BBC Master 128 Troubleshooting Guide
(Redirected from Master Turbo Troubleshooting)
Fault-finding for the BBC Master 128 and the machines built on its board (Master Turbo, Master 512, Master AIV and Master ET). It follows section 7 of Acorn's Master series service manual and Acorn's Master 128 Repair Guide, with IC numbers from Acorn's circuit diagram 0143,000; repairers' reports from stardot are added where they fill gaps.[1][2] For the Master Compact, which has a different board, see BBC Master Compact Troubleshooting.
Before you start
[edit | edit source]Acorn's first checks, before any component-level work:[1]
- no loose connectors or broken cables;
- no broken or shorting tracks;
- every socketed IC seated correctly;
- the supply working and reaching the components concerned;
- clean logic levels: TTL above 2.4 V for 1 and below 0.5 V for 0, CMOS above 3 V and below 1 V.
Its hints: never solder with the machine switched on; remove all user ROMs first (and put them back afterwards); use sharp meter probes that push through the solder resist; always suspect connectors. Disconnect any second processor or internal co-processor before starting.[1][3] The peripheral bus controller IC21 buffers the data bus and isolates parts of the system from each other, so a data bus fault may show up in an unexpected place: Acorn's example is a failed user VIA stopping the system VIA.[1]
Switch-on
[edit | edit source]Connect both a UHF television and an RGB monitor; using both shows up faults in the leads and wrong CSYNC polarity. Switch on and compare with this table.[3]
| Symptom | Probable cause | Action |
|---|---|---|
| Noise on both screens, no beep, keyboard power LED off (7.1.1) | Supply dead or its fuse blown | Check the supply; Acorn says to try a replacement anyway. Listen to it: a ticking supply should be swapped for another.[2] |
| Power LED lit, nothing else (7.1.2) | Board fault | Core checks below |
| A display with rolling lines or diagonal stripes (7.1.3) | CRTC or sync | Check the 14 MA lines (IC22 pins 4–17) and 4 RA lines (pins 35–38) reach the CRTC multiplexer IC31; VSYNC (IC22 pin 40) every 20 ms; HSYNC (pin 39) every 64 µs; CSYNC at IC25 pin 8 and SK9 pin 4. If only RGB is wrong, change the CSYNC polarity with LK5 |
| Cursor stuck at top left (7.1.4) | Keyboard, or a co-processor | Disconnect any co-processor; check the keyboard connections to the encoder IC16; replace IC16 |
| "Acorn MOS" and "Acorn ADFS" only (7.1.5) | ADFS waiting for a disc, or bad configuration | Press CTRL F BREAK, then fit an ADFS disc or type *CONFIGURE NODIR. Otherwise switch on holding R, then CTRL F BREAK, to reset the configuration
|
| "BASIC" but no > prompt (7.1.6) | IRQ stuck | Check that IRQ at IC14 pin 4 is not stuck; check the address and chip select lines of the system VIA IC8 |
| > prompt, keyboard dead (7.1.7) | System VIA, encoder or keyboard; IRQ held | Check IC8 and IC16 connections, try another keyboard, check IRQ (IC14 pin 4). The memory controller IC20 and the clock chip IC11 also drive IRQ; LK4 isolates IC11 |
| Some keys dead (7.1.8) | Keyboard matrix or encoder connections | Several keys: connections to IC16, or a break in the middle of a row or column. One key: a break at the end of its row or column, or the switch itself |
| CTRL BREAK behaves as BREAK (7.1.9) | Clock chip power-down circuit | Check the parts around IC11; see C11 below. IC11 CE (pin 13) should be high with occasional low pulses, AS (pin 14) low, DS (pin 17) low; R/W (pin 15) must change on reads and writes |
| Starts in terminal mode and *CONFIGURE will not change it (7.1.10) | As 7.1.9 | As 7.1.9, then check every connection to IC11 |
| Random dots or characters (7.1.11) | Memory system, rarely the DRAMs themselves | Check CPU address lines (IC14 pins 9–20, 22–25) to the multiplexers IC29, IC30 and IC33 at CMOS levels; address stable at least 15 ns after RAS falls; noise on lows under 500 mV; data bus to the DRAMs |
| Two or more copies of what you type (7.1.12) | CRTC addressing | Check test points TP7–TP20; MA13 (IC22 pin 17) static in any one mode; IC22 to IC31; C0 and C1 from the 74LS259 IC10 (pins 4 and 5) to IC31 (pins 38 and 39) |
Several dead keys can also come from the slow data bus that the keyboard shares with other chips. On one Master 512, line SA0 (system VIA IC8 pin 2, also to the keyboard encoder IC16 pin 34, the sound chip IC12 pin 10 and the clock chip IC11 pin 4) was held at about 0.04 V, giving a corrupt configuration, a constant background tone and large parts of the keyboard dead. Two new 6522s made no difference; the 146818 clock chip, sitting near old battery leakage, was holding the line down, and a new one cured it.[4]
Core checks: dead machine
[edit | edit source]For a machine with power but no picture, Acorn's sequence (7.1.2) is:[3]
- Check that every IC has +5 V and 0 V on the right pins, and make good any dry joints.
- Feel for hot ICs and look for shorted tracks. Watch for a short from −5 V to the I/O controller IC15, which latches it up. Replace any IC that got abnormally hot, even if it works afterwards.
- Check the 16 MHz oscillator output at IC43 pin 2 and that it reaches the VIDPROC at IC42 pin 8: a reasonably clean 4 V peak-to-peak signal. The oscillator is half of IC43 with crystal X2.
- Check 1, 2, 4 and 8 MHz from the VIDPROC at IC42 pins 4, 5, 6 and 7. If any is missing or not a clean square wave, replace IC42.
- Check for a clean 2 MHz clock at the CPU, IC14 pin 37.
- Check that no data or address line is shorted to another, to +5 V or to ground.
- Check that the system ROM is plugged in.
- Check RAS (pin 5) and CAS (pin 16) at the main DRAMs IC17 and IC23: clean 4 MHz square waves. RAS comes from the D-type IC28 pin 9; CAS from 2 MHz through two NAND gates in IC34.
- Check that the auxiliary DRAM CAS line is low when inverted at IC36 pin 8; if high it suppresses the main DRAMs.
- Check the multiplexed DRAM address lines (IC17 and IC23 pins 6–8 and 10–14) switch after RAS with about 20 ns hold time. If the DRAM timing is out, check that C26 is fitted and correct.
Warning: if the 16 MHz clock stops, RAS and CAS can be held low and the DRAMs overheat and fail. Acorn says not to leave a machine with a dead clock switched on for more than 30 seconds at a time.[3]
| Point | Expect |
|---|---|
| IC43 pin 2 → IC42 pin 8 | 16 MHz, about 4 V peak-to-peak |
| IC42 pins 4, 5, 6, 7 | 1, 2, 4, 8 MHz square waves |
| IC14 pin 37 | 2 MHz CPU clock |
| IC17, IC23 pin 5 and pin 16 | RAS and CAS, 4 MHz |
| IC36 pin 8 | Low (auxiliary CAS inverted) |
| IC22 pin 40, pin 39 | VSYNC every 20 ms, HSYNC every 64 µs |
| IC14 pin 4 | IRQ not stuck high or low |
Long beep and a blank screen
[edit | edit source]A Master that gives a continuous beep with no picture is a common report on stardot. Repairers suggest this order:
- Measure the supply under load. One machine with this fault read +4.72 V; the acceptable range with a 12 V 21 W bulb as load is 4.85–5.15 V.[5]
- Check that the CPU's reset input (IC14 pin 40, also IC39 pin 1) is high except while BREAK is held. 1024MAK names the system VIA IC8 and the 4464 DRAMs (IC17, IC18, IC23, IC26) as the most likely component failures after the supply, and the MOS ROM IC24 seated badly as another cause.[6]
- Reseat the keyboard connectors: a badly connected keyboard also gives the long beep with no output.[6]
- Suspect the DRAMs. One repairer found two Masters that would not start had failed 4464s that worked when warmed and failed again when cool; he socketed the replacements. Use 120 ns parts.[7] Acorn's own guide adds that two or more bad data bits within one nibble point to the DRAM connected to that nibble.[2]
- A test ROM in place of the MOS can show whether the CPU and memory run at all; tom_seddon's beeb_test_os has a Master version.[7]
Memory and sideways RAM
[edit | edit source]- Sideways RAM faults (7.2): check all connections to the memory address multiplexers, RAS and CAS at the DRAMs, and links LK18 and LK19.[8]
- Memory faults are rarely the DRAMs themselves according to Acorn; check the address lines and timing first (7.1.11).[3]
- A memory test program is rarely needed; Acorn suggests pressing BREAK repeatedly while watching each memory controller output (AA and AT lines from IC20 to the multiplexers IC29 and IC33).[9]
Video
[edit | edit source]Connect a television, an RGB monitor and a composite monitor together to see which outputs are affected.[8]
| Symptom | Check |
|---|---|
| Picture rolls vertically on every output | VSYNC at IC22 pin 40, through the exclusive-OR gates IC25. Trigger the oscilloscope from IC22 pin 40 to see VSYNC within CSYNC |
| Picture broken into moving diagonal lines on every output | HSYNC at IC22 pin 39 through IC25 |
| Fault on composite only | R147 to R150 and the signals to the output transistors |
| Fault on UHF only | The resistor network from IC40 pins 12–19 (R69, R75, R77, R78, R84, R85, R87, R89) |
| No colour on UHF | VC2 set for 17.7345 MHz at IC43 pin 12; IC40 pin 11 and the summing node (L1, C81, Q12, C88, R140); +5 V at IC40 pin 20; IC40 last |
| Only one colour, RGB | RGB from the VIDPROC (IC42 pins 14, 12, 10); in modes 7 and 135 only, from the SAA5050 (IC32 pins 24, 23, 22) to IC42 pins 13, 11, 9 |
| Only one colour, UHF | RGB into IC40 pins 6, 7, 8; R69, R75, R77, R78, R84, R85, R87, R89, R90 |
| Only modes 7 and 135 fail | 6 MHz into the SAA5050 from IC39 and IC25 (no part of the waveform narrower than about 40 ns); pull-ups R51, R54, R56; data bus to the PBC IC21 pins 21–28; VSYNC for half-height text |
Acorn's later fixes for Teletext break-up and for missing colour on the composite output are under Acorn's field fixes below.[2]
Clock and configuration
[edit | edit source]| Symptom | Probable cause | Action |
|---|---|---|
| Time shows strange characters | Clock never set | Set the time and try again[8] |
| Clock does not count | 32.768 kHz oscillator | Check VC1 is in circuit and 32.768 kHz is at IC11 pin 2; check X1, VC1, C6, C7, R17, R24[8] |
| Clock loses a lot of time; settings lost | Flat battery or poor PL8 contact | PL8 on all three pins and the leads soldered to the clips; at least 2.6 V at the clock chip with the mains off; if not, check FET Q4[8] |
| "This is not a language" and a * prompt | Language not configured | *CONFIGURE LANG 12 and *INSERT 12; if it recurs at the next switch-on, replace the battery pack[10]
|
| *CAT clicks but the drive does not spin | DFS not configured or unplugged | *CONFIGURE FILE 9 and *INSERT 9[10]
|
| Configuration lost at power-off with a good battery | Early clock chip power-down circuit | Acorn's C11, D8 and Q3 change (below)[2] |
A configuration reset is done by switching on with R held, which displays "CMOS RAM reset" and asks for BREAK, then CTRL F BREAK.[3][11] Battery replacement and the PL8 corrosion that a leaking pack leaves behind are covered in the maintenance guide.
Interfaces
[edit | edit source]| Interface | Symptom | Checks |
|---|---|---|
| Analogue port | Wrong results on all channels | Plug wiring; C49, which Acorn's text calls the integration capacitor (on the circuit diagram C49 is the 1 µF tantalum on the reference, and the integration capacitor across the CI pins, 4 and 6, is C36, 33 nF)[12]; about 1.8 V reference across IC49 pins 8 and 9; inputs 0 to 2.5 V; a pin connected to more than 5.5 V is likely to have damaged IC49; chip select from IC15 pin 35 to IC49 pin 23; 1 MHz at IC49 pin 2 |
| RS423 | Not working at all | Format and baud rate (reset to the configured values at power-on and CTRL BREAK); the connector can be inserted upside down; selects from IC15 pins 34 and 36; 1.23 MHz at the SERPROC IC48 pin 25; −5 V at the driver IC50 pin 8; CTS (IC48 pin 20) and RTS (pin 23) |
| Cassette | Will not load | Rx clock at the ACIA IC45 pin 3; about 1.2 V peak-to-peak at the SERPROC CASIN, IC48 pin 12; the static protection network and the LM324 |
| Cassette | Will not save | Tx clock at IC45 pin 4; data at CASOUT, IC48 pin 27 |
| Disc | Not working at all | Both cables home; another drive; select inputs of the 1770/1772 (IC3 pin 1) and 74LS174 (IC4 pin 1); FDCON and FDC from IC15 pins 1 and 4 |
| Disc | Drive light on, no catalogue | ADFS or DFS selected; disc formatted; Read Data at IC3 pin 19; pull-ups R1, R3, R5, R6; 8 MHz at IC3 pin 18 from IC19 pin 4 and IC42 pin 7; DRQ (IC3 pin 27) pulsing every 32 µs and inverted at the CPU NMI (IC14 pin 6); INTRQ (IC3 pin 28) |
| Disc | Catalogue corrupted, or data unreadable | C27 (220 pF, DRQ to ground) where the board has a place for it; only one drive with terminating resistors |
| Disc | Wrong data, or "Bad FS map" | Memory controller IC20 address inputs (pins 21–28) and its AA and AT outputs to IC29 and IC33 |
| Disc | Reads but will not write reliably | The built-in formatter is DFS only; write data and write gate at IC3 pins 22 and 21 (a drive must be connected, as the pull-ups are in the drive) |
| Printer | No output | *CONFIGURE PRINT 1; pull-up R11 on the user VIA IC6 pin 40; port A (IC6 pins 2–9) through the 74LS244 IC5; strobe at IC6 pin 39 and through the 7438 IC2 pins 11 and 8
|
| User port | Faults | Software that pokes the I/O processor from a co-processor; the 6522 shift register on the CA lines is not supported; a cable left trailing on carpet can damage the user VIA IC6 |
| 1 MHz bus and Tube | Device not found | Configuration (for example a floppy configured as default so a hard disc seems dead); pages &FC and &FD not selected internally; cables under 1 m; READY from IC15 pin 17 to IC14 pin 2 and IC21 pin 11; IC15 pin 38 to IC21 pin 5; IC20 address lines; IC15 pins 18 and 19 |
| Internal co-processor | Not recognised; or only a cursor at top left | Configuration (*CONFIGURE TUBE, INTUBE); both connectors mated; IC15 pins 18 and 19; boot ROMs seated; try the co-processor in another Master |
| Sound | None | Speaker plugged in to PL9; no short or very low impedance (under 4 Ω) on the audio output socket. Sprow adds: reseat the purple −5 V supply lead[10] |
| Sound | Strange sounds on key presses | Connections to the sound chip; press BREAK (mains noise); on some early computers IC10 pin 15 was connected to PRST and should go to +5 V |
| Cartridge port | Some cartridges not recognised | Electron cartridges are not all compatible; links; worn or bent edge connector contacts (replace the connector) |
| Econet | Not working | Both module connectors fitted (the longer one has two spare pins at its left, and a displaced module will not work); NMI at IC14 pin 6 (TP3) not held low |
Sources: Acorn's service manual, sections 7.5 to 7.17.[9][13] A machine on an Econet that reports "No clock" needs the network clock; off a network, remove the ANFS ROM.[10]
Acorn's field fixes
[edit | edit source]Acorn's notes to the Master 128 Repair Guide list four faults found in service and their fixes:[2]
| Problem | Fix |
|---|---|
| Loss of configuration | Replace C11 with 100 µF (0610,100); replace D8 with a 1N4001 (0794,001); fit a 1.8 kΩ resistor (0502,182) between base and emitter of Q3 on the solder side |
| Break-up in mode 7 | Add 100 pF to C30; needed only on machines bought before 3 February 1986 |
| Machine crashes after running for a while; confirm by heating the 74S04 IC43 to about 55 °C (usually National Semiconductor parts) | Change C85 and C86 from 100 pF to 10 nF (0629,010) |
| No colour on the composite output | Fit a 470 pF capacitor between the east legs of R137 and R153, beside the modulator, on the component side |
Power supply
[edit | edit source]| Symptom | Probable cause | Action |
|---|---|---|
| Bang, smoke and a smell, machine may keep running | Failed X2 suppression capacitor | Replace both X2 capacitors with class X2 parts of the same value[14] |
| Slow to start, or starts only after several seconds | Start-up capacitor, or the 180 kΩ resistors R2 and R3 that charge it | Replace the start-up capacitor; measure R2 and R3[15] |
| Dead, and one primary capacitor stays charged | A bleeder resistor open circuit | Measure R2 and R3 (180 kΩ); one was found open circuit in each of two reported repairs[16] |
| Ticking, power LED flashing | Start-up capacitor; secondary capacitors; overload or short on the outputs | Replace the start-up capacitor first, then the secondary capacitors; test with the board disconnected[17] |
| +5 V low under load | Ageing capacitors, especially the start-up capacitor, or the feedback circuit | Bulb load test; renew the start-up and secondary electrolytics[5][18] |
| Machine runs but no sound or cassette | −5 V missing | Check the purple lead and the −5 V output[5] |
Capacitor values and the removal sequence are in the BBC Master 128 Capacitor Replacement Guide.
Related pages
[edit | edit source]- BBC Master 128
- BBC Master 128 Maintenance Guide
- BBC Master 128 Capacitor Replacement Guide
- Acorn Master Turbo (co-processor fault finding)
- BBC Master Series Service Manual
- Capacitor Failure Symptoms
References
[edit | edit source]- ↑ 1.0 1.1 1.2 1.3 Acorn Computers, British Broadcasting Corporation Master Series Microcomputer Service Manual, part no. 0443,004, issue 1, April 1986, 7 "Fault finding", pp. 29–31: points to check first (connectors, tracks, seating, supply, TTL and CMOS logic levels), hints for repair, test equipment, and the warning that the peripheral bus controller IC21 can disguise data bus faults.
- ↑ 2.0 2.1 2.2 2.3 2.4 2.5 Acorn Computers, Master Series Hardware Specification, hosted as File:Acorn BBC Master Series Hardware Specification.pdf: "The Master 128 Repair Guide", PDF pp. 17–33 (including "listen to the power supply ... is it ticking?", the CTRL G beep test, and "if two or more data bits are faulty within a nibble, suspect the DRAMs"), and "Additional information relating to Master Repair Guide", p. 50.
- ↑ 3.0 3.1 3.2 3.3 3.4 3.5 Acorn Computers, Master Series Microcomputer Service Manual, 0443,004, 7.1 "Switch on", 7.1.1 to 7.1.12, pp. 32–37.
- ↑ stardot.org.uk, "Master 512 keyboard issues - 6522? (Solved - bad RTC)" (November–December 2024, retrieved 2026-10-02). Boydie: SA0 stuck low, constant tone, many dead keys; replacing IC8 twice made no difference; cutting pin 4 of IC11 freed the line and a new 146818 completed the repair. 1024MAK: the lines SA0 goes to (IC8 pin 2, IC16 pin 34, IC12 pin 10, IC11 pin 4) and how to measure a stuck line. Secondary source.
- ↑ 5.0 5.1 5.2 stardot.org.uk, "Master 128 and the long slow beep of doom" (May–October 2019, retrieved 2026-10-02). FelisSapien: power LED lit, blank screen, long low beep, +4.72 V. 1024MAK: the 12 V 21 W bulb test, 4.85–5.15 V acceptable; the main logic runs from any regulated 5 V supply; +12 V is only for disc drives and −5 V only for cassette and sound; R2, R3, C9 and Q2 in the supply. Secondary source.
- ↑ 6.0 6.1 stardot.org.uk, "BBC Master 128 beep of death" (June 2021 – March 2022, retrieved 2026-10-02). 1024MAK: check the reset line at IC14 pin 40 or IC39 pin 1; after supply faults the likely failures are the system VIA IC8 and the 4464 DRAMs IC17, IC18, IC23, IC26; check the MOS ROM IC24 contacts. marcusjambler: 16 MHz from IC43 to IC42 pin 8 and the divided clocks at IC42 pins 7, 6, 5, 4; he has replaced the 1772 and a DRAM on Masters. WrightStuff: a badly connected keyboard also gives the long beep and no output. Secondary source.
- ↑ 7.0 7.1 stardot.org.uk, "Faulty master mainboards" (May 2024, retrieved 2026-10-02). Ukwebb: three boards with relay click, continuous low beep and no video; system VIA swapped without effect. maniacminer: two non-starting Masters had failed 4464 DRAMs that worked when heated; 120 ns grade. hoglet: screen suspect DRAMs in a working Master with socketed RAM. tom_seddon: a Master-compatible test ROM, beeb_test_os. Secondary source.
- ↑ 8.0 8.1 8.2 8.3 8.4 Acorn Computers, Master Series Microcomputer Service Manual, 0443,004, 7.2 "Sideways RAM", p. 37; 7.3 "Video", pp. 38–40; 7.4 "Real time clock", p. 40.
- ↑ 9.0 9.1 Acorn Computers, Master Series Microcomputer Service Manual, 0443,004, 7.5 to 7.8, pp. 41–46: analogue port, RS423, cassette and disc interface.
- ↑ 10.0 10.1 10.2 10.3 Sprow, "The most common beeb power on problems", Sprow's webpages, updated 9 November 2010 (retrieved 2026-10-02). Community source: ADFS mount hang (*CONFIGURE NODIR and NOBOOT); "No clock" on Econet machines; "This is not a language" (*CONFIGURE LANG 12, and replace the battery pack if it recurs); "ACORN" in capitals meaning MOS 3.50; DFS missing (*CONFIGURE FILE 9, *INSERT 9); no speaker sound (speaker plug, purple −5 V lead); garbled time from a flat battery.
- ↑ Acorn Computers, Special Instructions for Replacing Batteries on the BBC Master AIV Computer, part no. 0461,004, issue 1, October 1986, hosted as File:Acorn BBC Master AIV Battery Replacement 0461,004.pdf, p. 2: hold R at power-on until "CMOS RAM reset / press Break to continue" appears.
- ↑ Acorn Computers, Master 128 main PCB circuit diagram, drawing 0143,000/C, sheet 2, in File:Acorn BBC Master Series Technical Drawings.pdf (PDF page 2): IC49 (µPD7002) with C36 33 nF across CI pins 4 and 6, and C49 1 µF tantalum on the VREF line with D17 to D19.
- ↑ Acorn Computers, Master Series Microcomputer Service Manual, 0443,004, 7.9 to 7.17, pp. 46–49: printer, user port, 1 MHz bus, Tube, internal co-processor, sound, modem, cartridge port and Econet.
- ↑ stardot.org.uk, "BBC Master 128 Astec PSU - Cap Advice" (June 2022, retrieved 2026-10-02). C1 10 nF X2 and C2 100 nF X2 on the common Astec supply. Secondary source.
- ↑ stardot.org.uk, "Master 128 PSU schematic" (2017, retrieved 2026-10-02). Slow start cured by replacing R2 and R3 (180 kΩ), one of which had drifted to 275 kΩ. Secondary source.
- ↑ stardot.org.uk, "BBC Master Power Supply Questions" (December 2020, retrieved 2026-10-02), and "Master PSU circuit diagram" (October 2021): R2 found open circuit in each, the supply dead and a charge left on one primary capacitor. Secondary source.
- ↑ stardot.org.uk, "BBC Master Power Supply Ticking" (September 2024 – January 2025, retrieved 2026-10-02). 1024MAK: start-up capacitor first, then the secondary capacitors; cmorley: ticking can be the crowbar repeatedly firing; Ukwebb: a supply that holds 5 V unloaded but cycles on load usually needs its output capacitors. Secondary source.
- ↑ stardot.org.uk, "Master PSU variants" (July–August 2023, retrieved 2026-10-02). 1024MAK: a +5 V output that falls with load points to the feedback circuit, the supply to the switching transistor or the transistor itself, assuming the electrolytics and especially the start-up capacitor are good; only +5 V is regulated; a failed feedback loop fires the zener and SCR crowbar on +5 V. Secondary source.