Sharp MZ-2500 Upgrades


Amongst the machines I am currently renovating is a Sharp MZ-2500 or Super-MZ. It is said to be the best 8bit computer made with advanced graphics capabilities, 256KB RAM, automated tape deck, floppy disk, hard disk (optional) and a Z80 running at 6MHz.

This machine not only runs MZ-80B/MZ-2000/MZ-2200 software but makes a big jump in terms of video, audio, memory and expansion for its own software. I'm looking forward to seeing if I can further upgrade it with the tranZPUter or tranZPUter SW design to upgrade the Z80 CPU to 24MHz and provide SD services accordingly.

The MZ-2500 already has colour video output with 64K Graphics VRAM, Character VRAM and a PCG with Kanji support so it makes no sense to adapt the Video Module to this machine.

After working with the MZ-2500 for the last few months, the machine is very difficult to upgrade other than using an expansion card. There is no space for a tranZPUter unless I copy Sharp's 8086 expansion for the MZ-2000 and use a 40pin fly lead and site the tranZPUter elsewhere.

I thus decided not to pursue the tranZPUter upgrade path for this machine and just get to know it so that I can add an MZ-2500 emulation into my FPGA Sharp MZ Series emulator in due course.

Renovation Begins

This is the first (of two) MZ-2500 I’m working on, externally it doesnt look too bad but a long time ago someone stuck packing tape over it and it has aged and discoloured the plastic. Sheet

Side image where the packing tape terminates. Sheet

Internally the machine looks good, very little signs of rust which I’ve encountered on many machines from Japan. I’m guessing the sea air, like the UK, attacks the metal work when they are stored. As can be seen, the machine is fully loaded with all the options. It even has an MZ-2000 Kanji ROM card which seems out of place, I will put this card in my MZ-2000! Sheet

First expansion board, an MZ-1R37 640KByte EMM card. This card adds 640Kbyte additional RAM which is typically used for a RAM disk. Sheet

Next up is am MZ-1MIO board, this card adds a 4096 colour capability to the MZ-2500 using a colour palette. Sheet

Next is an MZ-1R28 board, this board adds Hiragana to Kanji translation dictionary. Sheet

Next is an MZ-1R26 128KB RAM expansion card, increasing the MZ-2500 memory from 128KBytes to 256KBytes. Sheet

Finally an MZ-1R27 64KByte Video RAM expansion card, increasing the MZ-2500 video RAM to 128KBytes. Sheet

First stage of renovation after strip down was to attend to the plastics. The metal work is generally fine, a few scratches which are part of the machines character (a monitor no doubt once stood on the machine top) but the plastics are ghastly.

It took quite a while to remove the packing tape, it had set like super glue and after removal and full clean, the discoloration wasnt, to me, acceptable. Where the packing tape was located the machine was almost original colour and everywhere else was varying shades of yellow. Not something I would have prefered to do but a little retrobrite had to be applied to get the plastic to a uniform colour. It is still a slight yellow but uniform across the whole front bezel now. Harsher retrobriting would have brought the front bezel to original white colour but I considered this and wanted the machine to have some of it's character rather than a new out of box assembly clone.

Front Bezel after renovation. Sheet

Lid, only needed a good clean. Sheet

Side, the power switch was very discoloured so I brought it into line with the front bezel. Sheet

Opposite side, just needed a good clean. Sheet

Rear panel, amazingly in very good condition, the fan was black with dust and soot but a good clean sorted the panel out. Sheet

Second stage of renovation was to attend to the power supply. I'm not a great fan of the old switch mode PSU's as they were a black art, more modern devices using control IC's are preferable. Consequently I made a conscious decision to replace bare minimum parts and test the critical capacitors.

After removal of the PSU it was fully striped down and cleaned in a bath of Isopropyl alcohol, using a paint brush and tooth brush to clear and dislodged the years of 'electronic killer' gunge! The switch and external power socket were cleaned in Isopropyl alcohol and then with contact cleaner. The metal work was washed down in Isopropyl alcohol, cleaned up and then fine rust oil rubbed all over it to prevent the onset of rust.

I then set about desoldering and checking the electrolytics, replacing any which were out of tolerance by +/- 10%.

Finally, power up and check each of the voltage lines for noise and value.

Pictures below are from the 2nd unit under renovation, I forgot to take photos of the first machine as I was so engrossed!!

Sheet Sheet Sheet

All internal metal work and bottom casing were cleaned as above, ie. with Isopropyl alcohol and treated to a wipe down with rust oil to prevent degradation. The second machine being renovated is Very rusty so additional soaking in rust penetrant was necessary.
Next up was to attend to the motherboard and expansion boards and replace all tantalum and electrolytic capacitors. Tantalum had to be replaced, plenty of examples where one has gone short circuit and taken out circuitry, some of which, ie. custom VLSI IC's are not replaceable. The electrolytics where replaced as a precautionary measure as I've encountered a number on previous machines which have gone open circuit or lost their capacity, for example the audio section on an MZ-80A didnt work due to failed capacitors.

Another item which needs attention is the NiMH battery. On some machines they leak and damage the PCB often leaving the machine unserviceable. Luckily on this machine it was intact (Brown cylindrical device in bottom left corner below) so only required a replacement.

I didnt take photos of the mainboard, before/after, so below are just some mainboard photos for reference. The orange components are the tantalum bead devices.

Sheet Sheet

At this stage I connected up the mainboard to the PSU and keyboard then powered on and a rush of joy was felt after seeing the 'loading error, press C..' message. Stage one was successful.

Ever forward and onward, the 3.5" floppy drives.... once serviced I can boot a game :)

Servicing of the floppy disks amounted to disassembly and clean of all the metal work as above, cleaning of the control boards in Isopropyl alchol (the main control board) and gentle wiping and cleaning with a cotton bud of the head and mechanism. I lubricated the head control mechanism then put everything back together half expecting to delve into the wonderful world of floppy head alignment or track zero sensor replacement.

Connection to the mainboard Floppy daughter card and PSU, a prepared floppy disk (using Kryoflux) and power on.... The pleasant familiar sound of a healthy floppy disk seeking and stepping and the startup screen to Xevious... all good. I tried numerous other floppies, ie. Fenix, CP/M etc and all worked fine.

Unfortunately, the second MZ-2500 machine floppy drives had succumb to rust so this will require a lot more hard work, for now something to do on a rainy day! Maybe I can get some photos as I didnt take any of the floppy drives!!

The part I had been dreading, after numerous hours working with them on the MZ-2000, the Cassette Drive.....

The cassette mechanism for the MZ-2500 is basically the same as the MZ-2000 but with stereo heads (as it can playback/record audio cassettes and telephone recordings) and more control circuitry. It comprises of 2 control boards, whereas the MZ-2000 only has 1.

As found out on the MZ-2000, the mechanism doesnt age well, the pictures below are from the second MZ-2500 I'm renovating, which externally looked better than the first, but inside was rusting and decaying.

The ringed items are the main cause for concern, they are nylon cogs and with age start to crack where they are under greatest stress.

A full strip down is essential, right down to disassembling the head and drive modules. After washing with Isopropyl alcohol the first stage is to identify cracked cogs and repair them with plastic cement. I was at one stage going to get these parts 3D printed but decided against it. Another point of issue is the pinch roller, this degenerates and becomes like tar. Luckily you can still buy pinch rollers from China, so a quick order of 13x8mm pinch roller with 6mm rubber height and 2mm rubber diameter sorts this problem out.

Once the cogs are repaired, you need to set about greasing all the friction points, this part is very necessary as any friction, no matter how small can lead to some very interesting malfunctions. The flywheel top and bottom spindles need greasing, the solenoids need a little lubrication and the motor may need a drop of lubrication oil, on the 2nd MZ-2500 this was necessary due to rust.

The belts, if they have a deformation in them (ie. not round) they need replacing. The main belt is generally fine (ie. motor to flywheel) but the two smaller belts need replacing. The tape counter belt may need replacing if it has stretched or kinked, this belt is important because the tape counter sends pulse signals to the controller so it can determine if the tape is in motion, no pulse and it will stop the motor and disengage solenoids.

Re-assembling the cassette mechanism you can manually turn the flywheel and engage each solenoid to verify its smooth action and there should be very little additional force required on the flywheel during actions such as engaging the tape heads, and force means something hasnt been lubricated. I reassembled the complete tape drive and decided to test it, rewind, ffwd, play all worked fine, in MZ-2500 mode I could even play a music cassette, just one thing didnt work and it was the most important, the IPL couldnt see or interact with the cassette drive, either to eject the tape or engage play/rew/ffwd. This was also the case on the 2nd MZ-2500 and after researching some Japanese sites, a common fault but no cure was offered. I spent an age probing the 8255 and buffer drivers but couldnt find a cause, if I manually forced a signal active then the cassette drive would respond, ie. play. I then decided it was the input signals showing the CMT status but these 'seemed' in order albeit the high/low logic levels werent definitive. In exasperation and something I was trying to avoid given the sheer number, I set about replacing all the capacitors in the CMT control boards, a long task! After re-assembly, BINGO! the drive could now be seen and the IPL would correctly engage and boot off the cassette.

These pictures are taken from the 2nd MZ-2500 I own and comparing with the first, this machine was one of the first to come off the production line. Many manual corrections on the motherboard and additions to the tape drive, such as the F2181AF-1 card including point to point wires across the control circuits show it to be a first production or pre-production unit.

Top view of the first MZ-2500 cassette drive. Sheet

Side view of the second MZ-2500 cassette drive showing the circuit boards stacked on each other. Note the rust on the motor, this machine spent it’s days by the seaside I’m guessing! Sheet

An important point during disassembly, the electret microphone is below the cassette draw damper and control board, you need to desolder the two fine wires before disassembly else you risk damaging it. Sheet

The tape counter control circuit, this is important as it informs the microcontroller that the tape is turning. If no signal is sent it assumes a tape jam and stops the motor and disengages the solenoids. Sheet

The partly diassembled mechanism of the 2nd MZ-2500, you cant see the cracks but one exists on the clutch and some plastic c-clip equivalents which keep the cogs/wheels in place. Sheet

First level circuit board, similar to the MZ-2000 control board but sited on top of the mechanism. On this one there is additional wiring and components not seen on the first leading me to believe this was a first or pre-production unit. Sheet

The motor and rear side of the cassette mechanism before disassembly. Note the additional F2181AF-1 board which doesnt appear on the 1st MZ-2500 or the MZ-2000. Note the rust on the motor casing but underneath rust is prevalent on the flywheel and other metalwork. Sheet

Putting it all together, I was able to test the machine fully. I tried the MZ-80B mode and it worked beatifully, I tried the MZ-2000 mode and it wouldnt boot, it would 50/50 detect a tape image and when it did, after reading the tape header would stop then restart and would never complete the boot. I tried the MZ-2500 mode and BASIC M25 etc would just hang. Took a while to determine what was the issue, so one Caveat Emptor to other MZ-2500 owners who have boot problems, repair your mode selection switch. If you test the boot mode switch you will notice the resistance in the 3 settings may differ, they should all be circa 180R but some go upto 500R at least on my module. I also had a crackly speaker, this was despite washing the module in Isopropyl alcohol and using contact cleaner. It took a lot more cleaning before the switch functioned correctly and the crackling on the speaker stopped. Seems this module is more susceptible to oxidization.

The finished main unit, all cleaned with rust prevention and capacitor replacement. I clean up all the wires and connectors as well as generally there is a black oily gunge over everything, a simple wipe with a tissue will evidence this. Sheet

Side view during reassembly. Sheet

Final attention was turned to the keyboard. I completely disassembled it including the key caps, cleaned it in Isopropyl alcohol (it was filthy including 2 girls hair clips) and I used contact cleaner in each individual switch tapping them multiple times to allow the contact cleaner to work. The keyboard had a bit of damage on the left side, seems it may have been put up against a radiator or similar as there are about 6 small cylindical indentations but other than this, the keyboard looks perfect and fully works. Rust oil needed to be wiped on the metal frame as rust has started to show so prevention at this early stage should see the keyboard lasting a lot longer.

The photos below show the keyboard just after disassembly.





Finally, the gallery of working shots. This machine is a joy to behold.

In order to test the various modes of the machine I made cassettes using the TZFS extensions I’ve recently added which allow reading/creating all Sharp MZ tape formats on an MZ-700 or MZ-80A tape deck.

To create the floppy disks I used Kryoflux with the following parameters:

dtc -d0 -dd1 -wv1 -w -fsuper\ mz\ demo\ 1_d88 -e79

It is necessary to convert the D88 format image to Kryoflux raw stream image format before writing to floppy.

MZ-80B Mode - boot BASIC SB-5510 from tape/CMT unit. Sheet

MZ-2000 Mode - boot BASIC MZ-1Z001 from tape/CMT unit. Sheet

MZ-2500 mode - boot FENICS from floppy disk. Sheet

MZ-2500 mode - boot BASIC M25 from floppy disk. Sheet

MZ-2500 mode - boot Super MZ Demo from floppy disk. (see video for demo running). Sheet

MZ-2500 mode - boot Xevious from floppy disk. Sheet

MZ-2500 Demo movie running the SuperMZ Demo Program

This machine I believe was stored for a long time probably in a shop hence the packing tape which was probably a sales leaflet banner across the front. It came with all the original manuals, a Japanese monitor video cable, keyboard overlays and stands, power cable and 2 of the 4 original disks, the missing ones would have been the BASIC M25 and BASIC S25. Sheet

The machine also has its original packaging, the top layer containing the keyboard, manuals and cables and the bottom layer the machine itself. Sheet

The original box, some luck person on floor 2F I guess use this machine back in the 80’s/90’s. Sheet

Next Steps

I bought two MZ-2500 machines thinking I could get at least one working, they both looked in good order from the auction photos but as seen, the 2nd machine is in bad condition, this is the risk of auctions where items are sold as 'junk' and expensive for this particular machine! The mainboard and PSU on the 2nd machine works but it seems that someone scavanged it as there is no expansion card frame (unless these were optional extras, as I found out was the case on MZ-80B's expansion frame recently) or expansion options. The floppy drives are in a bad state and both need work and realignment, worse case is I have to adapt more modern 3.5" drives if possible, the cassette mechanism needs a lot of work.

It was always the intention to sell one of the machines if I could get both working, the first machine is perfect and is more a collectors item, the 2nd has a working motherboard and suitable for testing and development projects but not really for sale other than 'junk'! My thinking is thus to sell the first machine for the right price, I prefer to keep it for myself but if a collector is interested then I'm willing to part with it. That being said, it is dependent on getting the 2nd machine to a useable state, if this is not achievable then I will keep the first otherwise it means buying a third MZ-2500 from Japan (when they come up for auction) and taking another risk that it will work. Watch this space as I will add the 2nd machine renovation details in due course.


Where I have used or based any component on a 3rd parties design I have included the original authors copyright notice within the headers or given due credit. All 3rd party software, to my knowledge and research, is open source and freely useable, if there is found to be any component with licensing restrictions, it will be removed from this repository and a suitable link/config provided.


This design, hardware and software, is licensed under the GNU Public Licence v3.

The Gnu Public License v3

The source and binary files in this project marked as GPL v3 are free software: you can redistribute it and-or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

The source files are distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see