Sinclair QL Microdrive Troubleshooting Guide

The two Microdrive slots on a Sinclair QL: MDV1 on the left, MDV2 on the right.

This guide covers fault diagnosis and repair of the Sinclair QL's Microdrives and their cartridges. It follows Section 4 para 2.4 and Appendix D (Microdrive Fault Finding) of the Sinclair QL Service Manual, which Sinclair wrote specifically for this mechanism.

Microdrive reliability was one of the QL's most damaging launch problems, and it remains the subsystem most likely to need attention on a surviving machine.

How a Microdrive works

A QL Microdrive cartridge. Inside is an endless loop of magnetic tape running over a felt pressure pad.

A Microdrive cartridge holds a continuous loop of magnetic tape. The drive's motor turns a rubber drive roller (a capstan) that presses against the tape via a pinch wheel in the cartridge, pulling the loop past the read/write head. There is no rewind: to reach a given sector the drive simply waits for it to come round again.

Formatting writes sector headers (26 bytes, containing the sector number and medium name) and data blocks (512 bytes plus a checksum byte), separated by blank gaps. The gaps matter: the drive electronics use them to tell real data from tape noise, and the ZX8302 raises an interrupt as each gap passes. Sectors are written in descending order from 255, and because the tape is a loop the higher numbers get overwritten — the highest surviving sector number is what QDOS reports after FORMAT mdv1_. A healthy cartridge formats to a little over 200 sectors.

Sector headers are never rewritten after formatting, which is why a cartridge that formats to a low sector count should be re-formatted or retired rather than patched.

The drives are controlled through the ZX8302 at registers $18020 (control/status), $18021 (interrupt control) and $18022/$18023 (data, tracks 1 and 2). Drives are selected by clocking a token along a shift chain — set the select bit, then toggle the clock bit once per drive number. Once selected, the motor runs.

First checks

Before suspecting the electronics, check the supplies.

Microdrive supply test points (MDV2 references in brackets)

Supply	Test point	Expected
+5 V logic	IC30 (IC29) pins 11, 7 and 9	+5 V DC ± 0.25 V, no discernible ripple
+9 V motor	Pin 6 of the ribbon connector	+9 V

Note two limitations before going further:

• The read/write head cannot be replaced. Every other part of the drive can be, but a failed head means fitting a complete replacement Microdrive unit.
• If the motor has overheated, check the baseplate for buckling. A buckled baseplate means the whole drive unit must be renewed.

Mechanical checks

The Service Manual recommends doing these before any signal testing, as they are the most common cause of trouble. The mechanics are interdependent, so over-adjusting one can upset another.

Drive roller

The motor drive roller must not be distorted in any way, and must have the specified gap between it and the collar beneath it. A roller that has gone hard, shiny, oval or sticky with age is the classic cause of tape slipping, speed variation and read errors — the QL's equivalent of a perished drive belt.

Clean the roller with isopropyl alcohol on a lint-free swab, rotating it so the whole circumference is cleaned. If it is distorted or has degraded into a sticky surface, it must be replaced; reproduction rollers are made for these drives.

Motor position

An incorrectly positioned motor causes tape slipping and speed problems. Sinclair's procedure is to check and set the position using the motor location jig (Figure 4.1 of the Service Manual). If tape speed problems persist after the roller has been cleaned or replaced, motor position is the next suspect.

Side spring

With a cartridge spinning in the drive, push the cartridge away from the roller so the side spring is compressed, then release it. The cartridge should return and spin freely against the roller without slipping or jitter. The top of the spring should be reasonably level and must not foul the chassis.

Head position

The digital head should sit with both front feet and at least one rear foot touching the chassis. It is not always possible for the fourth foot to touch, because of manufacturing adjustment of the head itself. A head tilted up, down or sideways compounds any head-spring problem.

The face of the head must be clean.

Head springs

Weak or distorted head springs cause poor cartridge compatibility — the drive reads its own tapes but not others'. Viewed from the front of the drive, the top of the springs should appear roughly level with the bottom of the lower head guides.

The springs can be tested functionally with the signal test program: write to a tape and note the peak-to-peak read level, then twist the cartridge clockwise and release, and anti-clockwise and release. The level should return to roughly the same value. If the level drops significantly after settling for about 7 seconds in either direction, there is a problem.

Cleaning the head

Head cleaning is the single most useful routine maintenance job on a Microdrive, and the Service Manual lists a head cleaner among the servicing materials.

1. Power off and remove the cartridge.
2. Using a lint-free swab lightly moistened with isopropyl alcohol, wipe the face of the head. Wipe in the direction of tape travel rather than scrubbing across it.
3. Repeat with a clean swab until no more oxide comes off.
4. Clean the drive roller at the same time — oxide and rubber residue there cause as many problems as a dirty head.
5. Let the alcohol evaporate fully before inserting a cartridge.

Do not use abrasive cleaners or sharp tools on the head face. The head cannot be replaced if damaged.

Cartridges

Cartridges are consumable and are usually in worse condition than the drives. A drive that fails every cartridge may be faulty; a cartridge that fails in a known-good drive is simply worn out.

The felt pressure pad

Inside each cartridge, a small felt pad presses the tape against the head. Over decades this felt hardens, compresses or falls off — its adhesive perishes — and the result is weak or intermittent read signal, exactly mimicking a worn head or weak head springs.

Re-felting is the community repair: open the cartridge, remove the old pad and its residue, and fit a new pad of equivalent thickness and softness, using a small amount of a suitable adhesive. Work carefully — the tape loop must not be creased, twisted or contaminated, and the pad must sit square so it presses the tape evenly across the head.

Signs that point to the pad rather than the drive:

• The same cartridge reads poorly in more than one known-good drive.
• Read level (measured with the signal test program) is low for this cartridge but normal for others.
• The cartridge reads better if gently pressed while running — a sign the tape is not being held firmly to the head.

Tape condition and handling

• Tape that has become sticky, shed oxide, or been stored in a hot or damp place should not be run — it will contaminate the head.
• A jammed tape or tight cartridge tape is listed by Sinclair as a mechanical fault requiring attention to the clamp spring, rollers and motor.
• Listen when a cartridge is inserted: clicking or crunching sounds indicate a weak clamp spring.
• Damaged cartridges and "microdrive incompatibility" are treated together in the manual with tight tape, noisy operation and format failures — all point at the mechanics rather than the electronics.
• Check the write-protect tab before assuming a fault: a protected cartridge cannot be written or formatted.

Spinning a drive for test

To check a drive mechanically or with a scope you need it to run continuously. Sinclair's program does this — key it in with a known-good pre-recorded tape in MDV1:

100 DEFine PROCedure sedes(n,flag)
110 reg=98336:pc_sel=2+flag
120 FOR i=1 TO n
130 POKE reg,pc_sel
140 pc_sel=pc_sel&&253
150 POKE reg,pc_sel
160 pc_sel=2
170 END FOR i
180 END DEFine

200 DEFine PROCedure start_mdv(n)
210 sedes n,1
220 END DEFine

300 DEFine PROCedure stop_all
310 sedes 8,0
320 END DEFine

start_mdv(n) turns on drive n and keeps it spinning; stop_all stops all drives.

With a drive spinning:

1. Using an oscilloscope, check that a signal is present from the read head (INA at IC29) and trace it through to the RAW inputs on the ZX8302.
2. Nothing from the head — the head is faulty (drive replacement).
3. Input to IC29 but no output — faulty IC or headboard component.
4. Check the headboard connector the same way.
5. Repeat on the second drive. If both drives appear to work, suspect the ZX8302 — it is socketed, so it is easy to substitute.
6. Check the clock signal on pin 25 of the ZX8302.

A known-good ZX Microdrive plugged into the external Microdrive port is another way to decide whether the fault is in the machine or in the drives.

Signal test procedure (Appendix D)

Sinclair's SIGNAL TEST program writes a steady 100 kHz tone to a cartridge (erasing simultaneously) for 8 seconds, then switches to read, so the electronics and mechanics can be assessed against a stable waveform. A two-channel oscilloscope is required, set to add CH1 and CH2 with CH2 inverted to measure the differential signal.

Load the program, remove the program cartridge, insert a known-good blank, and select the drive. The QL beeps when it switches from writing to reading; the space bar stops and starts the process.

Write check

While writing, check the TTL data lines at ULA pins 19 (24). If those are correct, look at pins 4 and 5 (14 and 15). If the signal is not present there, the fault is a faulty ULA or head.

Read check

After writing, switch to read mode:

• Check data lines at ULA pins 19 (24). A waveform that is very unstable in the X axis indicates a mechanical fault.
• If the waveform is poor or absent, check the peak-to-peak signal at pins 4 and 5 (14 and 15).
• Signal present and correct there — the fault is in the intermediate read stages.
• No signal at all, with the write check correct — the head is faulty.
• Low voltage at this point (for example 150 mV pk-pk) indicates a worn head, which causes intermittent soft read errors rather than outright failure.
• A waveform very unstable in the Y axis (widely differing pk-pk levels) points to the erase function — or simply to a poor tape.

Erase check

Erase problems show as residual magnetism from previous recordings not being removed during writing. The microswitch is usually the culprit and can be checked with a multimeter with the power off. To verify with the signal test program, measure at HBC pin 11:

Erase check — HBC pin 11

Condition	Not write-protected	Write-protected
During write/erase cycle	approx. 5.6 V	approx. 0.8 V
During read	approx. 9.0 V	approx. 0.1 V

Symptom tables

These are the Service Manual's Microdrive fault tables. Where a component is given as "X(Y)", X is the MDV1 reference and Y the MDV2 equivalent.

Microdrive symptom to remedy

Symptom	Remedy
Screen displays "Put a blank tape in MDV1 (MDV2)" even when a tape is present	1. Check/replace TR7 (TR6). 2. Replace IC30 (IC29).
Screen displays "Failed microdrive test"	1. Check for a write-protect tab on the cartridge. 2. Check headboard connector HBC1 (2). 3. Check the microswitch. 4. Check/adjust the motor. 5. Check R35 (R34), R37 (R36), C16 (C15), C18 (C17). 6. Renew IC30 (IC29). 7. Renew IC23. 8. Renew IC22.
Microdrive does not operate when selected	1. Check TR7 (TR6), TR5 (TR4). 2. Check the motor. 3. Check the mechanics. 4. Check HBC1 (2).
Tight cartridge tape · noisy operation · fails to format · jammed tape · damaged cartridges · microdrive incompatibility	Mechanical: (a) weak clamp spring — listen for clicks/crunching; (b) faulty or maladjusted rollers; (c) adjust the motor. If the motor has overheated, check for a buckled baseplate and renew the complete unit.
Head failure	Renew the complete Microdrive unit
LED failure	Renew the LED (it is push-fitted)
Unreliable loading/saving, machine locking up	Check the modification state. 1. Renew IC30 (IC29). 2. Renew IC23. 3. Check motor position and roller.
Continuous running of MDV1 (MDV2)	Renew TR7 (TR6)
Microdrives do not activate at all	Renew IC23 (ZX8302)
Erase function not working	1. Check TR3, D28. 2. Renew IC23.
MDV1 (2) does not run	Check TR7 (TR6), TR5 (TR4)

Faults on early Microdrives

Two faults are documented on some earlier Microdrive ULAs, both since remedied in production:

• Pin 1 latch-up — at power-on or during writing, the pin voltage sometimes latches at about 4 V and does not return to its normal signal or quiescent value.
• Data line not tri-stating — the faulty drive itself appears to work normally, but when inactive its data lines stay at about 4.5 V instead of going tri-state, preventing the other drive from working. Worth remembering when MDV2 is dead and MDV1 seems fine, or vice versa.

Removing a Microdrive

1. Unplug everything and open the case (see the Sinclair QL Maintenance Guide).
2. Remove the heatsink next to the Microdrives by releasing the +5 V regulator screw; keep the M3 × 10 mm screw with its brass plain and crinkle washers.
3. Remove the three pan-head self-tapping screws securing the drive — two at opposite corners of the chassis and a third accessible from the underside of the lower case. Do not confuse these with the countersunk motor screws.
4. Lift the drive out and free its two ribbon cables from their sockets by pinching the ribbon next to the socket and pulling straight up.

All from the Service Manual disassembly section.

Practical order of work

1. Check the write-protect tab and try a different, known-good cartridge.
2. Clean the head and the drive roller.
3. Verify the +5 V and +9 V supplies.
4. Run the mechanical checks: roller condition and gap, motor position, side spring, head position, head springs.
5. If a specific cartridge is at fault in several drives, re-felt or retire it.
6. Use the spin program and a scope to establish whether the head is producing signal.
7. Use the signal test program for the write, read and erase checks.
8. Only then substitute IC30/IC29, then the ZX8302 (IC23), then the ZX8301 (IC22).

Modern alternatives

Because Microdrive cartridges are consumable and increasingly scarce, many owners fit a solid-state replacement such as QL-SD, which presents SD-card storage to QDOS. This does not repair the drives, but it removes the need to rely on them for day-to-day use while the original mechanisms are preserved.

Related pages

• Sinclair QL
• Sinclair QL Maintenance Guide
• Sinclair QL Troubleshooting Guide
• Sinclair QL Capacitor Replacement Guide