Technology and retro computer blog

Amiga 500 rescue

I recently rescued this Amgia 500 from being recycled.  It looked like it'd been dragged through a river.  It was reasonably complete and surprisingly not broken.  I also managed to get the power supply, but both cords had been cut off.  I figured it most likely wouldn't work but it'd be fun to tear it apart and see what was inside.


The first thing I did was brush the dirt off the top to see what was under it.  There were a couple of missing keys but it started to look reasonably good.  With most of the thick dirt off I started to take it apart.  I noticed that the warranty stickers were still in tact so it doesn't look like the machine had ever been opened before.  This was a good sign as I expected that it was complete on the inside.


Amiga Amiga Amiga


On getting the top cover off it started to look pretty bad, it was really dirty inside and the RF shield was very rusty.  I took the shield off to expose the motherboard which looked a little better.  I could see that the bottom part of the RF shield was also pretty rusty, so I figured the motherboard would be damaged.  It looked like water had leaked into this machine at some stage and sat there for quite a while.  When I removed the motherboard I was surprised to see it had survived due to the plastic insulation sheet between the motherboard and the bottom RF shield.  After a quick brush off the motherboard looked really good.  There was some surface rust on the rear connectors but that was all.


All the chips were in tact and it also had a memory expansion module installed.  The clock battery on the expansion module had leaked and there was some leakage on the clock circuitry.  The internal floppy drive looked reasonably good.  I removed the memory expansion card and the floppy drive.


Amiga AmigaAmiga


I removed the bottom part of the RF shield, the rust from the bottom shield had covered the bottom part of the plastic case.  With the functional bits removed I figured I'd do a quick test to see if the machine would power up.  With the cords cut off the original power supply I used an ATX power supply instead.  I looked up the pinouts for the power connector, I recall from past reading that the Amiga 500 power supply can be replaced with an ATX supply so I knew it had all the necessary voltages.  I plugged in an old Amiga RGB to scart cable that I had in a box of parts.  I haven't had an Amiga 500 for over 20 years but I managed to get an RGB/scart cable with a box of parts I picked up a while back.  Luckily I kept the cable.  I connected the cable to a small LCD TV that has a scart input and on initial power-up all that I got was a white screen.  I did a web search about Amiga's with white screens and there were several reports relating to water damage and chips not seated properly.  I re-seated all the chips but no change.  I eventually had the idea to connect the Amiga's mono video connector to the composite input on the TV and I was greeted with a workbench disk screen.  The issue with scart cable was that my LCD TV didn't support RGB through scart.




I found this very motivating as it showed that the machine was doing what it should.  I cleaned up the floppy drive and plugged it back into the machine and on subsequent power tests I got it to the point where the drive would spin ready for a disk.  This all looked very promising.  Not having been involved with Amigas for over two decades I had no software to test it with so I started doing some web searches on getting software.  There are utilities around that will transfer software from a pc to the Amiga but these only work if you boot the Amiga with a Workbench disk first.  I borrowed a copy of Workbench 1.3 and managed to firstly boot the machine and then use it to copy the existing disk, everything worked fine.  With evidence that the machine seemed to work ok, I got stuck into cleaning the case and other parts.








Once cleaned up the case came out looking really good.  I got the rust off of the RF shield and the bottom case and re-assembled everything. I also created a better power plug for the ATX supply.  I extracted four pins out of a round DIN connector and soldered them to an ATX molex lead.  Since re-assembly I've connected the Amiga to my pc and used Amiga Explorer to transfer disk images across to it.  The process worked fine.  The memory expansion card also works fine, I removed the leaky battery and cleaned up the board, it now reports an extra 512k of memory.  The Amiga boots reliably and I've since tested it with a couple of games which worked fine.


1st November 2015

After the success with the main machine, I tested the power supply.  I hooked up a mains cable and checked the output voltages and all voltage were fine!  I've re-attached the mains cable to the power supply and I'll need to get hold of an Amiga square 5 pin power connector to fix the cable to the machine.  These plugs are not easy to come by so I'll have to attach the makeshift version until I can get one.  The A500 was perfectly fine, aside from rigging up a power supply and cleaning nothing special was done to get this machine to work.


Amiga Amiga

15th November 2015

I've now replaced the missing keys and created a good makeshift Amiga power cable.  I've also added a Gotek floppy emulator which allows loading of ADF images from USB stick.  I mounted the Gotek on top of the Amiga using some double sided tape.  I passed the data and power cables for the drive through the back of the case.  I still have the original floppy installed but just un-plugged.  I think this is a much better solution than hacking up the case to fit the Gotek internally.  This way you can revert the Amiga back to it's original configuration.

Amiga Amiga Amiga

Macintosh Sony SuperDrive head spring repair

I recently obtained a couple of compact Macs with Sony SuperDrives.  In both cases the drives wouldn't read disks.  I did the usual servicing of the insert/eject mechanism and both inserted and ejected disks fine, but they still wouldn't read disks.  In both cases the issue with the drives was that the top head was not touching the disk.  I managed to diagnose this by connecting the drive to the computer with the cover off and watching as it tried to read the disk.  Then I simply pressed down on the head with my finger so it would touch the disk.  Once I did that the drive would start to read and write.

This seems to be a common problem with these drives because they use a leaf type spring on the top head.  What typically happens is when the machine is stored away for years on end with no disk inserted, the drive head is held in the open position and this causes the head to stay in this raised position.  The other cause would no doubt be people unwittingly raising the head too far in order to clean between the heads.  This would cause the leaf type spring to bend upwards.

I'm glad to say I've found a fairly easy fix to this problem.  I've repaired both drives and they are now working fine.  So I thought I'd write a guide on how to fix this head spring problem.  The fix involves two things:


  1. Bending the leaf spring on the top head back down
  2. Adjusting the internal spring inside the head to further drag the top head down on the disk.


In order to perform these tasks you need to remove the head from the drive.  This is a fairly simple process.  There are two small screws holding the head down.  These screws clamp down the slide bar on the right side of the head (looking from the back of the drive).  Then there are two small card cables which simply pull out of their sockets.  These need to be removed.  Once both those things are done the head can be removed from the drive.

To bend the leaf spring back down again, you need to insert a small flat head screw driver towards the back of the head assembly.  There is a small metal plate where you can rest the screw driver.

Sony SuperDrive


Push it in enough to spread the heads apart.  Once that's done you push down gently on the top head.  The screwdriver acts as a fulcrum forcing the spring to be bent back down.

Sony SuperDrive


The second fix involves moving the position of the small spring inside the head to increase the tension on the top head.  On the bottom of the head assembly there are three small adjustment steps that enable you to increase the tension on the spring.  But I've found that even moving the head to the third position is not enough to provide enough downward pressure on the top head.  So my solution was to notch out a fourth position further away.  I used a broken off piece of hacksaw blade to create the new notch.  This increases the tension on the spring enough to provide the required down force.

Sony SuperDrive


Internal spring moved to the fourth position.

Sony SuperDrive


Macintosh LC575 disassembly for donor parts

I've been looking for a hard drive and floppy drive for my SE/30 restoration, so as luck would have it I found this LC575.  I offered the owner $20 which was accepted.  I hoped to get the hard drive and floppy out of it as both would make nice era appropriate replacements for the drives in the SE/30.  I'm glad to say both worked out well.  I managed to salvage a 250meg Apple SCSI hard drive and floppy super-drive from this machine, both work, but in need of a clean-up, as would be expected.  I also managed to remove the internal cd-rom drive.  I haven't tested that yet but no reason to expect it won't work.  Before tearing this machine apart I did try to get it to boot, but on power up all I could hear was the dong of the screen.  The screen was just blank and no activity from the hard drive or floppy could be heard.  The only attempts I made to get it working was to check the fuse on the power board, that was ok.  I also tried the trick of switching it on and off really quickly, I'd learned this trick when trying to get my LC475 working.  If the clock battery is dead the LC475 won't start up, but switching it on/off quickly will get it to boot.  But no luck with the LC575.  Not wanting to spend too much time trying to get the thing working I just started tearing it apart.  I must admit it's quite good fun to disassemble stuff.

These aren't too hard to take apart, a few Torx T15 screws on the back and the back just lifts off.  The floppy and cd-rom drives can be removed through a drop-down cover at the front, no screws required.  The hard drive and logic board come out through a cover on the back which does have a couple of screws but were missing from this one.  The internal drive/board case just un-clips from the front bezel and can be removed.  A fairly interesting case design.  Case parts and monitor will probably go to recycling.  The logic board looks in fairly good shape and has the ram and CPU on it which are most likely good, so I'll hang on to that in case someone wants it.  See pics below.


Macintosh SE/30 restoration

A few days ago I picked up an SE/30 from an ebay auction.  The seller was located near me so that was a bonus.  It was in fairly good physical condition but the case was very dirty and had paint stains on it.  The seller showed the machine at the disk prompt so that was a good sign as many of these machines are in a much worse functional state.  On getting the machine home, I tried to boot it using the Floppy Emu disk emulator but unfortunately I couldn't get it to work reliably.  I tried booting from several OS versions 6.05, 6.08 and  a couple of times I did manage to get it to boot to the desktop.  But it was pretty clear that the machine had issues. 


Over the last couple of days I disassembled it and cleaned it inside and out.  The case has come up reasonably good.  I managed to get all the paint stains and other marks off.  I cleaned the floppy drive mechanism and the drive is now able to insert and eject disks properly.  On re-assembly the boot problems still persist.  None of the logic board capacitors have visibly leaked, there is oxidation on some of the contacts of each capacitor and also on some of the ic's but the board is generally clean and in good condition.  So I was hoping to avoid re-capping right away.

These are the things I've tried on it so far:

- Boot using Floppy Emu and physical floppy, Floppy Emu will boot to the desktop on occasion, but machine crashes shortly after.  Physical boot floppy is not recognised
- It came with 5meg or ram, I've swapped the ram with other ram and tried different configurations, doesn't seem to change anything.  All ram is reported through the finder when I've been able to boot to the desktop.
- I've tried re-seating all removable chips,  no change
- Just prior to making this post, I removed all socketed chips on the logic board and then thoroughly washed the board with bi-card/white vinegar and then detergent.  It's in the process of drying out at the moment.  I'll re-try it once everything dries.


The SE/30 lives

I've been reading a few posts online which lead me to tinker with the Quantum 40s drive.  I figured as the platter motor was stuck maybe the head was too.  I removed the cover from the drive again and removed the magnet from the head mechanism and then adjusted the head tension screw.  Before re-assembling the hd I plugged it in with the cover off.  The head now moves freely, I hadn't seen it move before.  Not expecting much I plugged the drive back into the SE/30 and on second power cycle the machine started to boot!  It now boots reliably.  The drive is rather slow and sounds fairly sick, but the good news is that with a healthy and faster drive this machine should work fine.  I haven't replaced the caps yet and I'm still not getting a bong on start up, so this has given me some good motivation to spend some time re-capping the board.  I still can't get a response from the Floppy Emu.  With it booting now to the desktop, I tried loading a disk image from the Floppy Emu and there was no response at all.  The more I read into it I think the bourns filter might be the issue with reading floppy disks.  Some good progress.



11th October 2015

 I've managed to repair the floppy drive from the SE/30!

 I've been doing some reading on issues with these Sony super drives and determined that the issue with this one was that the top head was not making contact with the disk.  I put the drive in my LC475 with the cover off and lightly pressed down on the top head with my finger and it started to read properly.  I even did the same to format a disk and that worked too.  This is a common problem with these drives, because they sit around for years with no disk inserted, the head spring stays in the up position for a very long time.  The other cause is that people lift the top head to clean it and this can bend the spring up.  The spring on the top head is a bit of sheet metal screwed to the top of the head assembly so bending down is not an easy job because you can only bend down until the top head touches the bottom head, and this is not enough travel to bend it back into position.


To repair it I had to remove the head, this involves removing the two small screws holding down the slide bar on the head and also removing the two small card cables attached to the head.  Removing the head is fairly simple.  There is a second spring inside the head assembly which draws the two heads together.  This spring has three adjustment steps so you can increase the downward pressure on the top head.  So I  set the spring to the highest step.  Thinking this may fix it I put the head back into the drive and tested again.  Unfortunately even though the spring setting had lowered the top head so it was now touching the disk, the pressure was not enough.  Pressing down on the head with my finger was still required to make the head work.


So my solution was a novel one, I cut another notch on the bottom of the head assembly to make a fourth spring position for the internal spring.  I used a small hacksaw blade and cut a tiny notch further away from the last spring position.  Putting the spring in this fourth notch increased the pressure on the top head just that little bit more.  On testing the drive again unfortunately it was still not enough pressure to get the head working properly.  So I removed the head again and this time came up with another idea to bend the top spring into place.  I stuck a small flat screw driver underneath the sheet metal top spring so that the two heads were spread apart, then I pushed down on the top head.  This caused the sheet metal spring to be bent downward slightly.  Once again I put the head back into the drive and tested again and success, it now reads and writes without any assistance! 

See my detailed post about the head spring repair below.


13th October 2015

A further update on the SE/30.

I installed the Quantum 250 meg hard drive that came out of the LC575 into the SE/30.  I also re-installed the newly repaired Sony SuperDrive back in.  I then installed system 7.1 using some newly acquired (ebay), LC475 install disks.  The new hard drive had some old files on it so the first install worked but I got some extension errors on boot, some old files were messing with the new install.  I then tried using the system 7.1 disk tools floppy to wipe the drive but for some reason the 7.1 disk tools disk wasn't happy being booted in the SE/30.  It complained that it didn't work with the system version?  So I used the disk tools disk from system 6.0.8 and I was able to re-format the drive using that.  I then did a clean install from the 7.1 disks and it worked perfectly.  I now have a clean install of 7.1 running off of the new hard drive.  New drive is much more quiet and quick than the original, SE/30 boots in only a few seconds with 7.1 and seems to run really quick.


All in all the SE/30 is working well now.  It boots reliably and screen looks great.  I still haven't done the re-cap, I'm a little hesitant to do it when the machine is working so well.  The problems that still persist are:


  • No sound.  Most likely related to caps from what I've read.
  • Can't seem to use the external floppy connector.  I tried connecting a floppy drive to it and it wouldn't even power up.  The floppy emu works somewhat but continuously comes up with read errors.  Not sure if this might be related to caps or possibly a power supply issue


I guess the last reasonably simple fix to address the remaining problems  is to re-cap.  Would be great to have the sound working.

24th October 2015

I replaced all the surface mount capacitors on the logic board yesterday and I'm glad to say it went well.  I wasn't getting any sound before, now I get clear and loud sound through the headphone jack.  But unfortunately I'm not getting sound through the internal speaker.  I plugged in another internal speaker (smaller), and I get faint sound through that.  So I checked the voltages at the plug that connects to the logic board (with the plug removed),  I'm getting -11.1v and +11.62v on pins 7 and 8 on the connector.  The 5v supplies are fine.  So I suspect that there's some issue on the analog board or power supply.

The SE/30 generally runs great, screen is good, SCSI port works fine, internal drives work fine.  The only things I have issues with are the internal sound and the external drive connector.  Whenever I connect a drive or Floppy Emu to the external connector it's very flaky.  Not sure why the internal drives are not affected.  To fix the 12v supply I'll look at doing a re-cap of the analog board.

Commodore VIC20 & 128 haul

I saw an ad on a local site for a Vic20 and Commodore 128 bundle. What really caught my attention was that both the original boxes for the computers were included. Both boxes are in very good shape. I've been wanting to get a Vic20 again since it was my first computer so it was a good opportunity. There was also a brown 1541, a cassette unit and a bunch of other stuff. I didn't realise until I got everything home how much extra stuff there was, mainly all the original brochures and manuals that came with the computers and the original Vic20 poster in mint condition.

Dungeon Hacks: How NetHack, Angband, and Other Roguelikes Changed the Course of Video Games

Dungeon Hacks, David L. Craddock

I was contacted recently by David L. Craddock who requested the use of a screen shot of Rogue that I have in my early Macintosh disk archive.  The screenshot was for use in his then upcoming book Dungeon Hacks.  He offered a pre-release copy of the book for review which I was happy to accept.  The book covers the development of the RPG game genre of Roguelikes, named after Rogue the game that started the genre.

I really enjoyed reading this book.  It's extremely well researched and written and has a great deal of interest in it, not only for people into this genre of RPG gaming but for anyone interested in computer and software history.  For example, early pioneers of Rogue were directly in contact with people like Dennis Ritchie who was instrumental in the development of Unix and C. 

I was never one who got into Rogue or any of the directly related RPG games on the Unix platform, but I did relate to many of the games and people mentioned in the book.  For example, there‚Äôs a section in the book on Sword of Fargoal on the VIC20.  My first computer was a Commodore VIC20 and one of my all-time favorites on that computer was Sword of Fargoal.  I recall at the time I had to borrow $90 from family members to buy the game and the 16k memory expansion cartridge required to run it on the VIC20.  This was a lot of money at the time.  I recall the great excitement of playing that game and finally finishing it.

I highly recommend Dungeon Hacks for anyone interested in RPG's and the history of game development.  Many of you I'm sure will relate to at least some of the games and platforms they ran on.


Dungeon Hacks is now available for purchase on Amazon here:


Also here:

David L. Craddock, the author of Dungeon Hacks


Commodore 64 Epyx Programmer's Basic

 Epyx Programmer's Basic, originally Hesware Graphics Basic for the Commodore 64.  Read about it here:


Epyx Programmer's Basic Toolkit


Epyx_Programmers_BASIC_Toolkit.pdf (6.77 mb)

Epyx_Programmers_BASIC_Toolkit_Addendum.pdf (134.20 kb)

Epyx_Programmers_BASIC_Toolkit_Reference_Card.pdf (930.23 kb)

 D64 disk image: (86.90 kb)


Hesware Graphics Basic

 D64 disk image: (89.05 kb)


FDC+ Enhanced Floppy Controller for the Altair 8800

The FDC+ is a new floppy disk controller board for original Altair 8800 computers.  It was developed by Mike Douglas, see the FDC+ website and Altair Clone website for more details.  The original Altair 8800 floppy disk controller was a two board s100 configuration which only connected to the original MITS 88-DCDD drive.  In contrast the FDC+ allows original Altair 8800  computers to connect to a variety of floppy disk drives from the original MITS 8" floppy drive to more recent TEAC FD-55GFR 5.25" drives.  It also allows the streaming of disk images to the Altair via a serial link, so you can use disk images without a physical disk drive at all.


My FDC+ arrived yesterday and I spent some time today getting it up and running.  My goal  was to use a Teac 5.25" drive to act as an Altair 8" drive and to create a CP/M boot disk.  The FDC+ comes with several adapter pcb boards for use with different drives.  You select the required drive configuration and then solder the appropriate connectors onto the bare pcb.  The pcb for use with TEAC 5.25" drives requires two IDC sockets, the first is a 50 pin socket for connection to the FDC+ board and the second is a 34 pin socket for connection to the drive.  I created the required adapter for the TEAC drive.


To get things up and running quickly and easily I decided to configure the Altair with just the CPU, FDC+ and SSMIO4 serial card.  I enabled both ram and prom on the FDC+ and configured it for the correct drive type.  I have a working Teac FD-55GFR(149) so I set the jumpers as per the FDC+ manual.  I used a really short 50 pin scsi cable to connect the FDC+ to the adapter board, it actually works well because it just suspends the adapter board next to the FDC+.  I then used a standard pc 34 pin floppy cable from the adapter board to the drive.

To get CP/M onto disk I used the following approach, I booted the Altair and ran the hex loader program in the FDC+ prom.  Using the hex loader I then loaded PC2Flop.  I then ran PC2Flop and created a CP/M boot disk.  One thing I noticed is that I'd configured my drive to be drive 1 not 0.  Pc2Flop worked fine once I selected drive 1.  After creating the boot disk, I examined the combined disk boot loader and tried to boot, but realizing that my drive was configured to be drive 1 not drive 0 I had to correct the jumpers on the drive.   Once I did that I was able to boot CP/M.


I used DS/DD 5.25" disks in the TEAC drive and while it initially worked, I found on subsequent attempts to create various disks, the process would quite often fail.  After some discussions Mike suggested removing the LG jumper from the TEAC drive.  Once I did this it made the disk creation process quite reliable.  Although in further discussions it was suggested that ideally high density 5.25" disks should be used instead of double density disks.  I have yet to try the HD media to compare results.  But for now DD disks seem to be working fine.



Compaq Portable

I saw this Compaq portable advertised in my local area, not knowing much about these machines I did a bit of research on it. I discovered that it was not only the first IBM pc clone but also Compaq's first product from the early 80's. I really like these old "portables" with built in CRT's so I snapped it up. The price was fairly cheap, so I bought it without doing much testing. The seller didn't have any working boot disks, so I just tested that it powered up and displayed a missing boot disk prompt on the screen, that was good enough for what I paid.


Compaq Portable Compaq Portable Inside Compaq Portable Inside


Compaq Portable specifications
Manufacturer Compaq
Release date January 1983
CPU Intel 8088
Speed 4.77MHz
Ram 128k to 640k
Storage internal 5.25" 360k floppy x 2
Expansion 3 slots
Ports Serial, parallel, composite video out
OS Compaq/MS DOS


Compaq Dos (177.65 kb) (1.05 mb)


On getting it home and doing some initial testing, I was happy to get it to boot from the A drive with a DOS 3.3 boot disk, but on the down side, I couldn't get any of the keys on the keyboard to respond. No really being able to do much more at that stage I pulled it apart and and gave it a full clean out. On pulling apart the keyboard I discovered that is uses sponge foam as a spring inside the keys.  I've never seen these before, but in doing a bit of reading I found out they are used on several other models like the Sol20 and Apple Lisa.  It was pretty clear to see why none of the keys were working as both the sponge material and the Mylar on the key pads had totally perished (see photo below).


I read a great guide on the Solviant website about replacing keys on the Sol 20.  The guide suggested making your own pads using sponge discs topped with aluminum foil and clear plastic tape. I didn't have any sponge material on hand so I just used some padded double sided tape instead. I create a test key and tested it on the keyboard pcb and all keys worked. I then proceeded to make similar pads for all keys on the keyboard. This took a couple of hours but once re-assembled, I had a fully working keyboard. The keyboard fix was good as I can now use the machine, but I'm not happy with the feel of the keys. The double sided tape that I used does have some spring to it, but it's way too hard compared to the sponge material. So at some stage I'll redo the pads using a more appropriate material.

With a working keyboard I did some further testing on the machine and found that the B drive didn't read any disks. I pulled it out and cleaned it internally and also did a thorough head clean, on re-installing the B drive it now reads some disks, but it's not perfect. I've noticed it gets to a certain track on the disk and then comes up with read errors. I'll need to extract that drive again and check the rails for debris. I can launch some programs from that drive now so it's better than before. Everything else on the machine seems to work well, the display is good, there's some geometry issues, but the green screen display is very sharp and there's no sign of any burn in. The machine has the floppy controller/printer card, an AST SixPakPlus clock/memory/serial upgrade card installed.  Memory reports at 640k, I also replaced the 3v battery on the AST card and I found the AST clock utilities on web, the clock now saves the date and time.

Epson HX-20 Revival

I recently saw an Epson HX-20 for sale on a local trading site, what initially caught my attention was that it had an integrated printer and micro tape drive.  Very cool!  Not knowing much about these machines I did a bit of research on it and was impressed to hear it's considered the first laptop.  I generally like to collect machines that have some historical significance, and limit my main collection to stuff from the late 70's and early 80's.  It fit the bill nicely, so I snapped it up.

Epson HX-20
Epson HX-20 with case
Epson HX-20 lcdscreen


  Epson HX20 specifications
  Manufacturer      Epson
  Release date   November 1981
  CPU   Two Hitachi 6301
  Speed   614 kHz
  Ram   16k to 32k
  Rom   24k (3 x 2764, 8k eproms)
  Storage   Cassette internal 
  Expansion   1 optional 8k 2764 eprom slot 



  Rom basic


 120x32 pixel LCD display



SkiWriter rom   SkiWriter rom


As is quite common with these old machines it was untested and had no adapter, but it's condition was great and it came with the hard case and an extra eprom.  I was hoping that getting it running would be as simple as finding a suitable adapter, I have an adjustable multi-fit adapter which did fit but on initial tests it wouldn't power on.

From reading posts on various sites, I found that these machines require a functional internal battery to operate even with the adapter plugged in.  So I proceeded to take it apart and found not surprisingly that the internal NiCd battery had perished (see pic above).  Luckily the leakage of the batteries was contained within the plastic wrapping of the battery pack.  The corrosion had also made it's way up the battery leads and started to lightly corrode the connector on the motherboard, but the board connector was easily cleaned. 

I removed the battery and cut off the battery leads, they were totally corroded and could not be saved.  I tried removing the battery leads from the connector on the lead side, but they were stuck with corrosion.  So I soaked the connector in white vinegar for a couple of hours and was able to remove the corroded leads.  I made up new leads using the wires from a pc 3.5" floppy connector.  The floppy connector pins fit perfectly into the HX-20's battery lead connector. 

Prior to making up new batteries, I was eager to test whether the machine worked, so I connected my adapter directly to the battery leads and plugged it into the mother board. I set the adapter to 4.5v and the machine came to life!  The adapter I have is voltage switchable but is only 300ma.  There wasn't enough power to run the tape drive or the printer but I could run Basic and enter in a simple Basic program.  Three items were displayed on the lcd: 1 Monitor; 2 Basic and 3, there was an optional rom installed called Entecard.

I found the posts below which discuss making a new battery pack for the HX-20:

I managed to find the exact same batteries at my local parts store, so I made up a new battery pack and attached the new leads.  I partly re-assembled the machine with the new battery and plugged in the  adapter and was able to monitor the battery voltage.  I saw the voltage on the battery going up as the adapter was powered, so this was a good sign that the batteries were charging. 

I fully re-assembled everything and after a few hours of charging I was able to get both the printer and the micro-tape drive working!  The printer ribbons that came with the machine were long dead but replacements are still widely available.  The printer paper is also readily available as it just uses 2.25" cash register paper.  There was also an original Epson micro-cassette that came with the machine which still reads and writes ok.

Now that I had a fully functional machine I started looking for software that would run on it.  There's plenty of Basic programs available on various sites, but the main one I wanted was Skiwriter.  It then occurred to me to check the extra rom that was supplied with the machine, and it was an original Epson Skiwriter rom.  It was in a sealed plastic bag so it may have never been used.  I removed the rear plastic cover replacing the installed Entecard rom with the Skywriter rom. On power-up I was greeted with the Skiwriter option on the menu.

Skiwriter ROM

Skiwriter is a word processor application for the HX20.  The zip file below contains an image of the Skiwriter eprom in both .bin and .hex versions.  To use Skiwriter on your own HX20, create a 2764 8k eprom using either the .bin or .hex image file. (14.06 kb)

Skiwriter Manual

This is the only electronic copy of the Skiwriter manual known to exist.