Amiga 500 Keyboard Repair

An Amiga 500 seen from above. The keyboard is a separate assembly inside the case, with its own controller, and plugs into the main board at CN13.

The Amiga 500 keyboard is a separate assembly (Commodore part 312502-01) with its own processor, which scans the key matrix and sends key codes to CIA U7 on the main board over the 8-pin connector CN13. Commodore's service manual covers the main-board side and the keyboard's self-test codes; Sams Computerfacts covers the keyboard's own circuit, with a 6570-036 controller on the unit it examined. Three kinds of keyboard were fitted: the common Mitsumi type with a single membrane and two graphite contacts per key, a Samsung type with a multi-layer membrane, and a rare type with mechanical switches on a circuit board, known for its two Commodore keys in place of the Amiga keys. Identify yours before ordering parts.

Symptoms

• The Caps Lock light flashes after power-up. One flash at switch-on is normal. Commodore's flash codes: one flash, the keyboard processor's ROM; two flashes, its RAM; three flashes, the watchdog timer (a 74LS123 or associated circuitry). With a flash code, the fault is normally on the keyboard itself. The Amiga Hardware Reference Manual adds four blinks for a short between two row lines or the special control keys.
• Perry's Amiga Troubleshooter chart puts one or two flashes down to the 6570-036 keyboard controller, and three to the 555, 74LS123 or 74LS27 on the keyboard.
• Keyboard dead with no flash code: the fault is more likely on the main board. Perry lists CIA U7 (keyboard data on pin 39, clock on pin 40), Agnus and Gary (U5).
• Ctrl and both Amiga keys do not reset the machine: Perry's most likely cause is Q2 on the keyboard, then the 74LS27.
• Some keys dead or needing hard presses: worn or broken membrane traces, or dirty contacts. Sams says to check the keyboard's connector CN1 and the key contacts.

Tools and parts

• A multimeter, a logic probe or oscilloscope, and a separate 5 V bench supply for testing the keyboard on its own.
• A Phillips screwdriver; the keyboard's base is held by many small screws.
• For a Mitsumi keyboard, a replacement membrane of the right type: there are several A500 membranes, and Amedeo Valoroso's original was a Mitsumi A-500(E)56 A619B.
• For a Samsung keyboard, conductive silver paint, fine wire and UV-curing resin, as used by hackup.net.
• Commodore's A500/A2000 System Test disk, which includes KEYTEST, a full keyboard test.

Procedure

Remove the keyboard

1. Remove the six T10 Torx screws along the front and rear edges of the bottom case and lift off the top case.
2. Disconnect the keyboard connector CN13 from the main board and the keyboard's ground lead from the disk drive, and lift the keyboard out.

Check the main-board side

Commodore's keyboard connector CN13, numbered from the front of the board: pin 1 KBDATA, 2 clock, 3 reset, 4 +5 V, 5 key, 6 ground, 7 status (the power LED), 8 in use (the drive LED). With the keyboard connected, Commodore checks:

Commodore's keyboard circuit checks (Service Manual A500, section 7)

Step	Measure	Expect	If wrong
1	CN13 pins 2 (KBDATA) and 1 (KBCLOCK)	+5 V with negative pulses when a key is pressed	U7, RP501, CN13
2	CN13 pin 3 (KBRESET)	0 V rising to +5 V shortly after power-up	The reset circuit (U42, Q711)
3	CN13 pin 7 (STATUS)	0 V rising to +5 V when the power LED comes on	Measure U7 pin 3 (LED): good there means Q502; bad means U7, Q502, U38
4	CN13 pin 8 (INUSE)	0 V, toggling high each time a disk is inserted	Measure U36 pin 6 (MTR0): good there means Q503; bad means U36, Q503

Sams adds that if the keyboard works on the bench but not in the machine, check for pulses at pin 7 of CIA U7 (missing: the address decoder U37), then the joints on CN13 and U7 itself.

Test the keyboard on the bench

The keyboard matrix and controller on Sams' schematic: the 6570-036 (U1) with its 3.0 MHz crystal Y1 and capacitors C12 and C13, the matrix lines to the keyboard connector CN1, and the Caps Lock LED driver Q2. The pencilled pin numbers are a previous owner's.

Sams' sequence, with the keyboard unplugged from the computer:

1. Connect a separate 5 V supply to the keyboard lead: +5 V to pin 4 and ground to pin 6 of the CN13 plug.
2. Check for pulses at pins 37 and 39 of the keyboard controller (U1). If they are present, the keyboard is working; go back to the main-board checks.
3. If they are missing, check for a 3.0 MHz waveform at pin 11 of U1. If it is missing or off frequency, check the crystal Y1, C12, C13 and U1.
4. If the clock is present, hold Ctrl and both Amiga keys and check pin 3 of the reset timer (U3): it should be high while the keys are held and go low when they are released. If not, check U2, the reset switch transistor Q1 and the parts round U3.
5. If U3 is correct, check pins 8 and 12 of U2 while the keys are held: both should be low while held and go high on release. A wrong reading at pin 8 means U2. If pin 12 is wrong and pin 13 reads low, U2; if pin 13 is not low, check the parts round U4, and U1.
6. If only some keys fail, check the connector CN1 on the keyboard and the key contacts.

Sams' designators (U1-U4, Q1, Q2) are those of the keyboard it examined.

Replace a Mitsumi membrane

Amedeo Valoroso's procedure on his A500:

1. Remove the screws from the back of the keyboard and separate the base from the key assembly. Keep the small parts in labelled bags.
2. Clean the parts; Valoroso varnished the rusty areas of the metal base.
3. Compare the new membrane with the old one: the key positions and traces should match, and Valoroso checked some traces for continuity before fitting it.
4. Lay the new membrane on the back of the keyboard in the correct orientation.
5. Refit all the back screws. Do not overtighten them, and tighten them in a scattered order across the keyboard.
6. Refit the keyboard's controller board. Open the latch of its membrane connector, insert the membrane tail, and close the latch by pushing it towards the connector.
7. Refit the clear plastic sheet behind the board and a cable tie round the keyboard lead.

Repair a Samsung membrane trace

The Samsung membrane's traces run between its layers, and its flex tails cannot be pulled from their sockets without damaging the contacts. On hackup.net's keyboard three keys at the end of one trace had stopped working; the repair was:

1. Remove the many small screws from the bottom of the keyboard and lift away the bottom half with the frame, membranes and circuit board. Set aside the top half carefully: the small springs sit loose on the plungers.
2. Hold the frame upright and gently spread the two foils of the top membrane with a third-hand tool.
3. Find the break by scraping the lacquer off the trace at a few points with a craft knife and checking continuity each time.
4. Expose about 20 mm of trace round the break.
5. Apply conductive silver ink over the exposed trace and lay two or three single strands of 30 AWG wire in it as reinforcement.
6. Let the ink dry; hot air at 100 °C speeds it up.
7. Cover the ink and wires with a thin layer of UV-curing resin and cure it.

Refit the keyboard

Reconnect CN13 and the ground lead to the drive, lay the keyboard in the bottom case and refit the top case with its six T10 screws.

Testing

Run KEYTEST from Commodore's System Test disk, or a keyboard test such as the one in AmigaTestKit, which Valoroso used, and press every key. Check that Ctrl and both Amiga keys reset the machine, and that Caps Lock gives only its single flash at power-up.

Related pages

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