Amiga 500 Troubleshooting Guide

An Amiga 500 with an external drive and a joystick

Fault-finding on the Amiga 500, built on two sources. Commodore's Service Manual A500 (PN-314981-04, October 1990) gives the start-up sequence and a signal-by-signal troubleshooting procedure for the Rev. 5 and Rev. 6A/7 boards. Ian Perry's Amiga Troubleshooter chart (1993) maps symptoms to chips on the A500 and A2000. Commodore takes all readings with an oscilloscope, to a ±10 % tolerance.

Before anything else, remove every add-on (memory, hard drive, external drive, printer, modem) and try the machine with a known good power supply and drive. For the supply itself see the maintenance guide.

Start-up colours

Kickstart 1.3 changes the screen colour as it starts. Commodore's sequence:

Amiga 1.3 start-up sequence (Commodore)

Stage	Screen
Hardware set up, screen turned on	Dark grey
ROM checksum bad	Red
Spurious exception before the error handler is set up	Yellow
Local memory configuration (boundaries, size) fails	Green
Custom chip register check fails	Blue
Software running	Light grey, then the Workbench prompt

Wikipedia's Kickstart article says blue is used by Kickstart 3.0 and higher; Commodore's own A500 manual lists it in the 1.3 sequence.

The keyboard has its own self-test, reported on the Caps Lock light. Commodore's Amiga Hardware Reference Manual gives one blink for a ROM checksum failure, two for a failed RAM test, three for a failed watchdog timer test, and four for a short between two row lines; the keyboard sends the blinks in bursts about once a second. One flash when the machine is switched on is normal. Perry's chart blames the keyboard's 6570-036 controller for one or two flashes, and the 555, 74LS123 or 74LS27 on the keyboard for three.

Preliminary checks

Commodore's minimum conditions for any picture at all. Points marked (P) are measured as power is applied.

A500 basic preliminary checks

Step	Measure	Expect	If wrong
1	CN12 pin 1 (right pin)	+5 V	Power supply section
2	CN12 pin 4 (left pin)	+12 V	Power supply section
3	U38 (MC1488) pin 1	−12 V	Power supply section
4 (P)	U5 (Gary) pin 41, RST	0 V rising to +5 V shortly after power-up	Reset section
5 (P)	U37 (74LS32) pin 11, RESET	As step 4	Reset section
6 (P)	U37 pin 8, IORESET	As step 4	Reset section
7	U1 (68000) pin 15, CLK	7 MHz	Clock section
8	U33 (74F04) pin 4, CCK	About 3.58 MHz	Clock section
9	U33 pin 6, CCKQ	About 3.58 MHz	Clock section
10	U33 pin 10, CDAC	About 7 MHz	Clock section

If all ten are right and there is still no picture, work through the 68000, CIA and video checks below.

Power

Commodore's power supply checks follow each rail through the line filter LF1 to the supply. If a rail is missing at CN12 with the drive unplugged as well, measure LF1's output pin (pin 7 for +5 V, pin 5 for +12 V, pin 6 for −12 V) and its input pin (2, 4 and 3). Good at the input but bad at the output means a faulty filter; bad at the input means a faulty supply or a component shorting the rail. If the rail returns when the drive is unplugged, the drive is at fault.

Power symptoms

Symptom	Check	Source
Black screen, no power-up	Try a known good supply; 68000 U1 (too hot to touch means bad); Agnus U2; crystal X1 (no clocks); U3 and U33	Perry
Machine works but the serial port does not	−12 V missing; the machine runs without it	Perry
Rails good with drive unplugged, bad with it connected	Faulty drive assembly	Commodore

Reset and clocks

The power-on reset comes from the 555 timer U42 through transistor Q711 to Gary's KBRESET input (U5 pin 5). The keyboard can also pull it through CN13. With KBRESET good, check RST at U37 pin 12, then RESET (pin 11) and IORESET (pin 8) on U37; Commodore lists U5, U1, U2, U37, U7 and U8 as possible causes. Sams gives the expected behaviour at 68000 pin 18: low for about half a second after power-up, then high.

All clocks start at Agnus. Commodore's checks:

• 28 MHz at U2 (Agnus) pin 34. If missing, suspect the oscillator X1 or U2.
• 7 MHz at U2 pin 38. If missing, suspect U2, U1 or U4.
• CCK at Gary pin 28, CCKQ at pin 27 and CDAC at pin 26. Missing at Gary points at U5, U2, U3, U4 or U33; present at Gary but missing after U33 points at U33.

Sams gives the NTSC figures as 28.636 MHz at Agnus pin 34, 7.159 MHz at pins 37 and 38, and 3.5795 MHz at pins 39 and 40.

Processor and CIAs

Commodore checks the 68000's control lines in turn. HLT (pin 17) should sit at +5 V; UDS and LDS (pins 7 and 8), R/W (pin 9), AS (pin 6), DTACK (pin 10), VMA and VPA (pins 19 and 21) and E (pin 20) should all show activity. The manual notes that the 68000 may start with control signals around 2.5 V, but levels that do not reach close to +5 V indicate a fault. Most failing steps point at U1, U5 (Gary) or U2 (Agnus).

The two 8520 CIAs are U7 and U8. U7's TICK input (pin 19) should carry vertical sync and U8's horizontal sync; a missing TICK points at Agnus or the CIA. Chip selects come through U37 (74LS32) from address lines A12 and A13.

Common chip-level symptoms (Perry's chart)

Symptom	Check, in order
Green screen	Agnus U2; the A501 or other RAM expansion; Gary U5; RAM on the main board (a chip too hot to touch is bad); the U2 socket; U7 and U8
Dark grey goes to light grey but no Workbench prompt	CIA U7; Paula U3; EMI301 (1 Ω), or R309 on Rev. 6A/7 boards; Agnus U2; U37 (74LS32); the U2 socket
Screen breaks up after 15 to 30 minutes	Agnus U2; Denise U4; the U2 socket
Locks up when an icon is double-clicked	Agnus U2; Denise U4; CIA U7; U37

Agnus sits in a PLCC socket, and Perry lists the socket itself for several faults. Reseat it with a PLCC extraction tool.

Video

Denise (U4) produces digital RGBI, buffered by U40 and U41 (74HCT245). The video hybrid HY1 turns that into analogue RGB and composite video. Work back from the output:

Video checks

Fault	Commodore's check	Perry's suspects
No or wrong analogue RGB	HY1 outputs AR, AG, AB, then U40/U41 outputs, then their inputs. Good at HY1: EMI filters or connector CN9.	Denise U4; U40, U41; the video hybrid
No composite	HY1 composite output to CN10
Wrong colours (red, green or blue missing)	U40/U41 inputs and outputs	Denise U4; U40, U41; video hybrid
No vertical or horizontal sync on the RGB port	RP403 pins 5 and 3; then Agnus U2
Scrambled picture, loss of composite sync	U41 B7 output and A7 input	Agnus U2; U41; RP403
Black screen with the machine running		R405 and R406 (4.7 Ω); Denise U4; video hybrid; U40, U41

Floppy drive

Try a known good drive first, and clean the heads; see Floppy Drive Repair. Commodore's checks for the internal drive:

• Motor: MTR and SEL at Gary pins 7 and 16, MTRON at Gary pin 46, then U36 (74LS38) pin 6 and transistor Q503 to CN11.
• SIDE (pin 12), STEP (pin 10) and INDEX (pin 24) on CIA U8.
• RDY (pin 7), TRK0 (pin 6), WPROT (pin 5), CHNG (pin 4) and the LED output (pin 3) on CIA U7.
• Read data DKRD at Paula pin 37; write data and write enable through Gary pins 45 and 44 and U36.

Perry's chart gives CIA U8 for a drive that is not recognised, U8, Gary U5 or U36 for motor problems, CIA U7 for disk-change and write-protect faults, and Paula U3, the 68000 or Agnus for read and write errors.

Keyboard, mouse and joystick

• Keyboard data and clock on CN13 pins 2 and 1 go to CIA U7. If they are missing, check U7, RP501 and CN13.
• If the machine cannot be reset with Ctrl and both Amiga keys, Perry's first suspect is Q2, then the keyboard's 74LS27. On the main board he lists CIA U7, Agnus and Gary.
• Mouse and joystick direction lines pass through U15 (74LS157) to Denise; the fire buttons go to CIA U7 through EMI415 and EMI425.
• Mouse pointer does not move but buttons work: EMI401 (5.1 Ω, the 5 V feed to the mouse), Denise U4, U15. Left button and fire button dead: CIA U7. Right button dead: Paula U3.

CIA failures show as dead fire buttons, disks not recognised or key presses not registered. Hot-plugging a joystick or mouse can short a pin to the connector shell and damage a CIA.

Audio

Paula outputs left audio on pin 31 and right on pin 30. Each channel passes through a section of U14 (LF347) and a filter stage, then the output capacitor (C324 left, C334 right) and an EMI filter to the RCA jacks. Commodore expects about 200 mV of audio at the jack side of the EMI filters. Check in order: Paula output, U14 inputs and outputs, the capacitor, resistors and EMI filter. Perry adds the filter FETs Q321 and Q331, and cracked solder joints on the jacks CN3 and CN4.

Serial and parallel ports

• Serial transmit fails: U38 (1488), Paula U3, CIA U8. Receive fails: U39 (1489), Paula U3, CIA U8.
• Printer does not work: CIA U8 and the EMI filters on the port; printing wrongly: CIA U7. Commodore checks the data lines through EMI512-EMI519 back to CIA U7.

Diagnostic software

Commodore's service manual describes an A500/A2000 System Test disk. Its commands include KEYTEST (keyboard and matrix), SCREENTEST (colour bars and grey scales), SHOW BALLOON (a HAM picture that shows up a bad Agnus or Denise), CUBEROTE, and SYSTEST, which tests the clock, the blitter, chip and fast RAM, sprites and disk I/O with options for A500s with and without an A501.

Related pages

• Amiga 500
• Amiga 500 Maintenance Guide
• Amiga 500 Capacitor Replacement Guide
• Amiga 500 Service Manual (314981-04)
• Sams Computerfacts CSCS26 Amiga 500
• Amiga 2000 Troubleshooting Guide