The Ace Place

While I'm waiting for the ACE3 board to comeback I thought I'd have a look at my plans for ACE4. This should include pageable RAM, the costs of RAM mean it is more expensive to fit the small RAM chips than go for a bigger one. While the Z180 has a fully-fledged MMU which can address up to 1 megabyte and map it, I'm going for something a bit simpler and that is to have the top 8k pageable. The circuit above is my current thinking. It should work as follows. For the bottom 56k, the #PAGED8K line should be high, this will force chip Z106, a SN74AHCT541N , active, which will put the normal A13, A14 and A15 address lines on to the memory chip bus(MA) while pulling MA16-20 low, thereby addressing the bottom 56k directly. If however #PAGED8K is low the chip Z105, a SN74AHCT574N ,  will be active and it will place on lines MA13-20 the data previously written to the chip, this will mean that any of the memory in the RAM chip can be accessed via this 8k block even that mapped to

I've sent off the order for the PCBs to make the latest development version of my Ace rehash, it should take about 14 days to get them back. All I have to do know is fund the chips to go in it, I've got a lot the Z80 and KIO and some of the 74 logic. There have also been some changes to the IO ports for the Keyboard and mouse they are now fully clamped down to just one address rather than taking up all the even IO, 48K a RAM is available. This fil with design and all the data is on GitHub here . The CSV file contains enough data to be loaded directly into Mouser's BOM system . Though some of the chips notable the ZIO are not available and neither are some of the transistors. One or two of the components can be obtained cheaper either on eBay or with Cricklewood .  This is untested and less than no guarantee is given. Delay Well, it was all going so well right until the transfer to Royal Mail for the final leg, the Chinese courier thing Royal Mail have t

In order to get some of the software working in advance of the hardware being ready and due to the lack of debugging support on the hardware I'm tying to rejig a simulator to provide some support. After trying to get several to work I ended up with CHIPS   which allows me to play with the driving the CTC, which I'll need to get working to support the 50hz clock. While none of the exiting ones is is perfect, the LC-80 emulator provides a good base, you can see them all at Tiny Emus . I made a simple ROM which is below. To get it to work the source is assembled to chips-test\examples\roms\lc80_2k.bin then the  chips-test\examples\roms\lc80-roms.h deleted at the project make is run. I am using Visual Studio so I thin make the individual project and run it in debug mode. To check that it is working some bytes in the RAM as 0x2000 are updated. CH0 equ 0h CH1 equ 1h CH2 equ 2h CH3 equ 3h IBASE   equ 200h CTCIBASE equ 0h SIOIBASE equ 8h  .org 0000h     di                    

One of the things that have changed between the original ACE and these models is that there is no RF Out, the UM1233 was used to give access to analogue TV stopped being made some time ago and so it wasn't practicable to include one in the design.   Fortunately the input to the modulator was straight composite, so in theory it just need rerouting into the composite socket of  a modern TV and this is indeed the case but it needs a bit of amplifying cation in order to work, so I have added the circuitry below, complete with a variable resistor to make sure it is not too bright. It is basic 1 transistor amp and the transitor already used in the crystal circuit(Q1) the image seems very stable and sharp overall an improvement on the original image. I may do a similar amplification on the audio and output that for use via the monitor/TV speaker.

The ACE2NOKBV2MK2 design is approaching completion soon I'll be sending off for test PCBs, the shape of some of the components might look a bit odd and the schematic has more components than needed. Not the bumps on the bottom of the biggest DIL chips. This is not an accident they are 2 chips overlayed. The 6116 chips come in two size the old HM6116LP variety is 15.24 mm wide and slow with access time between 200-120 NS, all of which are faster than the original ace chips, while the newer ICT6116 are a slimline 7.62 mm, faster and more expensive. From the underside, you can see I've parked the outline of the wider chips in red and in one case the narrower in blue. You do not have to fit both one but I'd be consistent and fit either. You can see there is a Turbo option this is simply the modification from here . I'm not sure the slower chips will work with the Turbo option enabled, in fact, I'm not sure it will work at all. In addition to the change

I was having problems with the test leads falling out from my Experimental Keynoard/PCB so I order so om KeyIF   PCB and though simple it works a treat.

For the basic system, all the chips are available from major suppliers, though some are not available in the LS format. The table below gives details of the speed and power requirements, it is from the Digital Logic Family Selection Matrix . Any technology that is as fast or faster the LS should work provided it is TTL level compatible, this is usual signified by the last letter in the abbreviation being T. As you can see LS technology is slow and consumes a lot of current the AHCT technology is faster and consumes about 0.04% of the power, which should allow the Voltage regulator to run cooler, it also seems to be a bit cheaper so is the one I would recommend. You should be able to mix the various TTL technologies but out of an excess of caution, it is best to stick with just one. Mixing faster a slower should not be a problem at the base speed but if you try and overclock you may run into a problem with the slower technology. A couple of the transistors are easiest and cheap

While I would like to get some boards made the main thrust of this that people can do it all themselves. I've been using JLCPCB to get my boards made, though there are several other manufacturers that will do the same job. To get the first boards for ACE2NOKB I used Eagle CAD and it cost me £7.44 for 5 boards, delivery is extra but should be less the £6. The boards seem to be good quality and this has been confirmed by other more experienced engineers than I. The process of ordering using KiCad is very simple, simply use the function to create Gerber files, then Zip them up and upload them. It is all described on the site. One other site I found allowed the upload of just the kicad_pcb file. There are several mistakes with tat board ost noticeably that I forgot to distribute the smoothing capacitors, others include the wrong pin ordering on a transistor and some very strange placements. The current version of ACE2NOKB for KiCAD has those problems fixed and has a single tr

The main thrust of my work at the moment it on ACE3NOKB   which includes some significant changes to use more modern components and fully populate the RAM as well as make some space for extra ROM. Included on the board is a Zilog KIO Z80C90 chip. This means that there will have to be a change in the interrupt structure which requires the extra ROM. It is all here but is ongoing work so less than 0 guarantees are provided. As you can see I've moved to PLCC for the Z80, which is the larger of the 2 brown sockets. The smaller socket contains NOR-Flash memory which will help in development by allowing me to update the firmware. The large empty socket is for the Zilog KIO this provides 2 serial channels, 24 GPIO pins and 3 CTC, one of which will handle the 50hz interrupt this will entail a move to interrupt mode 2 . To create space for the extra ROM I've moved the start of the system RAM up 1k so that the 4k between 12k and 16k can be used. The 2 chips next to the ROM socket a

The lack of a keyboard and edge connector on most of the boards is to allow them to be placed in small boxes with those things added via the pin headers provided. There are some designs for the separate boards that will be needed. If you need a keyboard and want to make one with Keyboard switches from Cherry then this Keyboard is what you need. It feeds into a pin header which can be connected to the pin header on the main PCB this should be common to any of the designs. Keyboard Interface If you have a Keyboard and switches that you can heavily modify then the small board you need this KeyIF . Once you have modified your keyboard solder the wires into J1 and use a ribbon cable to connect to the main PCB. Edge connector For the Edge main connector, it is the Edgeboard you need. This like the other boards it is connected by a ribbon cable and provides the same edge connector as the original ace.

In addition to the cut down ACE2NOKB   design, I've also created something a bit closer to the original in ACE1 though it uses real keyboard switches. It has the edge connectors and only the changes need to make the output composite compatible and to cope with a lower impedance speaker. Though I have not built it the core of the system is the same keyboardless version. It should be compatible with all the hardware packs. The ICs are all in approximately the correct position, though the routing has no connection to the original.