-made cables although these are usually SCART rather than anything else, and I do quite like the ritual of making cables up. One sticking point is that for some reason the ST uses a lot of very odd connectors (including the notorious 19-way D type connector). The monitor is no exception and uses a very strange 13-way DIN connector with the pins arranged in almost a square pattern:

About the only other thing I know of which uses this is the equally obscure power connector on the Yamaha CX5-M MSX computer. I have an interest in these too and will talk about these in future, but a word of warning if you like these is never buy a CX5-M without a power supply because it’s quite difficult to make a replacement because of the connector and the need for various positive and negative voltages.

So if I wanted to make my own cable I’d need to source the connector. Fortunately they are not that hard to find on ebay these days, but on looking around I found something even better:

This is a ‘new old stock’ (‘NOS’) Atari ST monitor cable which was sold by CPC back in the 1980s. CPC is still around (as part of the mighty Farnell empire) and these days are a general electronics retailer (pretty good too, I’ve had a fair bit of stuff from them). However back in those days they specialised in spare parts for also sorts of consumer electronics, and I remember the catalogue which had complete parts lists for Sinclair, Amstrad and other computers. They obviously handled others too, and this cable was specifically sold as a ‘DIY’ with the 13 pin DIN on one end and pre-prepared stripped wires at the other end.

The pinout for the ST is widely available but here it is for convenience:

There’s actually quite a lot of interesting stuff on here which you don’t always find, including a pin 8 which is used for SCART switching, pin 01 which carries audio (but only in mono) and a choice of composite, mono or RGB video signals. There is also (oddly) an audio in on pin 5 but I’m not quite sure what you’d use that for.

The Philips CM8833 range have quite a choice of different connectors depending on the exact variant but the one I have uses a 9 pin D connector:

V & H on the above refers to vertical and horizontal sync. Although the monitor does have an internal amplifier and speakers, you can’t feed it from this connector but you have to use separate phono inputs instead. This is a pain and so you need to run a separate set of wires either from the back of the computer (which has L & R phono outputs) or you could possibly run one out of the D connector but this would be awkward.

The only thing left to do is work out which wire goes where. You would hope the colour coding would make some kind of sense but you can’t take it for granted. So in case any one else happens to find one of these cables (there do seem to be some more on ebay, perhaps someone is clearing out a workshop) here is what I found (function on the left, wire colour on the right):

FunctionColour Atari ST8833-II
Red (R)Red73
Green (G)Light green64
Blue (B)Light blue105
Horizontal syncBlack98
Vertical syncYellow129
GroundBraid131

I had a few problems at first with getting the two sync wires the wrong way around but once I’d sorted that out it the picture looked great – really clear and bright and the famous ‘little green desktop’ look perfect.

I have a few ideas in mind for this – there’s an interesting project called ‘ACSI2STM‘ which emulates an ACSI hard drive using some low cost components. It does need a bit of hacking around (not least because of the limited availability of the 19 way D connector) but looks like fun. Meanwhile I’ve had a lot of other new toys in so more on those soon…

Rediscovering Retro

I grew up in the 1980s at a time when microcomputers were the new thing, and my lifelong interest in technology was really sparked by the excitement of that period. It really did feel as though anything was possible, and rate of change and innovation was far greater than it seems these days.

Over a long period of time I’ve maintained an interest in the tech from this period, and at the time I collected quite a decent amount of kit which I picked up cheap or was given by others. At that time, I was interested in anything I could get my hands on although I rapidly got bored and moved on to the next thing. By the early 1990s I had moved on largely from 8 bit to PCs and the emergence of Linux and the Internet – both very exciting – and so the collection was put away and neglected. I did get heavily into the emulation scene which emerged in the late 1990s and this seemed the perfect combination and so really there was no need for all the old hardware.

I never completely lost interest though and picked odd bits up here and there, but also went through phases where it all felt like a bit of a burden and I got rid of stuff which seemed of no interest, giving some away and selling others. Some of this I now very much regret – eg throwing out a big pile of Acorn manuals, giving away a BBC B and monitor, throwing an Acorn Atom in the bin(!) although to be fair it didn’t work and was missing half its casing. I also had an impressive Commodore PET collection – three in total, plus loads of disks, books, tapes, manuals and other peripherals – that I gave away. The trouble is that at that time (early 90s) it all seemed old fashioned and boring – after all, why try and write text based BASIC programs when you can be running full blown Unix on a 386 PC!

So that was pretty much where things remained for a long time. I had a few boxes full of stuff that I would look at sometimes, but I never really set it up or spent any time with it. I did buy things sometimes if I saw them but usually tired of them pretty quickly.

However, like many people the recent lockdown brought some significant changes. On a whim I was able to buy a Microvitec Cub monitor (as seen on 1980s BBC computer setups) for £6 off ebay:

I already had quite a lot of Acorn kit and so I got it out again to test the monitor… and all of a sudden I found I was really enjoying handling real hardware again. I was also surprised how much stuff I had that I had completely forgotten about and which really I have no idea where it came from.

So this has led me to really dig into my whole collection and I am really enjoying rediscovering what I had and also I’ve started buying more for the first time for years. Prices have gone up a lot (although I think there is something of a bubble caused by lockdown that may not last) but are still within reach.

The other thing I’ve also suddenly realised is that things which were once common and of little value – specifically CRT monitors – are now becoming rarer and are much more desirable than they once were. I never appreciated this before but there is something very authentic about using whole systems from the correct era. It just doens’t look right using LCD panels… so I’ve been buying those too!

There isn’t anything really unique about this but I am going to write some posts about my experiences, including any fixes or upgrades along the way as there are things available now that never were in the past. As usual this is partly for my own benefit and so that I can remember what I did and how I did it… but hopefully of interest to others too!

Repairing a Kawai K1 synthesiser

I’ve always liked music and music tech, although I’m not really much of a musician myself. I’ve written before about repairing an old Casio keyboard rescued from a skip. I’ve started looking around recently at old, unregarded and (importantly!) cheap synths to play with. The classic 1980s ones like the Yamaha DX range are now really quite expensive (£500 or so for a decent one) making me regret not looking at this before given that a few years ago you could have got one for next to nothing. However trawling ebay I found someone locally selling a Kawai K1 untested reasonably cheaply so I bought it.

I’d not really heard of this one but it appears that it was quite popular back in the day and they sold a lot of them. They seem to have been in the doldrums for years but there does seem to be some interest again. There are a few interesting websites with more information:

https://www.musicradar.com/news/blast-from-the-past-kawai-k1

https://kawaik1.wordpress.com/

The one I got was in a pretty sorry state when I got (and not _quite_ as described by the seller either!)

There is some damage to the casing on the front at the right and the rightmost keys are slightly damaged. It’s a bit shabby all over with some dirty keys and needs a good clean. It came without a power supply although it’s a standard barrel connector and a 12V power supply (although ‘centre negative polarity).

I was pleased to find that it powered up without an issue, although on trying to play there were quite a few issues. Some of the piano keys didn’t work, none of the control buttons worked and the joystick didn’t seem to do anything. I had a strong suspicion that the internal battery had failed too.

So first job was to open it up and have a good look. The Kawai blog above has good advice on doing this, basically a matter of removing all the screws on the base of which there are several different kinds. There is a good account of it here:

https://kawaik1.wordpress.com/2019/05/16/dissassembling-the-k1/

Once I got inside there were a few broken bits of casing rattling around but in general not too much obvious damage:

The main board has a copper sheet on the bottom, but you can get to the top of the board by unbolting the board and carefully flipping it over:

There’s nothing obviously wrong here other than a badly bodged backup battery install. It looks like someone has torn the straps off the old battery and put a new one in with sticky tape, but it’s not really worked. I could replace this with a proper tagged battery, but these are a bit of a fiddle and so instead I thought I’d use a battery holder. There seem to be quite a few of these out there now, and I found a lot of these out there:

They are very cheaply made, but let you install one or two CR2032 batteries in a case and solder the flying leads to the board. This should make it much easier to replace the battery in future should this ever be needed. So I cut down the old tags and soldered the wires on, fortunately the positive terminal was marked on the board. I stuck the holder to a convenient space on the board with some self-adhesive tape:

The next problem is the unresponsive piano keys and control buttons. For both of these it was simply a case of stripping everything down. These keys work by a rubber mat which pushes down on some PCB based contacts. I was able to unscrew the button board and pull the mat off, and clean the contacts up. It was the same approach for the piano keys, and the details are in the post linked to above. There was nothing obviously wrong with any of the contacts but there was a fair bit of fluff which I cleaned up. I had read on the internet somewhere that people had trouble with the copper sheet under the PCB so I took that off too although it didn’t make any different. However it gave me a chance to have a good look at the underside of the PCB which again looked fine.

There was also some damage to the top ‘C’ key and a stabilising post had snapped off. Unfortunately I broke this further trying to get it out and whilst the key still works it is a bit loose and gets stuck. However I can’t think I’ll have much occasion to use that particular key anyway.

The final thing to look at it is the joystick. I’m not entirely sure how it is meant to work but it seems to be working reasonably well in the Y axis but not in the X axis. These joysticks are actually fairly simple devices with two potentiometers:

From a few measurements they seem to be fairly standard linear potentiometers tracking from 0 to 4.2k. There doesn’t seem to a big problem with them from measuring with the meter, and similarly the cabling etc also seems fine. However I’ve sprayed a liberal amount of switch cleaner on it, the connectors at the other end and the socket on the main board. After leaving it overnight to let the cleaner soak in I gave it a good wiggle around in both axes before refitting it.

Then it was simply a question of putting it all back together again and testing it. Somewhat to my surprise… basically everything worked! The keys all work, the control buttons work (although they do need firm pressure) and the joystick also works smoothly in both axes. So it’s rather a shame about the physical damage, it would be great to find a new casing from somewhere although that’s a bit of a long shot.

The backup battery is also working correctly, so the next thing to do is reload all the stock factory patches and get some more interesting sounds out of it so that’s for part 2.

Repairing a Gameboy Cartridge

I’ve had a passing interest retro gaming for years, although it’s more about the tech than about actually playing the games (which I’m really not very good at). I have got a small collection of odds and ends, and one thing which I got more recently was a Gameboy Advance SP Zelda edition, rather like this one:

One of the nice things about these is that the play original Gameboy carts, and back in the 1980s there was nothing I wanted more than one of these. I fell in love with the accessibility and sheer playability of the games in (what seemed like) a small and compact package. I spent a lot of time with them after borrowing them from friends but never had my own.

I did start looking out in second hand shops for cartridges though, just to see what I could get my hands on. I found a couple of interesting ones, but the best was a Pokemon Silver cart:

It was just the bare cartridge and it worked fine, although it was clear straight away that there was a problem. These cartridges use an early form of ‘save game’ technology, done very simply by using battery backed RAM. At the time I remember this seemed like magic as the cartridges were identical to normal ones and there was no sign of the battery externally. However, after 20 or more years these batteries have unsurprisingly run out of life, and so whilst the game worked fine it didn’t save anything and you had to start again each time. Sometimes these old batteries leak and damage the actual electronics, I previously bought a Pokemon Red for which that had happened and the game simply wouldn’t start.

Fortunately as this is a common problem it was easy enough to find what I needed for repair. The cartridges are held together with a single ‘security’ screw which needs a special screwdriver, and these are easy to find on eBay etc:

They often come in sets with other sizes for different Nintendo products, eg SNES cartridges. Make sure you get one for the cartridge, the actual consoles use another type of special screw (a ‘triwing’) for which you can also buy screwdrivers. I have some of them too… another story!

The other thing to do is make sure you have the right battery. There are at least two types (earlier games use a CR1616, later ones often the more common CR2025), and the easiest way to check is to open it up and have a look:

Printed on the board is the battery type (CR2025) and the date code shows it was made in January 2000 so shortly after the game was released. The battery is soldered in rather than socketed, so you do need to find a replacement with tags or else hack in a normal one. There are various suppliers on eBay and others who carry these so they are easy and cheap to find.

I took the opportunity to take the old battery off an clean the board up a bit. It’s in pretty good condition given its age:

Installing the new battery was pretty easy too, just a couple of blobs of solder, but as always make sure you get the polarity correct (in this case negative is the lower right and positive the upper left terminal). The replacement battery I had the contacts in a different place, so it looks different but the polarity is correct:

Then it’s just a matter of putting it back together and testing – which worked fine – so now it’s a question of actually getting on with the game!