Bygonebytes



By January 2020 I'd revived my Acorn Computer collection to the point where it's up, running and stable with all the basic add-ons I wanted. There is still a lot I'd like to add or alter but for now I can start looking through all the data and software at my leisure. Things still to do are: I would like to get a working Stop Press 64 board in the Electron which may a tall order, also I'd like to add a couple of switches to the Master to switch between my ROM board and the internal Sideways RAM, upload the Gerbers for the TTL RGB Switch and BBRAM board, complete a 3D case for the TTL RGB Switch, add a User Port to the Electron and so on..but for now I have a clear work area to get started.

Electron setup:

Electron Issue 4 with Plus 1, Master RAM Board, Retro Hardware AP6, 2 off 32 banks of Battery Backed RAM, Retro Hardware ATI, RPi 2nd Processor, PRES AP3 & 4 Floppy Disc Interface, 2 off Gotek floppy disc emulators.

BBC B+:

BBC B+ with 64K RAM upgrade (BBC B+ 128), ATPL Sidewise + ROM board, ZIF socket, RPi 2nd Processor, 2 off Gotek Floppy disc emulators.

BBC Master:

BBC Master with 32-16 ROM board, 2 off Gotek Floppy disc emulators.

Common to all systems:

5 input/1 output TTL RGB Switch and 15" LCD colour monitor with TTL input.

What follows is my five month journey from the inspiration to unpack my kit to having three stable and upgraded systems to work with.



Introduction

When searching for something Raspberry Pi related but nothing connected to Acorn computers I stumbled across a reference to a Raspberry Pi second processor. Bouncing around a few links brought me to Sundby.com which sold the PiTubeDirect signal converter and links for the Raspberry Pi 2nd processor emulator software. All this gained my interest so I opened my retro museum and unpacked loads of my Acorn stuff - This lot of computers to the right.
It turns out that my interest in the Raspberry Pi had led me rather belatedly to the Raspberry Pi Second Processor emulator project for Acorn BBC/Electron computers, my other obsession! So I thought that I'd buy the interface and give it a go. I ordered the Sundby PiTubeDirect interface external kit which I'll use with a Pi Zero. Learning of this project this rather late in the day made it fairly easy to put together as it's well documented, anyway I'm getting ahead of myself, I had to find out if my collection of Electrons, BBC B+ and Master were still working after being in storage for such a long time. First was the BBC Master..




Quick Links: Master | BBC B+ | Electron | Disc Drives | RGB Switch | Battery Backed RAM | UVIPROM EPROM

BBC Master 128
A week later I switched on the BBC Master for the first time in many years, needless to say the batteries were exhaused and therefore the configuration lost. The batteries were dated 2001! amazingly after 18 years they hadn't leaked. As it's been a while since I've used an Acorn machine I had to re-aquaint myself with the *configure command and now it's all working perfectly.

Since the Master was OK I re-manufactured my 32-16 ROM board from the '90s so I could install my inter series ROMS inside. More info on the BBC Master page.




The Master 32-16 ROM Board Gerber and Schematic files.

BBC B+ 128
Another few days had passed before I got the chance to switch on the BBC B+. All was well for the first ten minutes then I could hear a faint buzzing sound, not long after the dreaded smoke genie appeared! I think I was quick enough to switch it off before there was a lot of damage. The smoke came from the power supply area so I disconnected the mains supply and removed the PSU and striped it down - the fault was fairly obvious, the suppression capacitor C2 had smoke damage as you can see in the photos. I've ordered a replacement and we'll see what happens next..

After a quick look round the internet it seems that it's a good idea to replace the smaller 10nF capacitor C1 as it's the same type. Looking at the capacitor it is showing signs of age/stress/swelling and I don't want to dismantle the supply again so it'll be replaced..

A few weeks later I received the parts and replaced both suppression capacitors, a tricky task as once the capacitors have swollen it's difficult to get them out due to the tightly packed neighbouring components - you can see the problem in the photos. I put the power supply back in the BBC case and before connecting up tested each of the power lines - all was ok. So that's the BBC B+ back up and running - good for another 40 years!

On getting the Raspberry Pi 2nd processor up and running on the Electron I thought I'd build one for the BBC B+. I've got no space to put my Acorn 2nd Processors on my desk and with the Pi 2nd processor being able to fit under the BBC it would be prefect for me. I have had some experience soldering surface mount components but I'm never very happy with the results so needless to say I am not keen on using surface mount devices. I thought I'd simplify the interface by making a Level Shifter PCB using standard DIL devices. Even when using the larger devices it still fits neatly under the BBC.

I said it would last another 40 years..it hardly lasted another 40 days until it wouldn't power up properly! Removing C9 it could be seen that it had swollen on the underside so I replaced it then rebuilt & tested the supply. All is well again but for how long...










Electron
Unpacking my box of Acorn computers revealed two Electrons, three Plus 1's and a Pres Plus 3/4, when tested the Electrons and the Plus 1's worked ok except for the disc drive PSU and three keys on one of the Electron keyboards.. The Electrons are issue 6 and 4 and the three Acorn Plus 1's are all issue 2's.

Edit: end of 2019, when looking through some boxes for my Marconi trackerball I found a third Electron / fourth Plus 1 and more data disc's. They were hidden amongst my Amiga 1200/600's. Still no sign of the trackerball though.. Opening the Electron shows that it's an issue 6 and has had a few modifications added, two are around the composite video output, one on the cassette interface and a piece of foam/cardboard placed in front of the speaker to reduce the volume. They won't affect my use of the machine so I'll leave them as is. Switching on revealed nothing unexpected, the Plus 1 does have a copy of the PRES Plus 2 ROM though. I think this Electron would have been given to me when I was involved with the EUG..there was a lot of trading of computers and parts back then so much so that an EUG member, whilst on holiday in Scotland, randomly turned up at my door and bought a disc drive! There are some photos of this Electron below the MRB photos along with some ROM cartridges.

I replaced the PSU and used switch cleaner on the keys so everything is now up and running. My aim for the PiTubeDirect & Raspberry Pi Zero is to get a stable 2nd processor working on the Electron. Years ago I bought a Slogger 2nd Processor which worked fairly well but over time became too unstable to use. I only used the 2nd processor for large documents written in View and occasionally with Starword, I eventually replaced it with the Master RAM Board. I no longer have either of these add-ons so the 2nd processor emulator is an interesting replacement.

To run the 2nd processor emulator on an Electron it requires a TUBE interface. Research brought me to the Retro Hardware ATI which not only has the TUBE interface but two 16K sideways RAM banks which will come in handy for running View. I assembled the PiTubeDirect interface which was a bit fiddly due to the surface mount components. Care has to be taken mounting the 40 pin headers as it's easy to solder them in the wrong way round. Also when creating the SD card the instructions state that the card should be formated 'standard FAT' which I took to be FAT16. This can be done but it's a bit tricky with cards greater than 4Gb so I also tried FAT32 which works just fine. With everything assembled it all worked first time, I have run the test programs but yet to try View etc.

Just had five minutes to try an old EUG disc I found, issue 18, the opening screen of the EUG being munched by a dot - what an amazing speed increase with the second processor! I wasn't expecting much of a speed increase with View and I wasn't dissapointed, the only upside is that 30462 characters are available in any screen mode. So far nice and stable..

To help bring my Electron back to what I used in the early 90's I've fitted an AP6 ROM expansion board to my Plus 1 and a Master RAM Board (MRB) to my Electron issue 4.














Disc Drives
Each machine has it's own disc drive(s), the Master has twin 5 1/4" drives, the BBC B+ a single 5 1/4" drive and the Electron a 3 1/2" drive. There is a disconnect with this arrangement which is that the Electron is isolated, I cannot transfer programs/data between the Electron and the other two computers. I decided to solve this problem by adding a 3 1/2" drive to the BBC B+ giving it both sizes of drives.

I found an old 3 1/2" drive and ribbon cable from by PC spares (junk), a USB power supply and made up a USB to Floppy 5V power cable. The cable was made up from old PC drive 'Y' cable - two molex and one floppy connector, a resolderable USB connector and some 2-core 3A mains cable. I cut the floppy connector off leaving plenty of wire, soldered and sleeved it onto the 3A cable then soldered the USB connector to the other end. The use of the fairly heavy cable is to ensure minimal volt drop accross the cable when the drive is taking peak current ie when the stepping motor is stepping. Connecting the drive to the Electron PRES Plus 3 with the ribbon cable and plugging in the usb power adaptor for testing, no problems were encountered and as the drive defaults to drive 1 this makes it easy to connect to the existing BBC B+ drive which has its drive set to 0 - no link swapping or cable core twisting was required. All I need now is a 34 way IDC connector to press onto the existing 5 1/4" drive ribbon cable.




And to add to the mix is this GOTEK Floppy drive emulator - I intend to plug it into the BBC B+ along with the 5 1/4" drive so I can copy years worth of documents that have been stored away since the 80's. Then it'll move around the computers as required.



Having spent a day using the GOTEK I will consider replacing the old disc drives with these units. I've been copying data onto the GOTEK from floppy discs that have been stored in drawers for around 35 years..1984 or even earier! One paricular make, Parrot, hasn't stood the test of time but they have all suffered in such a way that the disc struggles to spin, probably due to the disc surface swelling. I think I've worn out the heads on at least one drive! The photo above is the pile of about fifty discs that are no longer readable, thousands of man hours of work gone.

I have decided to replace all the disc drives after I've transferred the data from my 5 1/4" and 3 1/2" discs on to the Goteks USB drive. The task of copying is moving along slowly, I've completed the 3 1/2 disc transfers which were mostly ADFS - *Treecopy made the process simpler. I've collected over the last few months enough Gotek drives which I have upgraded and flashed with FlashFloppy 2.13. I followed a couple of the guides available on the internet that detail how to update the firmware and to add the OLED screen, buzzer and rotary switch. It turned out to be a very easy process so the Gotek drives are ready to replace the old drives. I'll start by swapping the Electron drives with each machine getting two Gotek drives which will give them drives 0 (and 2) and 1 (and 3). The photos show the drives being tested.




TTL RGB Switch
Working with several machines at once with only one monitor is a bit of a chore, constantly swapping the RGB cable over is not ideal so I've decided to resurrect another old project from the early 90's - my RGB switch. The original switch had two inputs switched to one output, the new one has five inputs letting me easily select which computer to display. The output selection is by rotary switch. The photos below show the build from board layout through testing then mounted and in use.

While testing the upgraded Electron I used a small program I'd written to test the RGB switch which highlighted a fault on the Green output, photo below, it should be a clean green screen - no interference. Having already tested the RGB switch I discounted this and thought that the fault would be internal to the Electron, somewhere around the ULA, either the chip or a tarnished socket. I looked out a few spare ULA's and tried them with no change so I'm forced to consider the fault is external to the Electron. With four computers plugged into the switch I tried removing them in various combinations till I found what if anything made improvements. Some combinations made it better and others made it worse but none made it perfect so my next move is to look at the diodes and individual HC125's on the switch.

With the new six input version built (see below) I tested it with all combinations of inputs and it works perfectly so I think I'm on the right track. The difference between builds are that the diodes on the six input switch are 1N4001 instead of OA47's so I'll swap them first and if necessary I'll then look at individual HC125's. The OA47's may be forming part of a tuned circuit picking up some high frequencies..

I replaced the diodes which fixed the problem. I think my theory was correct that the germanium diodes were picking up noise from the switched mode PSU's in the BBC, Master and monitor.

Having used the five input switch fairly extensively over the last few months I decided to make a couple of changes to the layout. The changes have come about partly through feedback and use. The new switch now has six inputs and they're grouped closer together, the sockets are designed to be cascaded thus leaving room for a sixth input which natuarly fits the six way rotary switch. These changes hasn't affected the overall dimension of the PCB. I will now get on and see if I can design a suitable case.






Battery Backed RAM modules
Now that I have the new AP6 board in my Plus 1 I thought I'd resurrect another module from the past - the BBRAM128. This module gave eight 16K RAM banks, one at a time. Bank selection was via a BCD thumb wheel switch. At first I thought that just a professionally made PCB would give this module the facelift it needed but I came across a the AS6C4008 SRAM chip which would give me thirty two 16K RAM Banks. - a ridiculous amount!

When designing this new prototype I ended up making quite a few changes, I removed most of the pull up resistors on the data & address lines, put in a smaller battery, increased the RAM size and reduced the board size. The only component that was difficult to obtain was the OA47 germanium diodes, I found equivalents but they are getting quite expensive.

To test the board I had a couple of 128K x 8 SRAM chips left from the '90s but I also ordered some from China. The old chips worked and the new ones didn't..maybe they need the pull-up's? So when it comes to the AS6C4008 I have ordered some from CPC-Farnell and a couple from China to see if I get similar problems. At this point I thought that I could use one of the 32K sockets on the new AP6 and switch two ROM's in at a time, also that there should be a newer method, one with a smaller component count, to feed the battery power to the SRAM chip. Looking around the DS1210 Nonvolatile Controller Chip seems to be the most popular but it is also now obsolete but seems to be readily available so the board is away for a second prototype. I've added some pads for SIL pull-up resistors just in case they're needed and a link to the board so it can be used in either a 16K (32x16K banks) or a 32K (16 x 32K (2x16K) banks) socket. The advantage of using the 32K sockets in the Plus 6, 5/6 or 13/15 is that you can use software lock/unlock instead of a switch to make the RAM read only.

The AS6C4008 (512Kx8) SRAM chips arrived from CPC and tested out ok with my first prototype board so I can now store and select 32 ROMs at a touch of a button..well two! I don't think I've got 32 Electron ROM's!!

The second prototype board is now built and tested and works well in the AP6 ROM sockets 5/6. I think I'll leave both boards in the Plus 1 at the moment and mount the selection switches to the right hand side in front of the cartridge slots.

The old design with the small battery didn't last long, about four weeks before the voltage fell below the threshold to hold the memory content so I've abandoned that design and I'm just using the DS1210 design. Still soak testing, one configured 16x2x16K in socket 5/6 and the other 32x1x16K in socket 14. I have closed the lid on the Plus 1 and added photos of the cut-out and switch mounting. All nice and neat now and I've added a few ROM images while I gain confidence that the batteries will last.

Just got round to fitting the read/write switch for RAM position 14.







UVIPROM EPROM Programmer
The UVIPROM EPROM programmer is a great little programmer, I bought it many years ago and modified the accompanying ROM software to work on the Electron. The software was written for the BBC/Master Mode 7 which when used in the Electron with its mode 6 it messed up the screen formatting, I altered this and a few other things to get it working for the Electron. The programmer was so well used that the Vcc switch has just about fallen apart so I simply replaced it with an equivalent switch. I used the programmer to program the 27C010 (or equiv.) EPROMs for my ROM128 boards via a ZIF adapter I made - a bit rough and ready but it worked! I currently don't have a User Port (yet) for the Electron so I used the BBC for testing.




Historical Letters
I have more or less completed the disc copying to the gotek USB and have un-earthed a few letters which date some changes to Electron upgrades I was using and also a bit of a rant to Micro User over the constant discontinuing of their publications.

The AP2 ROM - on the 27th September 1988 I returned the AP2 to PRES which I'd bought on 5th July of the same year listing a few incompatibilities with my Electron setup. I stated that the disadvantages of interfering with the system outweighed the advantages. In hindsight I'm not sure how many of these disadvantages would stand up to scrutiny today!

Stop Press 64 - On the 8th January 1992 I replied to Andrew (Slogger) returning two EPROMs, the MRB OS3.00 and a previous release of Click. I had just received the new Master RAM Board OS3.10 which enabled me to get SP64 up and running, however, this revealed another weakness in Click - certain actions within SP64 invoked Click thus I had to turn off the mouse within Click before using SP64. Another version of Click would be on its way..

Micro User - During 1992 Acorn magazine publications went through radical changes and in September 1992 I got a bit frustated with this and wrote this letter. Needless to say it was never published.