Sinclair QL Troubleshooting Guide
This guide gives fault diagnosis for the Sinclair QL. It follows Section 4 (Fault Diagnosis and Repair) and Appendix D of the Sinclair QL Service Manual, which Sinclair published at full component level, so the QL is unusually well documented for fault-finding.[1]
Microdrive faults are covered separately on the Sinclair QL Microdrive Troubleshooting Guide.
โ ๏ธ Before you start
- There are no mains voltages inside the QL. The highest voltage on the board is the 44 V peak-to-peak AC feed from the external power pack.[1]
- The board is fragile. Use a low-power, earthed, temperature-controlled iron and short dwell times; see the soldering rules in the Sinclair QL Maintenance Guide.[1]
- Never plug or unplug the RGB monitor lead with the machine powered on. See the RGB section below โ this has destroyed ZX8301 chips.[2]
- Hairline solder splatter can measure several ohms and produce very misleading symptoms โ inspect the board visually before chasing anything complicated.[1]
- When substituting a suspect component, fit the suspect part into a known-good machine, not the other way round, so a faulty part cannot damage a working one.[1]

Normal power-on behaviour
At switch-on the yellow power LED lights and the machine should display, at bottom centre of the screen:[1]
F1...MONITOR F2...TELEVISION c 1983 Sinclair Research Ltd
Reaching this screen means most of the system is working. If it does not appear, switch off and try two or three more times before diagnosing โ the QL can lock up on start-up and appear lifeless.[1]
A useful quirk: if only 32 KB of the 48 KB ROM is readable, the splash screen appears with the correct window layout but no text in the windows.[3]
Reading the start-up screen
The QL's power-on screen is a genuine diagnostic. The Service Manual identifies these categories:[1]
| What you see | What it means | Where to look |
|---|---|---|
| Completely blank screen | Often a faulty ZX8301 | IC22 is socketed โ substitute it. If no spare, check the CPU clock and the RGB outputs are present |
| "Tweed" pattern โ fine green, white, red and black dots moving quickly down the screen | The RAM test has started, so the ZX8301 works and ROM is being read and executed | Start fault-finding at the RAMs |
| Geometric pattern (e.g. red and green vertical bars), machine goes no further | ZX8301 works but ROM code is not executing and RAM is not being overwritten | Missing 8301 clock to the CPU, faulty RAM, or an absent CPU |
| RAM test completes, geometric pattern, then blank | ROM, IPC communication or interrupt problem | ROM fault; ZX8302โIPC comms (check crystal X4 oscillating, activity on IPC pins 29 and 35); or a continuous interrupt to the CPU |
| Random flickering patterns with excited beeping | ROM not being read correctly | Suspect the ROM; check the ROM enable signal on ZX8301 pin 33 |
| Plain white screen | RAM test part 1 failed (cannot store data) | RAM, or an address/data line short |
| Plain green screen | RAM test part 2 failed (data not retained) | Refresh failure, or a short on an address or data line |
| Almost-white or almost-green screen with narrow vertical stripes or a pattern of dots | A single bit is faulty | One RAM chip or a shorted data/address line โ use the probe method below |
| Tweed pattern after RESET that reverts to white | Failure found in the second bank of RAM | Second RAM bank |
A white or green screen is actually good news: it means the RAMs are broadly working and the machine is close to functional.[1]
Pinpointing a faulty RAM chip
Sinclair's method for finding the failed device when the screen shows stripes or dots:[1]
- Connect one end of a probe to ground.
- Touch the probe onto the data pin of each RAM in turn and watch the screen.
- Shorting a data line puts a vertical black line (or lines) on the screen. When that black line coincides with the existing pattern of dots or stripes, you have found the offending chip.
- Because the two RAM banks share the data bus, the fault could be either bank's device (for example IC1 or IC9). Change the memory-mapped-area device first โ IC1 in that example.
As a rule of thumb: a long vertical stripe suggests a shorted data line; dots suggest a faulty RAM chip.[1]
An alternative for non-socketed devices is a "test IC" made from an IC test clip with a known-good device (a 4164 for RAM) attached, bridged across each suspect chip in turn.[1]
The Minerva ROM is also useful here: its memory test flags the first faulty address, identifying the chip directly. The standard QL memory test will not detect RAM errors where null bits are written.[4]
Dead machine: supply checks first
A missing copyright screen means a fundamental failure. Check the supplies before anything else:[1]
| Function | Test point | Expected |
|---|---|---|
| ยฑ12 V regulator input | + side of D26 | 15.6 V AC ยฑ 2.0 V |
| +5 V regulator input | + side of C41 | +9.0 V ยฑ 2.0 V |
| +12 V regulator output | + side of C38 | +12.0 V DC ยฑ 0.25 V, no discernible ripple |
| +5 V regulator output | + side of C42 | +5.0 V ยฑ 0.15 V, no discernible ripple |
The external pack supplies 9 V DC at 2 A and 44 V peak-to-peak AC into J8; it is unregulated apart from a thyristor limiting the DC peaks, and has significant ripple by design.[1] On the board, IC37 (78L12) and IC36 (79L12) derive +12 V at 80 mA and โ12 V at 50 mA from the AC feed, and IC35 (7805) regulates the DC feed to +5 V.[1]
- Loss of power โ check IC37 and IC36.[1]
- +5 V low โ if below about 4.8 V, renew the 7805. Re-grease it and make sure the heatsink screw is tight and the whole heatsink gets hot.[4]
- Noisy PSU โ change the power pack.[1]
- Clean the mains plug, its fuse holder and the QL's power socket. The power socket often develops dry joints, particularly on machines used with the case off.[4]
Signals to check on a dead board
If the supplies are good, work through the key signals. The Service Manual lists these checkpoints:[1]
| Device | Pin | Signal |
|---|---|---|
| 68008 (IC18) | 28 | RESETOUTL |
| 68008 | 30, 31 | XTAL |
| 68008 | 25 | CLOCK |
| 68008 | 31 | DTACK โ must be present and regular or the RAM test can never start |
| ZX8301 (IC22) | 7 | CLOCK output (15 MHz crystal X1 divided by two) |
| ZX8301 | 12 | CSYNCL |
| ZX8301 | 30, 31, 32 | RED, GREEN, BLUE |
| ZX8301 | 33 | ROM enable (ROM0EH) โ absent means ROM cannot execute |
| ZX8301 | 8, 9, 10 | RASL, CAS0L, CAS1L (DRAM strobes) |
| ZX8302 (IC23) | 25 | Clock |
| IPC (IC24) | 29, 35 | Serial link to the ZX8302 |
If DTACK is missing, check for a short on the DTACK line and check the ZX8301 โ in particular DSMCL, which disables the 8301 when held high.[1]
Where you have no better starting point, compare IC23, IC18 and IC22 against a known-good board. If the fault persists, substitute the socketed devices in the order IC18, IC22, IC23, IC24, IC28, powering up after each change.[1]
โ ๏ธ RGB monitor output โ the QL's weak point
The RGB output is one of the most commonly reported failures on surviving QLs, and it fails in two distinct ways: mechanical damage at the socket, and electrical damage to the ZX8301.
Why it fails
The ZX8301 generates R, G, B, CSYNCL and VSYNCH and these run to the 8-pin DIN socket J7.[1] The critical design weakness is that the video pins are taken from the ZX8301 to the socket with no series resistance or protection, so anything that happens at the connector reaches the chip directly.[2] The consequences:
- Certain issues of the ZX8301 are very fragile, and simply unplugging the RGB connection while the machine is running has been known to destroy the chip.[2]
- Differences in earth potential between the QL and a monitor can blow the ZX8301 โ this has been reported specifically when leads are removed with equipment powered, and with certain monitors (Microvitec is named).[4]
- Sinclair's own manufacturing acknowledged the problem: Samsung-built QLs have protection diodes fitted piggyback onto the ZX8301's RGB video output pins as a production measure.[3]
The socket itself is a through-hole DIN connector on a board with fine tracks and no great mechanical reinforcement. Repeated insertion, a heavy DIN plug, or sideways force on the lead flexes the joints until they crack and the pads lift โ the same failure mode that affects the other rear connectors.[1]
Symptoms and what they point to
| Symptom | Likely cause | Action |
|---|---|---|
| Picture appears and disappears when the lead is moved or wiggled | Cracked solder joints at J7, or a lifted pad | Reflow all eight pins; inspect pads under magnification; repair broken tracks with wire links |
| One colour permanently missing (e.g. no red) | Broken track or cracked joint on that pin โ or that output of the ZX8301 has failed | Buzz out the pin from J7 back to the ZX8301; if continuity is good, suspect IC22 |
| No RGB picture at all, but the machine clearly boots (TV output works, Microdrives run) | Sync or all-colour loss at the connector, or a failed ZX8301 video section | Check CSYNC and RGB at J7; renew IC22 if signals are absent at the chip |
| No colour on the monitor at all but a monochrome image is present | Loss of the RGB lines with sync intact | Check the RGB pins and their tracks |
| Machine died immediately after plugging or unplugging a monitor lead while powered | ZX8301 destroyed | Renew IC22 (socketed); fit protection before reconnecting |
| No colour monitor output (general) | Service Manual: renew IC22 | Substitute the ZX8301 |
| Blurred picture; picture drifting; gradual colour loss in use | Service Manual: renew IC22 | Substitute the ZX8301 |
Sources: Service Manual fault-finding guide and the video notes.[1][2]
The 8-pin DIN pinout
| Pin | Signal | Level | Series resistor for a 75 ฮฉ input |
|---|---|---|---|
| 2 (and shell) | Ground | โ | โ |
| 3 | CSYNC (composite sync, active low) | TTL | 1.2 kฮฉ |
| 4 | VSYNC | TTL | 10 ฮฉ |
| 6 | RED | TTL | 680 ฮฉ |
| 7 | BLUE | TTL | 680 ฮฉ |
| 8 | GREEN | TTL | 680 ฮฉ |
Pinouts and resistor values from the Classic Console Upscaler Wiki, which notes that several conflicting QL pinout diagrams circulate, so verify against your own machine before wiring anything.[2]
The series resistors are not optional. The QL's outputs are TTL-level; driving a 75 ฮฉ RGB input directly, without the resistors, can damage the machine.[2]
Repairing the socket
- Work with the board out of the case and supported flat.
- Add fresh flux and leaded solder to each of the eight pins, then reflow them properly โ do not simply touch the iron on the old joint.
- Inspect every pad under magnification for lifting or a ring crack. A pad that has lifted but is still attached to its track will fail again; a pad whose track has broken needs a link.
- Repair broken tracks with insulated wire links soldered from the connector pin to the next solid point on the net (a via, component leg or IC pin), rather than trying to rebuild fine copper.
- Buzz out each pin to its destination before powering up.
- Add strain relief โ anchoring the lead or the socket so that force on the cable is not carried by the solder joints.
- Consider fitting series protection on the video lines. Samsung's factory fix was diodes on the ZX8301 video pins; series resistors in the cable protect against low-impedance loads.[3][2]
Prevention: always power the QL and the monitor down before connecting or disconnecting the video lead, and switch the QL and monitor on the same mains outlet or switched strip to avoid earth-potential differences.[4]
TV and composite video faults
The RGB and sync signals also feed IC28, which produces composite PAL for the internal UHF modulator M1; TR9 mixes the same signals for a composite monochrome output.[1]
| Symptom | Action |
|---|---|
| No TV picture | Try tuning slightly either side of the expected point โ the modulator may have drifted. Renew IC35 if the output is distorted or absent (Service Manual "TV output distorted / No TV output") |
| Picture but no colour on TV | Check crystal X3 is oscillating; look for colour modulation on the luminance signal. If chroma and luminance are both present but the TV will not lock, the crystal frequency is wrong; if no chroma at all, the chroma oscillator is not running |
| Chroma present, still no TV colour | Changing the value of C31 slightly may restore it; check RGB inputs to IC28, then consider renewing IC28 |
| No colour TV output | Check C35, C36, R85, R86; check C29, R54 and the components around IC28; renew modulator M1; renew IC28 |
| Dark TV screen | Renew modulator M1 |
| No black-and-white TV output | Renew TR9 |
| Fails colour test / colour fades | Renew crystal X3; check components around IC28; renew IC28 |
| TV picture drifting continuously | Check C23, C24, C25 |
To test whether the modulator itself is faulty, remove IC28 and feed the VIDEO signal from TR9 in via R85 โ this gives a low-contrast black-and-white picture.[1]
Keyboard faults
The keyboard is an 8 ร 8 matrix with partial diode isolation, on a membrane connected at J11 and J12.[1]
Most apparent keyboard faults are the membrane or the connectors rather than the ICs. If a whole row or column is dead, suspect a connector. If a pair of keys misbehaves, suspect a short between them; two dead rows likewise suggests a short โ check with an ohmmeter.[1]
To watch the scan, trigger a two-channel oscilloscope from one keyboard output (for example KB01) and confirm the other outputs go high at successive intervals; a short shows up clearly as disturbed scan timing.[1]
For the symptom-to-remedy table (membrane vs bubble mat vs IPC) and the membrane tail repair, see the Sinclair QL Maintenance Guide.
If the QL locks up but an on-screen clock keeps running, or a few keys do not work, suspect the keyboard ribbons โ clean the tail ends and sockets, and if the ends are damaged on the older type, trim 5 mm off.[4]
Serial, network, sound, clock and joystick
| Area | Symptom | Action |
|---|---|---|
| RS232-C / Printer | Fails the RS232 loopback test | Check J5, J6; renew IC25, IC26; renew IC23; renew IC24 |
| Network | Fails the network test | Check J9, J10; check TR2 and its associated components |
| Sound | No sound | Check TR1; check R104 (post-Issue 6); renew IC24 |
| Real-time clock | Fails the real-time clock test | Check R91, R80, C1, C53 |
| Joystick | Fails the joystick test | Check J3, J4, J2 |
| Reset | Does not reset | Check the reset switch and its spring |
| Display | Erratic display after warm-up | Check the RAMs |
| General | Locks up or crashes at random after prolonged use | Renew crystal X2; renew IC23; renew IC34; renew IC24 |
Source: Service Manual.[1]
Serial ports that will not transmit. If the +12 V and โ12 V rails are good but outgoing RS-232 does not work, fit the two 33 kฮฉ pull-ups from ZX8302 pins 19 and 21 to the โ12 V rail (mandatory modification (g)); this very often cures it. An open network socket โ centre pin not making contact with no plug inserted โ has also been known to affect the serial ports. Failing that, reseat or replace IC25 (1488), and try different ZX8302/1488 combinations.[4][1]
A loopback cable between J5 and J6 makes a good overall system test: RS-232 reception passes through the IPC, the ZX8302 and back onto the processor data bus, so if that works the whole system is close to proven.[1]
Random crashes and lock-ups
The QL has a reputation for instability, and much of it is contact-related rather than component failure:[4]
- Remove all expansions, external ROMs and upgrades and see whether the fault persists.
- Remove, clean and firmly reseat every socketed IC; clean the ROM cartridge edge connector with an eraser on the solder-side tracks. This is a very common cause of crashing.
- Verify the +5 V rail and the regulator's heatsinking.
- Check the machine is to the latest build standard and that the mandatory modifications are fitted.
- Run a thicker wire from the 5 V line at the regulator to the RAM 5 V rail (the thick track running from just above R68 and under IC7) if expansions are marginal.
- Make sure expansion connector pins are straight and the socket is clean.
- Switch all equipment on from one switch where possible, to avoid earth-potential differences between the QL and its monitor.
Where a second working QL is available, swapping all socketed ICs across and then replacing them one at a time is a quick way to isolate the culprit.[4]
Repair notes
- Use recognised desoldering and heatsinking technique; only a suitably earthed, low-power iron.[1]
- Ensure good contact between the IC35 regulator body and its heatsink, with thermal grease applied.[1]
- Renew the modulator as a complete unit.[1]
- Use proper IC removal and insertion tools; avoid contaminating pins by handling.[1]
- Inspect for solder splatter after every repair.[1]
Common fault catalogue
- Dead, no display, power LED on โ check the supply rails first, then IC22, IC24, the RAMs, IC34, IC18, IC33.
- Blank screen โ most often the ZX8301 (IC22).
- White or green screen โ RAM fault or an address/data line short; use the probe method.
- Stripes or dots on white/green โ single failed RAM bit; probe each RAM data pin.
- Flickering patterns with beeping โ ROM fault.
- Splash screen with empty windows โ only 32 KB of ROM readable.
- No RGB picture, TV works โ J7 solder joints, lifted pads, or the ZX8301.
- Picture comes and goes when the lead moves โ cracked joints at the RGB socket.
- Machine died on unplugging the monitor lead โ ZX8301 destroyed; never hot-plug video.
- Dead keys / lock-ups with clock running โ keyboard membrane or its tails.
- No outgoing RS-232 โ fit the 33 kฮฉ pull-ups to โ12 V on ZX8302 pins 19 and 21.
- Random crashes โ dirty IC sockets and ROM edge connector; marginal +5 V.
- Microdrive problems โ see the Sinclair QL Microdrive Troubleshooting Guide.
Related pages
- Sinclair QL
- Sinclair QL Maintenance Guide
- Sinclair QL Capacitor Replacement Guide
- Sinclair QL Microdrive Troubleshooting Guide
- Sinclair ZX Spectrum Troubleshooting Guide
References
- โ 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 1.13 1.14 1.15 1.16 1.17 1.18 1.19 1.20 1.21 1.22 1.23 1.24 1.25 1.26 1.27 1.28 1.29 1.30 1.31 1.32 1.33 1.34 1.35 Sinclair QL Service Manual, Sinclair Research Ltd. Source for the power-up diagnostic sequence, screen-pattern interpretation, supply test-point voltages, the RAM bit-probing method, the IC substitution order, the video fault notes, the fault-finding guide tables, the repair notes and Appendix D (Microdrive Fault Finding).
- โ 2.0 2.1 2.2 2.3 2.4 2.5 2.6 "Sinclair QL", Classic Console Upscaler Wiki. Source for the ZX8301 fragility warning, the 8-pin DIN pinout, the required series-resistor values and the video timing figures.
- โ 3.0 3.1 3.2 Kรถnig, Urs. "Sinclair QL Build Standards", sinclairql.net. Source for the 32 KB-ROM symptom, chipset and PCB issue variants, ROM version list, and the Samsung RGB protection diodes.
- โ 4.0 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 Firshman, Tony. QL Troubleshooting, TF Services; reproduced in the Sinclair QL Homepage archive. Source for the Minerva RAM-test advice, the socket-cleaning and edge-connector notes, the 7805 test and replacement advice, keyboard ribbon repair, the โ12 V serial-port pull-up fix, IC removal technique and the monitor-earth ZX8301 damage note.