Sinclair ZX Microdrive
| Sinclair ZX Microdrive | |
|---|---|
 |
|
| A ZX Microdrive with a cartridge in its slot | |
| Manufacturer | Sinclair Research |
| Type | Endless-loop tape cartridge drive |
| Release date | Announced 23 April 1982; deliveries from September 1983 |
| Discontinued | 1986 |
| Price | £49.95 (drive); cartridges £4.95, cut to £1.99 in 1985 |
| Interface | Sinclair ZX Interface 1 (ribbon cable to the first drive, connector blocks between drives) |
| Compatible | Sinclair ZX Spectrum, Sinclair ZX Spectrum+, Sinclair ZX Spectrum 128, Sinclair ZX Spectrum +2 (through the Interface 1) |
| Dimensions | 90 × 85 × 40 mm |
| Model | Head board S12100, base PCB S12101, ULA 2G007 |
| Power | 9 V from the computer's supply through the Interface 1; 5 V regulated on the head board |
| Storage | At least 85 KB per cartridge (Sinclair's figure) |
| Media | ZX Microdrive cartridge: an endless loop of 1.9 mm video tape |
The Sinclair ZX Microdrive is a small tape drive made by Sinclair Research for the Sinclair ZX Spectrum. Its cartridge holds an endless loop of narrow video tape that runs past a fixed two-track head, and Sinclair's manual promises at least 85 KB on every cartridge.[1] The drive connects to the Spectrum through the Sinclair ZX Interface 1, and up to eight drives can be joined side by side.[1] Announced with the Spectrum in April 1982, it reached buyers only in the autumn of 1983.[2] Two of the same drives, modified, were built into the Sinclair QL.[2]
Ferranti's files, released in the Ferranti Archives 2026, hold the transistor-level drawing of the drive's custom chip, the Ferranti ULA 2G007, and two of Sinclair's 1983 circuit diagrams of the drive. They are described under Hardware, together with Sinclair's 1985 servicing manual, which documents the production drive. Faults and repairs are under Troubleshooting. The QL's built-in drives have their own guide, the Sinclair QL Microdrive Troubleshooting Guide.
History
[edit | edit source]Sinclair Research showed a prototype Microdrive at the launch of the ZX Spectrum on 23 April 1982. The launch press reported a capacity of 100 KB and a price of about £50, and the early advertising called the medium a "microfloppy"; by the end of 1982 the advertisements described "a single interchangeable storage medium" instead.[2] In spring 1983 Sinclair's Nigel Searle said that the design was finished and that the company was waiting for custom chips made in the UK.[2] The drive's custom chip is a Ferranti part, the ULA 2G007,[3] and the transistor-level drawing of the 2G007 in Ferranti's files was revised four times between 29 March and 28 April 1983.[4]
According to former Sinclair staff interviewed by The Register, David Southward ran the project, Ben Cheese designed the analogue electronics and the ULA, John Williams the mechanics, and Rick Dickinson the case, which copies the Spectrum's styling. Martin Brennan designed the Interface 1 and its software. Cheese and Brennan argued that the spread of tape lengths and motor speeds made the promised 100 KB unsafe, and Sinclair settled on a guaranteed 85 KB.[2]
In July 1983 Sinclair announced that drives and Interface 1 units would ship from September, the first 1,000 going to people who had ordered Spectrums early.[2] Sinclair User gave the prices in October 1983 as £49.95 for the drive and £29.95 for the Interface 1, with cartridges at £4.95 each.[5] Sinclair cut the cartridge price to £1.99 early in 1985.[2]
The drive gained a reputation for unreliability. John Mathieson, a Sinclair engineer of the time, told The Register that variation in the heads and tapes made reliability "quite spotty", and that by trying a few drives one could usually find a good one.[2] Sinclair's own manual warns that cartridges "will not last forever".[1] The QL used two drives whose motors ran 25 per cent faster than the Spectrum drive's. After Amstrad bought Sinclair's computer business in 1986 it dropped the Microdrive in favour of 3-inch discs.[2]
Specification
[edit | edit source]| Item | ZX Microdrive |
|---|---|
| Medium | Cartridge holding an endless loop of video tape, 1.9 mm wide and 23 µm thick[5] (the 1985 servicing manual says "2 mm wide video tape"[3]) |
| Capacity | "Up to 100K bytes"; the capacity "should never be less than 85 kilobytes"[1] |
| Recording | Two tracks on a two-track ("stereo") head, alternate bytes on each track, frequency-modulation coding[3] |
| Tape speed | 30 inches (about 76 cm) per second[5] |
| Formatting time | About 30 seconds[1] |
| Drives per Spectrum | Up to eight; the drive nearest the computer is Microdrive 1[1] |
| Motor | Mabuchi EG-500AD-9F, 9 V, 2,400 rpm[3] |
| Supply | 9 V line from the computer's supply through the Interface 1; 78M05 regulator for the 5 V logic[6][3] |
| Custom chip | Ferranti ULA 2G007 (IC1 on the head board)[3] |
| Write protection | Removable tab on the cartridge, sensed by a micro switch in the drive[1][6] |
| Size | 90 × 85 × 40 mm[5] |
Cartridges
[edit | edit source]
A cartridge holds one reel of tape whose ends are joined into a loop. The tape is drawn from the centre of the reel, passes the opening in the edge of the cartridge where the head and the drive roller meet it, and winds back on to the outside of the reel.[2][7] Pulling tape from the centre strains it, and DataServe Retro gives that as the reason Sinclair used video tape, which is tougher than audio cassette tape.[7] Behind the tape, a small felt or sponge pad on a metal spring presses it against the head.[8]
Formatting checks the whole loop, marks any area that cannot be written or read so that it is skipped, and takes about 30 seconds. The capacity shown by CAT afterwards varies from cartridge to cartridge.[1] Snapping off the plastic tab on the side of the cartridge write-protects it, and a piece of tape over the gap reverses this.[1]
Sinclair's manual gives two rules: never take a cartridge out while the drive's light is on, and never switch the power on or off with a cartridge in the drive.[1] Sinclair's hints sheet explains the second rule: switching on or off with a cartridge inserted can erase the stretch of tape over the head, and the drive then reports "Microdrive not present" until the cartridge is formatted again.[9]
Cartridges wear out. The manual gives the symptom of an ageing cartridge as the computer taking longer and longer to find a file, and advises keeping back-up copies.[1] Sinclair claimed more than 5,000 operations per tape.[5] The QL uses the same cartridge with a different format.[7]
Hardware
[edit | edit source]The drawings
[edit | edit source]| Document | Date | What it shows |
|---|---|---|
| Ferranti Archives MD-FER-SCH-001[4] | Revisions dated 29 March, 19 April, 22 April and 28 April 1983 | Pen drawing of the ULA at transistor level, titled "MICRODRIVE 2G007". No company name, drawing number or printed title block. |
| Ferranti Archives MD-SIN-SCH-003[10] | 10 June 1983 | Sinclair's hand-drawn circuit of the whole drive, "(c) Copyright Sinclair Research Ltd", divided into the head board S12100 and the base PCB S12101. |
| Ferranti Archives MD-SIN-SCH-004[6] | 16 September 1983 | The same circuit drafted on a Sinclair Research drawing form ("TITLE MICRODRIVE"). The drawing number and issue fields are blank. |
| Ferranti Archives MD-FER-FOL-008[11] | Undated | The Ferranti folder, labelled "SPECTRUM - MICRODRIVE -", that held these drawings and the Interface 1 papers. |
| Servicing manual, Fig. 1.3[3] | October 1985 | Sinclair's production circuit diagram of the drive, CSK 10626. |
The September 1983 drawing settles several values that are over-written on the June sheet: R4 150, C6 330p, CLK on pin 21, COMMS IN on pin 22, and IC2 as a 78M05.[10][6] Neither drawing gives values for C1, C2, C7 or C9. The 1985 parts list gives C1 and C2 as 220 pF and lists C7 and C9 on the head board as not used.[3]
Head board S12100
[edit | edit source]The head board carries the head, IC1 (the 2G007) and IC2 (a 78M05 5 V regulator). IC2 takes the 9 V line from the base PCB at head-board connection 10, with C4 0.22 µF on its input, and supplies +5 V to IC1 pin 9 and to the resistors on IC1, with C3 0.47 µF on its output.[6]
The head L1 has three windings. Track A (L1 pins 1 and 2) goes to INA1 and INA2 on IC1 pins 8 and 13, and track B (L1 pins 3 and 4) to INB1 and INB2 on pins 6 and 10. The third winding (L1 pins 5 and 6) has D1 across it and D2 in series with the ERASE line.[6] The 1985 manual calls it the erase head. It is set apart from the write head; its current is switched in the Interface 1 (gate array pin 35 and transistor Q9) so that erasing starts before writing, and the current is about 25 mA.[3]
The 1985 manual describes what IC1 does. In read mode the signals from the two head windings are amplified differentially in two amplifier chains inside IC1, followed by a peak detector and a hysteresis stage, and leave as logic levels on DATA 1 (pin 24) and DATA 2 (pin 19). When the R/W line goes low the drive writes: DATA 1 and DATA 2 become inputs that drive the current sources for the two tracks. The write current is deliberately asymmetric, because the tape is erased with a DC field, and R4 and R5 on pins 11 and 7 set it; the manual says they would need changing for a different tape or head. The gain can be changed through the ratio of R9 and R8 on pins 15 and 14, though it should not need changing because only one type of tape was used. The replay signal across C1 and C2 is typically 400 to 500 mV at the low recording frequency and 250 to 350 mV at the high one.[3]
The 2G007 ULA
[edit | edit source]Ferranti's drawing of the 2G007 is a large pen drawing at transistor level. It shows two identical analogue channels, A (transistors 1 to 65) and B (101 to 165), each built from long-tailed pairs and current mirrors, amplifier stages, a comparator and a Schmitt trigger, and ending in an emitter-follower output, DATA A or DATA B. A logic section along the bottom takes COMMS IN, CLOCK and R/W through buffers. COMMS IN feeds a D-type flip-flop clocked by CLOCK, whose output drives COMMS OUT; NOR gates combine the flip-flop and R/W to gate the channels and to drive two outputs, G1 and G2, with 80K pull-ups. The sheet does not use the words read, write, head or track.[4]
| Pin | Name on the drawing | Connected to | Function (1985 manual) |
|---|---|---|---|
| 8, 13 | INA1, INA2 | Head winding, track A | Read input and write output, track A |
| 6, 10 | INB1, INB2 | Head winding, track B | Read input and write output, track B |
| 17, 18 | C1, C2 | Capacitor C1 | Replay signal measured across it |
| 1, 2 | C3, C4 | Capacitor C2 | Replay signal measured across it |
| 3, 4, 5 | – | R6 and R7 4K7, C6 330p, C7 (not fitted in production) | – |
| 14, 15, 16 | – | R8 and R9 4K7, C8 330p, C9 (not fitted in production) | Gain set by the ratio of R9 to R8 |
| 11 | – | R4 150 to +5 V | Write current |
| 7 | – | R5 180 to +5 V | Write current |
| 9 | VCC | +5 V from IC2 | Supply |
| 12 | 0V | 0 V | Ground |
| 21 | CLK | COMMS CLK bus line | Select clock |
| 22 | COMMS IN | Select line from the unit before | Shift-register input |
| 20 | COMMS OUT | Select line to the next drive; Q1 base through R1 | High when this drive is selected |
| 23 | R/W | R/W bus line | Low to write |
| 24 | DATA 1 | DATA 1 bus line | Track 1 data |
| 19 | DATA 2 | DATA 2 bus line | Track 2 data |
The QL uses the same chip. The QL Service Manual lists the Microdrive ULAs on the QL's two head boards, IC29 and IC30, as Ferranti 2G007 issue 3, includes a drawing of the 2G007's internal circuit, and gives the same pin numbers for the supply and data test points.[12] Its Appendix D describes two faults on some early 2G007s, since remedied: pin 1 latching at about 4 V at power-on or while writing, and data lines that stay at about 4.5 V instead of going tri-state when the drive is not selected, which stops the other drive working.[12]
Base PCB S12101: motor, LED and write protect
[edit | edit source]The base PCB carries the two edge connectors, the motor switch, and the connections to the motor, the LED and the write-protect micro switch, which are off the board.[6] When IC1's COMMS OUT (pin 20) goes high it turns on Q1, whose base is fed through R1 33K. Q1 pulls the base of Q2, a PNP transistor with its emitter on the 9 V line, low through R2 1K, and Q2 powers the motor through the R10 position, which is fitted with a wire link, and lights the red LED D5 through R3 820. D3 is connected across the motor and C5 47n across its supply.[6][3] Q2's output also feeds the erase winding through head-board connection 6 and, through D4, the write-protect micro switch, whose other side is the WR PROT line. A selected drive holding a cartridge with its tab intact pulls WR PROT high; with no cartridge, or with the tab broken off, the line stays low.[6][3]
A 22 µF capacitor from Q2's base to its collector slows the motor's start so that it does not snatch the tape. Both 1983 drawings show it, as C10 22u 16V.[10][6] The 1985 manual calls it C7 and lists it as a modification to be checked for and, if missing, fitted retrospectively; The Register also describes it as a later addition.[3][2] The 1985 parts list gives Q1 as a BC184 and Q2 as a Ferranti ZTX551; where Q1 is a Ferranti BC184P it is fitted the opposite way round to the board legend.[3]
Edge connector and the select chain
[edit | edit source]The bus runs the length of the base PCB between the two edge connectors, and drives plug together through connector blocks, so the lines pass from one drive to the next.[6][1]
| Contact | Signal | Contact | Signal |
|---|---|---|---|
| 1A | DATA 1 (IC1 pin 24) | 1B | DATA 2 (IC1 pin 19) |
| 2A | WR PROT (micro switch) | 2B | COMMS CLK (IC1 pin 21) |
| 3A | Not used | 3B | Not used |
| 4A | 9V | 4B | Select line (see below) |
| 5A | R/W (IC1 pin 23) | 5B | ERASE (erase winding through D2) |
| 6A, 7A, 8A | 0V | 6B, 7B, 8B | 0V |
The drawing notes "A IS COMPONENT SIDE"; the 1985 diagram adds that B is the solder side.[6][3]
Every line passes straight through except the select line. On the September 1983 drawing the line arriving at end A, named COMMS OUT after the signal of the unit before, goes only to IC1's COMMS IN (pin 22). IC1's COMMS OUT (pin 20) leaves at end B on contact 4B, named COMMS IN after the input of the next drive. End A is therefore the end towards the Interface 1.[6] The 1985 diagram labels the same contact with the drive's own signal, COMMS OUT 4B.[3]
The 1985 manual explains the scheme. Each drive's IC1 holds one stage of a shift register; the Interface 1 clocks a single 1 along the chain with COMMS CLK until it reaches the wanted drive. The drive whose stage holds the 1 has COMMS OUT high, which runs its motor, lights its LED and allows erase current and write-protect sensing; while COMMS OUT is low all of these are off. Only one drive may be in use at a time.[3]
Production changes
[edit | edit source]The 1985 manual lists these changes to the drive:[3]
- The 22 µF capacitor across Q2 (above), to bring the motor up to speed gradually. Check for it and fit it if missing.
- C1, C2, C6 and C8 changed from polystyrene to axial ceramic for convenience of production, and C1 and C2 changed from 150 pF to 220 pF. Check the value and change it if necessary.
- A 47 nF decoupling capacitor added between IC1 pins 9 and 12 (C10 in the head-board parts list, where it is typed as 48 nF).
- The micro switch positioning peg moved 0.5 mm towards the base edge. On early drives rough use could bend the switch's moving arm and make it misread.
- The ratchet roller, ratchet and ratchet spring replaced by two rollers, one rubber and one plastic. The ratchet stopped a loop forming in the tape when the roller is turned by inserting the cartridge.
- A harder material for the clamp spring, because the original sometimes lost its set and had to be bent back.
Electrolytic capacitors
[edit | edit source]| Reference (1983 drawings) | Board | Value on the drawings | 1985 parts list | Position |
|---|---|---|---|---|
| C3 | Head board | 0.47 µ | 0.47 µF 50 V electrolytic, axial | 78M05 output, +5 V |
| C4 | Head board | 0.22 µ | 0.22 µF 50 V electrolytic, axial | 78M05 input, 9 V |
| C10 | Base PCB | 22 µ 16 V | C7, 22 µF electrolytic, axial (no voltage given) | Q2 base to collector, motor soft start |
The drawings give no voltage ratings for C3 and C4; the ratings above are from the 1985 parts list, which marks no other capacitor in the drive as electrolytic.[6][3]
Troubleshooting
[edit | edit source]Sinclair's hints sheet asks owners to rule out the connections and the cartridge before deciding that a drive is faulty, and the community repair guides start in the same place: the cartridge, the head and the drive roller.[9][13] Use a cartridge known to work in a drive known to work as the reference, and change one thing at a time. The drive's head cannot be replaced; Sinclair's manual treats a failed head as the end of the drive.[3]
Symptoms
[edit | edit source]| Symptom | Probable cause | Action |
|---|---|---|
| "Microdrive not present" with a cartridge inserted | No cartridge, an unformatted cartridge, a jammed tape, a cartridge not seated against the head (often with a rattle), or the tape over the head erased by switching on or off with the cartridge in[9] | Push the cartridge fully home; try a known-good cartridge; format the suspect cartridge again (this erases it) |
| "Microdrive not present" with every cartridge | Q2 or IC1 in the drive[3]; a failed Interface 1 gate array[14][7] | Check Q2, then IC1 (the 1985 manual's order); try the drive on another Interface 1. See Interface 1 troubleshooting |
| Drive spins after CAT 1 and does not stop | A cartridge formatted on a QL[7] | Format it on the Spectrum (this erases it) |
| Files take longer and longer to find | An ageing cartridge[1] | Copy the files to another cartridge or to tape |
| Irregular running, poor formatting, low capacity | A tight or ageing tape, a dirty head, a dirty or worn drive roller[7][13] | Format the cartridge two or three times in succession; clean the head and roller. DataServe Retro expects a good cartridge to format to at least 80 KB |
| Screen stays white during the third stage of FORMAT | Noise on the write line[3] | Check the motor alignment, IC1 and the regulator IC2 |
| Border stays black during the first part of FORMAT | A clock fault in the Interface 1[3] | Check the Interface 1's crystal oscillator |
| Writing fails (the manual's test program stops in its write test) | R4, R5 or IC1[3] | Check R4 (150) and R5 (180), then renew IC1 |
| Reading fails (the test program stops in its read test) | C1, C2 or IC1[3] | Check C1 and C2 (220 pF; 150 pF on early drives), then renew IC1. If that fails, the manual's answer is a new drive |
| Drive reports write protection although the tab is intact | The micro switch; on early drives its arm can be bent[3] | Check the switch with a meter, power off; check its peg position |
| Drive selected but does not run | The motor or the mechanics: burrs on the head springs (look for scratches on the cartridge), a weak clamp spring (listen for clicks or crunching), or the ratchet assembly[3] | See below. If the motor has overheated, check the base plate for buckling; a buckled plate means a new drive |
| Tape looped out of the cartridge, tight tape, broken tape | As above[3] | As above |
| LED never lights | The LED[3] | The LED is a push fit in the top cover |
Cartridge pads and tape
[edit | edit source]The felt or sponge pad behind the tape is the usual failure on cartridges today. It is very fragile and often disintegrates in use; debris that gets into the cartridge can jam it for good, and without the pad the head does not make good contact with the tape and the bare spring rubs on it. Spectrum for Everyone recommends re-felting cartridges before use:[8]
- Lift the tape gently with clean tweezers to reach the old pad.
- Scrape the old sponge off the spring without letting debris fall into the cartridge.
- Clean the spring with a cotton bud and isopropyl alcohol, holding the cartridge tilted so that nothing runs inside.
- Cut a piece of fine self-adhesive felt to about the size of the metal pad and press it on to the spring.
- Take up the slack in the tape by rolling it gently round with the stick of a cotton bud.
The tape itself stretches. The Register describes each start of the drive, from rest to about 750 mm per second, as a sharp tug on the tape where it leaves the centre of the reel; the 22 µF soft-start capacitor on later drives was added to reduce it.[2] Oxide shed by the tape collects on the head and the roller.[2][13]
Opening the drive
[edit | edit source]- If two drives are joined, remove the joining plate underneath (four small screws). It covers the latch at the bottom of the front edge that the top case clips into, and leaving it on can snap the locating lug.[13]
- Lift the decorative top plate, which is glued on. Sinclair's manual says to lift one corner with a small screwdriver and discard the trim, fitting a new one on reassembly.[3] DataServe Retro lifts it intact, if necessary with a little white spirit run under the edge (wipe it off at once, as it can attack some plastics), and refits it with double-sided tape.[13]
- Remove the two screws under the plate, release the tongue at the back edge, and lift the top cover away. The LED in the cover stays connected to the base board by flying leads.[3]
- To reach the motor or the board components, remove the two screws holding the base board to the base and the screw underneath that holds the base to the head chassis. Do not strain the flexible cables between the head board and the base board. Do not separate the head board from the head chassis: the head alignment is critical and needs special tools.[3]
- On reassembly, do not trap the LED leads between the head board and the base.[3]
Head and drive roller
[edit | edit source]With the top off, the head is in the middle, with the drive roller to its left and the cartridge sensor to its right.[13]
- Clean the head with a cotton bud, or better a chamois stick, moistened with head-cleaning fluid.[13]
- Clean the brown oxide off the rubber drive roller in the same way, then wipe off the fluid with a dry bud.[13]
- Check that the rubber roller has not ridden up the motor spindle. DataServe Retro calls this a common problem that hampers the drive and can tangle and snap the tape; push it back down.[13]
- A well-used roller can wear a groove the width of the tape, which shows as inconsistent loading or the drive failing to find files with a clean head. Replace a worn roller; if the wear is slight, pulling it off and refitting it the other way up can help.[13]
Sinclair's manual adds the factory procedures:[3]
- When renewing the motor, renew the plastic and rubber rollers as well. A motor with a combined rubber roller and ratchet gets the later pair of rollers and loses the ratchet assembly. The gap between the rubber and plastic rollers is 0.25 mm. Set the motor's position with Sinclair's motor location jig before tightening its screws.
- Motor drive test: with a cartridge in and the motor running, push the cartridge gently away from the roller until the tape stops, then release it. It should return under the clamp spring and the tape should restart without rattling or noise. If not, and the clamp spring is right, adjust the head spring: unscrew the motor and press the head spring down on each side of the head to the reference dimension in the manual (Fig. 4.1 area), then refit the motor with the jig. Never adjust the later clamp spring, which has a loop below the chassis level.
- Burrs on the head springs and a weak clamp spring both stop the cartridge being pushed firmly into the drive position, and cause intermittent running and tape spill. File burrs off with a small flat instrument file, taking care not to damage the drive roller. Bend the clamp spring back with bull-nosed pliers so that it ends within the diameter of the adjacent head-assembly post.
Electrical checks
[edit | edit source]| Supply | Where | Expected |
|---|---|---|
| +5 V | IC1 pins 11, 7 and 9 | +5 V, no discernible ripple |
| +9 V | Pin 10 of the ribbon connector (head-board connection 10) | +9 V |
The manual's initial-check table gives the drive's +9 V on edge contact 4b; its own circuit diagram and both 1983 drawings put the 9 V line on 4A.[3][6] Check which contact carries 9 V on the board before probing.
The 1985 manual's waveform test: with the drive connected to a working system as Microdrive 1 and a cartridge inserted, enter OUT 239,252: OUT 239,254 to set it spinning. Look at the waveform at either end of C1 and C2, with the oscilloscope earthed on the body of the regulator. It should show packets of oscillation (one long, two short) separated by quiet areas. The oscillation should be about 0.5 V peak-to-peak, and the manual stresses that it should be steady; the noise in the quiet areas should not exceed 25 mV. Compare with a known-good drive.[3]
The circuit also points to suspects for faults the manual does not list. These are reasoning from the 1983 drawings, not documented repairs:[6]
- 9 V at head-board connection 10 but no 5 V at IC1 pin 9: the only parts between are IC2 (78M05), C3 and C4.
- Motor and LED both dead when the drive is selected: both are fed from Q2, so check Q1, Q2, R1 and R2 and whether IC1 pin 20 goes high, before suspecting the motor.
- LED lights but the motor does not run: the R10 link, the motor connections, D3 (a short across the motor) or the motor.
- A drive that cannot be selected stops every drive beyond it as well, because the select bit passes through each drive's IC1 (pin 22 to pin 20) and its 4B contact. Test a suspect drive on its own, then move it to the end of the chain. A drive whose data lines do not go tri-state when deselected (the early 2G007 fault described in the QL manual) would also disturb the other drives, since DATA 1 and DATA 2 are shared along the bus.[12]
- Erase faults: the erase current flows from Q2 through head-board connection 6, the erase winding and D2 to the ERASE line and the Interface 1's driver. Check D2 and the winding for continuity, then the Interface 1's Q9 and R30.
Replacements
[edit | edit source]OqtaDrive is a Microdrive and network hardware emulator for the ZX Spectrum and QL, which stores cartridge images instead of tape.[15]
Documentation
[edit | edit source]- Servicing Manual for ZX Microdrive, Interface 1 and Interface 2 (Thorn EMI Datatech for Sinclair Research, October 1985): description, disassembly, tests, fault tables, modification history, circuit diagram CSK 10626 and parts lists.
- ZX Interface 1 and ZX Microdrive owner's manual (Sinclair Research, 1983).
- Ferranti Archives 2026: MD-FER-SCH-001 (2G007 circuit), MD-SIN-SCH-003 and MD-SIN-SCH-004 (drive circuits), MD-FER-FOL-008 (folder).
Gallery
[edit | edit source]-
A ZX Spectrum on its Interface 1, with a Microdrive connected by the ribbon cable -
Three Microdrives joined side by side beside a ZX Spectrum, with a ZX Interface 2 and a Stonechip Echo Amplifier
Related pages
[edit | edit source]- Sinclair ZX Interface 1
- Sinclair ZX Spectrum and the Sinclair ZX Spectrum Troubleshooting Guide
- Sinclair QL Microdrive Troubleshooting Guide
- Ferranti Archives 2026
References
[edit | edit source]- ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 1.11 1.12 Sinclair Research, ZX Interface 1 and ZX Microdrive owner's manual, first edition 1983 (File:Sinclair ZX Interface 1 and ZX Microdrive Manual (1983).pdf). Source for the 100 KB maximum and 85 KB minimum, up to eight drives, the formatting time, write protection, the cartridge handling rules and the cartridge lifespan.
- ↑ 2.00 2.01 2.02 2.03 2.04 2.05 2.06 2.07 2.08 2.09 2.10 2.11 2.12 Tony Smith, "Infinite loop: the Sinclair ZX Microdrive story", The Register, 13 March 2013. Source for the announcement, the delays, the design team, the launch and prices, the tape path and stretching, the later 22 µF capacitor, the QL drives' speed and Amstrad's decision.
- ↑ 3.00 3.01 3.02 3.03 3.04 3.05 3.06 3.07 3.08 3.09 3.10 3.11 3.12 3.13 3.14 3.15 3.16 3.17 3.18 3.19 3.20 3.21 3.22 3.23 3.24 3.25 3.26 3.27 3.28 3.29 3.30 3.31 3.32 3.33 3.34 3.35 Thorn (EMI) Datatech Ltd for Sinclair Research Ltd, Servicing Manual for ZX Microdrive, Interface 1 and Interface 2, October 1985 (File:Sinclair ZX Microdrive, Interface 1 and Interface 2 Servicing Manual (Thorn EMI Datatech, 1985).pdf). Section 1 paras. 2.1 to 2.15 (how the drive works), Section 2 para. 3 (disassembly), Section 3 paras. 3.1 to 3.3 (tests and waveforms), Section 4 paras. 2.3 (modification history), 3.4 (supply checks and fault table) and 4.1 to 4.5 (repair), Fig. 1.3 (circuit CSK 10626), Tables 5.3 to 5.5 (parts).
- ↑ 4.0 4.1 4.2 Ferranti Archives 2026, MD-FER-SCH-001, hand-drawn transistor-level circuit of the Microdrive ULA, titled "MICRODRIVE 2G007", with revisions dated 29 March, 19 April, 22 April and 28 April 1983 (MD-FER-SCH-001). Source for the revision dates, the two channels with Schmitt triggers and DATA outputs, the COMMS IN to COMMS OUT flip-flop and the G1 and G2 outputs.
- ↑ 5.0 5.1 5.2 5.3 5.4 Reviews of the Interface 1 and Microdrive from Sinclair User, October 1983, and Your Computer, 1983, transcribed at World of Spectrum, "Hardware Feature #7" (retrieved 2026-10-18). Source for the prices, the 1.9 mm, 23 micron tape at 30 inches per second, Sinclair's claim of more than 5,000 operations per tape, and the 90 by 85 by 40 mm case.
- ↑ 6.00 6.01 6.02 6.03 6.04 6.05 6.06 6.07 6.08 6.09 6.10 6.11 6.12 6.13 6.14 6.15 Sinclair Research, ZX Microdrive circuit diagram on a Sinclair Research drawing form, 16 September 1983, Ferranti Archives 2026 (MD-SIN-SCH-004). Source for the 2G007 pin connections, IC2 78M05 with C3 0.47u and C4 0.22u, the three-winding head, the Q1 and Q2 motor stage, the micro switch and LED wiring, the edge-connector allocation and the select-line routing.
- ↑ 7.0 7.1 7.2 7.3 7.4 7.5 DataServe Retro, "The Sinclair Microdrive and Interface 1" (retrieved 2026-10-18). Source for the tape path and the choice of video tape, reviving tight cartridges by formatting, QL-formatted cartridges and Interface 1 ULA failures.
- ↑ 8.0 8.1 Will Woodvine, "How to Refelt ZX Microdrive Cartridges", Spectrum for Everyone (retrieved 2026-10-18). Source for the pad failure and the re-felting method.
- ↑ 9.0 9.1 9.2 Sinclair Research, "ZX Microdrive and Interface 1: Useful Hints and Tips" leaflet, transcribed by Wilko Schröter in 2004, World of Spectrum mirror on the Internet Archive (ZXMicrodriveAndInterface1-UsefulHintsAndTips.pdf). Source for checking the connections before assuming a fault and for the causes of "Microdrive not present".
- ↑ 10.0 10.1 10.2 Sinclair Research, hand-drawn ZX Microdrive circuit diagram, "MICRODRIVE 10 JUN 83", Ferranti Archives 2026 (MD-SIN-SCH-003). Source for the head board S12100 and base PCB S12101, the over-written values settled by the September drawing, and C10 22u 16V on the June 1983 circuit.
- ↑ Ferranti card folder labelled "SPECTRUM - MICRODRIVE -", Ferranti Archives 2026 (MD-FER-FOL-008). The archive catalogue lists the Microdrive and Interface 1 documents as its contents.
- ↑ 12.0 12.1 12.2 Thorn (EMI) Datatech Ltd for Sinclair Research Ltd, Sinclair QL Service Manual, issue 3, October 1985 (File:Sinclair QL Service Manual.pdf): Section 1 para. 7 (Microdrive circuit), Section 4 para. 2.4.2 (test points), Appendix D para. 4 (faults on early Microdrives), Fig. D9 (ULA 2G007 internal circuit) and the Section 5 parts list (IC29, IC30 "2G007 Issue 3", Ferranti).
- ↑ 13.00 13.01 13.02 13.03 13.04 13.05 13.06 13.07 13.08 13.09 DataServe Retro, "Servicing the Sinclair ZX Microdrive" (retrieved 2026-10-18). Source for opening the drive, cleaning the head and roller, the roller riding up its spindle, roller wear, and the 80 KB formatting guide.
- ↑ "Micro Clinic: hints for repairing Sinclair Spectrums", Television magazine, 1986, transcribed at World of Spectrum, "Hardware Feature Extra" (retrieved 2026-10-18). Source for the Interface 1 ULA failing when the interface moves or is disconnected suddenly, and for cleaning the edge connectors.
- ↑ OqtaDrive, project site (retrieved 2026-10-18).