Jump to content

Tandy 102 Troubleshooting Guide

From RetroTechCollection
Tandy 102 power supply and reset circuits (Service Manual Fig. 7-1)
The 102's power section from the service manual schematic (Fig. 7-1): adaptor jack CN9, backup battery and MEMORY POWER switch SW3, the DC/DC converter around OT2, and the reset circuit with C78 at bottom left

Fault-finding on the Tandy 102 starts from Tandy's service manual, which has a checking procedure for each of twelve functions and names the parts to check at each step.[1] Faults that owners have found since are added under the symptom they produce. Part references M, T, C, D and R follow the main board drawing (Fig. 1-4) and schematic (Fig. 7-1) in the manual.[2]

Before you start

[edit | edit source]
  • Fit fresh AA cells or connect the adaptor (Radio Shack 26-3804), and check that MEMORY POWER and the power switch are both on. The 102 will not run with MEMORY POWER off.[3] The adaptor is centre negative.[4]
  • By default the computer switches itself off after ten minutes without use.[3]
  • A 102 that has been stored for years may have a flat backup cell and scrambled RAM. Leave it on the adaptor with fresh AA cells fitted for a day, then try again.[5]
  • Do a cold start: hold CTRL and PAUSE and press RESET on the back. All files are erased.[1] Switching MEMORY POWER off does not always clear a scrambled RAM; Kevin Becker's 102, which would not start after a crash, came back with a cold start.[6]

Power

[edit | edit source]
Doesn't work at all (Service Manual, p. 5-3)
Check Expected If not, check
Converter transformer OT2, pin 1 3.6 to 8 V C82, C83, battery contacts, adaptor jack
VDD +5 V D13, C84, ZD1
VEE −5 V D15, C85, ZD2
VB +5 V T27, T28
T21 Oscillating T22, T13, C81, R126, R127, R140, T20
RESET, high +2.2 to 5.3 V T10, T11, T25, T9
RESET, low 0.8 to 0.3 V T10, T11, T25, T9
CPU clock 4.9152 MHz at the X1 pin, 2.4576 MHz at CLK X2, M19
Logic Runs once the above are right Try another LCD unit; check the ICs
  • ZD1 and ZD2 are the board markings beside D13 and D15 on the manual's drawing.[2]
  • The low-battery LED lights at 4.2 V ±0.1 V. At 3.7 V the CPU is interrupted and the converter is stopped; once that has happened, the batteries need replacing before the computer will run again.[7]
  • Jeff Birt has found the flux on Model T boards turning slightly conductive over the decades, causing faults from phantom key presses to a supply that will not start and a RESET button that does nothing.[8]

Display

[edit | edit source]
LCD faults
Symptom Probable cause Action
No display at all Supply, reset or LCD drive The manual's order: VDD, VEE and VB; the RESET signal; the LCD waveform, and the LCD power supply operational amplifier if it is wrong; the interface ICs on the bus including M17 and M25; the LCD connector; the HD44102 and HD44103 drivers on the LCD board, or replace the LCD unit.[1]
No contrast, dial has no effect Contrast circuit The contrast dial VR1 feeds the base of T23, whose output goes to V2 on pin 4 of the LCD connector CN7, with R154 and R128 to VEE.[9][10] Measure VEE as well.
Garbled screen after the backup cell has leaked Corroded bus traces Jeff Birt has seen a 102 where the leak had eaten the fine traces around the battery; a garbled LCD is likely to be bus damage, and such boards can be impractical to repair. If a cold start brings the screen back, the damage may be minor.[11]
A missing horizontal row of dots LCD glass Birt traced this on a T102 to the glass. The glass is laid out symmetrically: take it out of its frame, turn it through 180 degrees, and if the missing line moves with it the glass is faulty. Lining the glass up again on reassembly is fiddly.[12]
A partly missing column One driver output A bad contact or a failed driver.[12] Look for corrosion on the traces under the conductive rubber strip between the glass and the board; Stephen Adolph repaired an LCD by cleaning and repairing traces he found there under a microscope.[13]

Will not start up properly

[edit | edit source]
Reset (Service Manual, p. 5-6)
Check Parts
The reset circuit T11, T10, T9, T25, D20, D12 and C78
ICs that use RESET M19, M25, M22, M14, M27, M36 and M31; also the LCD unit and the RAM
  • C78 (3.3 µF) and R103 delay RESET until VDD is up.[7] Fugu ME100 traced a 102 that started only after the RESET button was pressed to a cracked trace to the reset capacitor, found by checking continuity. He finds that 102 boards develop hairline cracks in their traces over time.[14]

Memory

[edit | edit source]
  • The standard RAMs are M7, M8 and M9, 8 KB each; M6 is the optional fourth. The decoder M5 (TC40H139F) selects them from A13 and A14.[15][16]
  • Josh Malone's 102 showed 21,446 bytes free after a cold start, 8 KB short. Writes to one RAM's address range never stored, and there was no activity on that chip select from the decoder; he replaced the decoder, and when the fault returned later he reflowed the joints of the RAMs and decoder.[17][18] Club 100's RAMTST.BA and RAMCHK.BA programs test the RAM.[14]
  • Memory protection: VB should be +5 V with the power on and 2.0 to 4.0 V with it off. If not, check the converter, or the Ni-Cad, D11 and D22. Then check that RAM RST and RESET switch symmetrically, and check the RAMs.[1] Geoffrey Oltmans found a 102 that lost its memory at switch-off because leakage from the old battery had etched through the long trace along the board edge that carries backup power to the RAM. Measure VB at the power pin of each RAM, not at the battery.[19][20]

Keyboard, clock and buzzer

[edit | edit source]
Keys, clock and buzzer (Service Manual, pp. 5-5 and 5-6)
Symptom Check, in order
Keys do not work Keyboard connector seated; key scan signal at M25 PA0 to PA7 and PB1, and if absent, pulses every 4 ms from the TP terminal (pin 10) of M18 and at M25; the diodes on the keyboard; the return signal; M15, M3, M26, M16 and pull-up resistors R163 to R170; the individual keys
Clock does not keep time Supply to M18: +5 V on, 2.0 to 4.0 V off; 32.768 kHz oscillation (X1, C17, C18, M18); a 250 Hz pulse from TP; the clock-setting ICs M14 and M25 and the RESET signal
No beep or melody The LCD connector and the buzzer connector on the LCD board; M26 and M25 (PB5 port and TO terminal); the buzzer

Source: the checking procedures.[1]

  • The keyboard is scanned from the clock chip's timing pulse, so a dead clock also stops the keyboard. Philip Avery's 102 showed a random date and ignored the keyboard; there was no TP output from the clock chip because its chip select, fed from RESET, was open. Tying CS to RESET elsewhere on the board cured it.[21][22]
  • Dry joints at the clock chip M18 are another cause. Fugu ME100 found them on a 102 board that had been working. On two other machines he found a cracked joint and an empty one near the clock chip, and on the Model 100 the keyboard cables press on that part of the board when the case is closed, so the fault can appear only with the case shut.[14][23] Birt reports a friend's 102 that kept bad time until it had warmed up; replacing M14 cured it.[8]
  • Where only some keys fail, look at the keyboard's flat cable first: its contacts lift easily and replacement cables are available. A single dead key can be its diode; see General Maintenance.[24][25]

Ports

[edit | edit source]
Interface faults (Service Manual, pp. 5-7 to 5-10)
Interface Checks
Printer Strobe signals, M14 and M32; data signals, M25, M10 and M32; the connector and cable
Cassette A modulated waveform at pin 5 of the cassette connector while saving (else M34, M19, C63, C64); a digital signal at the CPU's SID pin while loading (else D5, D6, M30, M34, M19); the remote relay RY1, T6 and M14; the connector and cable
Bar code reader M34, M25, M19 and L5; VDD
RS-232C A ±3.5 to ±5 V signal at pin 2 when sending, with CTS (pin 5) low (else M22, M24, M35, C71, C72, C73); a signal at M22 pin 20 when receiving, with RTS (pin 4) low (else M22, M24, M33, D8, D9, D10); PB3 of M25 (pin 32) low (else M25, M34); the connector and cable
Modem The modulated signal at connector pin 5 (coupler) or pin 7 (direct) when sending, and the carrier at M31 pin 1 (else M22, M30, T4, T7, OT1, RY2); on receive, the signal at M31 pin 1 (else M29, M30, OT1, T2, T3, T5, D1, D2, RY2, then M22 and M31); the dialler, RY2, RY3, T7, T24, M36 and M13

Source: the checking procedures.[1]

  • B4 Me100 warns that leakage from AA cells left in the battery compartment can reach the board under the battery carrier, and has had an RS-232 failure on a 102 from it.[26]
  • Two wires soldered across the printer port area are a factory modification, both grounds, and Fugu ME100 found them on every 102 he had seen.[14]
[edit | edit source]

References

[edit | edit source]
  1. ↑ 1.0 1.1 1.2 1.3 1.4 1.5 Tandy, Service Manual: Tandy 102 Portable Computer, catalogue number 26-3803, Part V, Troubleshooting, pp. 5-1 to 5-11: the troubleshooting guide flowchart (pp. 5-1 and 5-2), checking procedures for "doesn't work at all" (p. 5-3), LCD (p. 5-4), keys and buzzer (p. 5-5), clock and reset (p. 5-6), memory protection and printer (p. 5-7), cassette and bar code reader (p. 5-8), RS-232C (p. 5-9) and modem (p. 5-10), and the check list (p. 5-11). Hosted as File:Tandy 102 Service Manual (26-3803).pdf, described at Tandy 102 Service Manual.
  2. ↑ 2.0 2.1 Service Manual: Tandy 102, Fig. 1-4, "Main PCB (Bottom View)", p. 1-5, rendered as File:Tandy 102 main board (Service Manual Fig. 1-4).png; Fig. 7-1, "Main PCB - Schematic Diagram", p. 7-1, power section rendered as File:Tandy 102 power supply and reset circuits (Service Manual Fig. 7-1).png.
  3. ↑ 3.0 3.1 Service Manual: Tandy 102, Part I, pp. 1-2 to 1-4: the power switch, the ten-minute automatic power-off, the low battery indicator, the adaptor socket for AC Power Supply 26-3804, the RESET switch, and the MEMORY POWER switch, without which the 102 will not operate.
  4. ↑ Bob Pigford, "Re: 102 Keyboard", M100 mailing list, 4 February 2017 (retrieved 2026-10-21). Community source. From the T102 owner's manual: cold start by holding CTRL and PAUSE and pressing the power switch or RESET; from Club 100's notes, the adaptor is centre negative.
  5. ↑ Roma, "Re: 102 Keyboard", M100 mailing list, 4 February 2017 (retrieved 2026-10-21). Community source. A 24-hour charge on the wall adaptor with four fresh AA cells has often revived 102s that had not been used for a long time.
  6. ↑ Kevin Becker, "Re: Model 102 died", M100 mailing list, 25 March 2018 (retrieved 2026-10-21). Community source. A Model 102 that would not start was recovered with a cold reset (CTRL, BREAK and RESET); turning off the memory power had not been enough.
  7. ↑ 7.0 7.1 Service Manual: Tandy 102, Part IV, "Power Supply Circuit", p. 4-24: VDD from D13 and C84, VEE from D15 and C85, feedback through D4 to T13; the low-power detector (LED at 4.2 V ±0.1 V, LPS signal to the CPU TRAP input at 3.7 V, converter stopped by T20 through M28); the reset circuit, in which C78 and R103 delay RESET until VDD is up, T25 shorts C78 at automatic power-off, and thermistor TH2 compensates for temperature.
  8. ↑ 8.0 8.1 Jeff Birt, "Re: M100 garbled screen", M100 mailing list, 4 July 2025 (retrieved 2026-10-21). Community source. A friend's T102 kept bad time until it warmed up, cured by replacing M14; the board flux becomes slightly conductive over the decades and has caused phantom key presses, a power supply not starting and a reset button not working.
  9. ↑ Service Manual: Tandy 102, Fig. 7-1, "Main PCB - Schematic Diagram", p. 7-1: the contrast circuit (VR1 50k B, T23, R154, R128, V2 at CN7 pin 4), the clock IC M18 (µPD1990AC) with X1, C17 and C18, and the power and reset circuits.
  10. ↑ Service Manual: Tandy 102, Part VI, parts list, p. 6-7: VR1, 50k B-curve, contrast; VR2, semi-fixed, modem.
  11. ↑ Jeff Birt, "Re: Fwd: Error on laptop. Nicd or something serious?", M100 mailing list, 25 March 2022 (retrieved 2026-10-21). Community source. A leaking NiCad in the T102 eats away the fine traces in that area, which can be catastrophic and impractical to repair; try a hard reset; a garbled LCD is probably bus damage.
  12. ↑ 12.0 12.1 Jeffrey Birt, "Re: M100 / 102 Wanted", M100 mailing list, 7 July 2021 (retrieved 2026-10-21). Community source. A partial missing column is one driver output's worth, from a bad contact or driver; a missing row on his T102 was the LCD glass; rotating the symmetrical glass 180 degrees shows whether the fault moves with it.
  13. ↑ Stephen Adolph, "Re: LCD Lines", M100 mailing list, 3 March 2020 (retrieved 2026-10-21). Community source. Missing lines cured by repairing corroded traces from one driver, found under the rubber conducting gasket with a microscope.
  14. ↑ 14.0 14.1 14.2 14.3 Fugu ME100, "Re: M102 weird behavior and bodge wires", M100 mailing list, 25 September 2017 (retrieved 2026-10-21). Community source. T102 boards develop micro-cracks in traces; a cracked trace to the reset capacitor meant a T102 only started after pressing RESET; dry joints at the clock IC on a board that had been working; two factory bodge wires at the printer port, both grounds, on every T102 he had seen.
  15. ↑ Service Manual: Tandy 102, Part IV, pp. 4-3 and 4-4: memory map, the standard RAMs M9, M8 and M7 and the optional M6; address decoding and bank selection by M5 and M14.
  16. ↑ Service Manual: Tandy 102, Part VI, parts list, p. 6-3: M5 TC40H139F decoder; M7 to M9 µPD4364C or TC5565PL-15 RAM; M14 TC40H175F; M18 µPD1990AC; M19 MSM80C85ARS.
  17. ↑ Josh Malone, "Place your bets: M5 or M6?", M100 mailing list, 10 July 2016 (retrieved 2026-10-21). Community source. A 102 with 8 KB missing (21,446 bytes free), no activity on the M5 chip select, and pokes into that RAM's range not stored; decoder or SRAM suspected.
  18. ↑ Josh Malone, "M102 weird behavior and bodge wires", M100 mailing list, 25 September 2017 (retrieved 2026-10-21). Community source. A long history of memory problems on his 102, first resolved by replacing the decoder IC; later cold starts reporting less memory; reflowing all the RAM and decoder joints changed the behaviour; two bodge wires at the printer port.
  19. ↑ Geoffrey Oltmans, "Re: Unresponsive Model 102", M100 mailing list, 28 December 2015 (retrieved 2026-10-21). Community source. Battery acid had etched a hole in the long edge trace that supplies backup power to the RAM; bridged with a jumper; check RAM power at the RAM's own power pin.
  20. ↑ Geoffrey Oltmans, "Re: Unresponsive Model 102", M100 mailing list, 28 December 2015 (retrieved 2026-10-21). Community source. Check RAM power at the power pin on the RAM itself, not at the battery terminals.
  21. ↑ Philip Avery, "Re: M100 stuck on the menu with weird date", M100 mailing list, 6 January 2019 (retrieved 2026-10-21). Community source. A T102 with random date data and an unresponsive keyboard: no timing pulse from the RTC because its chip select, fed by RESET, was open; fixed by tying CS to RESET elsewhere.
  22. ↑ Stephen Adolph, "Re: Dead Keyboard", M100 mailing list, 23 April 2021 (retrieved 2026-10-21). Community source. On a T102 whose keyboard was dead though a cold start worked: pin 10 of the clock chip is the timer pulse to the RST 7.5 input, and without it the keyboard service routine does not run; a dead X1 crystal would do this.
  23. ↑ Fugu ME100, "Re: M100 stuck on the menu with weird date", M100 mailing list, 6 January 2019 (retrieved 2026-10-21). Community source. On two machines he repaired, dry joints near the clock chip, one cracked and one with no solder in the hole; pressing on the area made the board work; the keyboard cables press down on the board when the unit is closed, so an open unit may work and a closed one not.
  24. ↑ Bert Put, "M-102 keyboard flat ribbon cable", M100 mailing list, 29 July 2020 (retrieved 2026-10-21). Community source. The 102's keyboard ribbon cable tracks moved when refitting it to its edge connector.
  25. ↑ Josh Malone, "Re: 102 Keyboard repair help RQ", M100 mailing list, 17 September 2022 (retrieved 2026-10-21). Community source. Each key on the 102 matrix has a diode; check it with a meter on diode test.
  26. ↑ B4 Me100, "Re: T-102 is here!", M100 mailing list, 13 March 2020 (retrieved 2026-10-21). Community source. Check the AA battery box for signs of a leak, which can destroy the PCB under the battery carrier and lead to RS-232 failure, as on one of his T102s; check the electrolytic capacitors for leakage.


Retrieved from "https://retrotechwiki.com/index.php?title=Tandy_102_Troubleshooting_Guide&oldid=29150"