Sinclair ZX Spectrum 128K

This is a ZX Spectrum+ 128K, nicknamed “toastrack” because of the shape of the prominent heatsink on the right side. The history of this model is quite interesting and would certainly fill a blog article of its own.

The toastrack is quite rare, but this machine here is even more special. It is an Arabic modification made by Sinclair Egypt, with a modified Arabic ROM, and Arabic letters on keyboard stickers. A switch at the front selects between the original Sinclair 128K ROM, and the modified Arabic ROM.

Sinclair Egypt modified other ZX Spectrum+ models in a similar manner, with a stacked ROM inside and a switch at the front. There is even a known ZX Spectrum +3 modificiation where the ROM could be selected by a boot menu instead of a physical switch. However it is unclear how many have been produced, and if they were actually for sale or just a proof-of-concept.

Configuration

• Built: 1986-02-20
• Mainboard: Issue 6U
• CPU: Z80A (date code 8605)
• ULA: 7K010E5 (date code 8540)
• Sound: AY-3-8912A (date code 8607)
• RAM: 128KB
• ROM: 27C512 EPROM containing the Original Sinclair 128K ROM and the Arabic v.3.1

Rarity

It is not known how many modified models were produced by Sinclair Egypt, and if they were ever sold to the general public. However, considering the rare sights, it must have been very few.

Trivia

• This model has an Arabic modification. Some of the keys have labels with Arabic letters, and there is a switch at the front that switches between the original and Arabic operating system. There is also an Arabic inscription made with a white permanent marker: “A gift from the Alumni Association”.
• Save from the switch and the inscription, the machine is in an excellent optical condition.
• It is a semi-official product made by Sinclair Egypt. The Arabic boot screen reads: “© 1987 Sinclair Egypt, presented by the Matsiko company, Sinclair and Amstrad agents Egypt. Made by Dr. Nabil Nazmi.”
• The keyboard membrane has been replaced by the previous owner. However it did not seem to work properly, so I replaced it with another replica.
• This machine once had a broken DC/DC converter, because TR4, TR5 and D15 have been replaced.
• There was an original ROM and a 27C128 EPROM inside, stacked onto each other. The EPROM contained version 1 of the Arabic modification. A switch selected the active ROM via the Vcc line. It is a miracle that this is working, because one of the chips is always powerless, but still connected to the address and data bus. It would have been better to switch the /OE lines, or use a single 27C512 EPROM and switch the A15 line. However, the first solution would require two additional pull-up resistors, while the latter one might not have been possible for license reasons.

Known Issues

• AY sound is still very silent (after replacing R115).
• Original PAL has a bug that can be fixed by using a modified GAL.
• Cannot put a vLA128 into the ULA socket. ULA needs to either get a heatsink, or the socket needs to be replaced.

Restauration Works

• Carefully cleaned the keyboard and the case. I made sure that the keyboard stickers, the label at the bottom, and the hand-written inscript were not damaged.
• Recapped the machine. C28 was replaced with a 47µF cap. C7 and C8 were replaced with 1µF MLCCs. This fix is supposed to enhance image quality, and indeed all disturbances were gone after that, and the jailbars were greatly reduced.
• Replaced R115 with a 1.65kΩ resistor. This is supposed to raise the AY volume and make it level with the beeper volume. However, the AY is still very quiet.
• Replaced the voltage regulator with an 78S05 drop-in replacement that is rated at 2A. I was lucky because the previous owner replaced the original regulator with a LM1085, which has a different pin configuration. I found out in time, and reverted the connector wires to the original 7805 order, to avoid a “bad surprise” for future owners.
• Replaced the original ROM stack with a single 27C512 EPROM. It contains the Arabic ROM v3.1 (twice, lower half) and the original ZX Spectrum 128K ROM (upper half). I changed the switch to pull the A15 address line either to Vcc or GND. In the original configuration, the switch was connected to Vcc (with a red wire that was removed) and provided it either to the ROM or EPROM. Of course I kept the ROM stack, to be able to restore the machine to its original condition if required.
• 3D printed a cap for the ROM switch.

History

• 2024-01-01:
  • Printed a cap for the ROM switch
  • Restauration completed
• 2023-12-31:
  • Replaced the keyboard membrane with a known-good brand
  • Replaced the original ROM stack with a 27C512 EPROM
  • Updated the Arabic ROM to v3.1
• 2023-12-30:
  • Cleaned the keyboard and the case
  • Recapped the machine
  • Modified C7 C8, C28, and R115
  • Replaced the voltage regulator with an 78S05
• 2023-12-22: Date of purchase