Jump to content

TRS-80 Model I Troubleshooting Guide

From RetroTechCollection
Radio Shack TRS-80 Model I rear panel
Rear of the TRS-80 Model I keyboard unit: the three DIN sockets J3 (tape), J2 (video) and J1 (power) at the left, and the expansion connector under its flap at the right (Binarysequence, CC BY-SA 4.0).

Fault-finding on the keyboard unit of the TRS-80 Model I. Radio Shack's TRS-80 Micro Computer Technical Reference Handbook has a troubleshooting chapter for the logic board (pp. 57-81), and the procedures below follow it. Designators (Z for ICs, C, R, Q, CR) are the handbook's.[1] Opening the case, cleaning and the supply adjustments are on TRS-80 Model I General Maintenance; the electrolytic capacitors are on TRS-80 Model I Capacitor Replacement Guide.

Before you start

[edit | edit source]
  • Unplug the Expansion Interface and anything else on the expansion connector, so that a fault in the add-ons cannot be mistaken for a fault in the computer.[2]
  • Plug the power lead into J1, the socket next to the power switch, and not into the cassette socket J3. All three sockets are 5-pin DIN types.[3]
  • Measure the rails first, with the meter's common lead on the right-hand side of C9: +12 V at the top end of R18 (11.4-12.6 V), +5 V at the left-hand end of R4 (4.75-5.25 V) and −5 V at the anode of CR2 (−5 V ± 5 %). The procedure and adjustments are on the maintenance page.[3]
  • Learn the signal states the handbook uses. A TTL low is 0.8 V or less and a high is 2.4 V or more; anything between is "floating", which on a scope shows as noise averaging about 1.5 V. Touching a floating pin with a finger makes the noise grow sharply.[4]

Garbage on the screen at power-on

[edit | edit source]
TRS-80 Model I section isolation flowchart (Technical Reference Handbook Figure 16)
Radio Shack's section isolation flowchart, Figure 16 of the handbook (p. 63).

A screen filled with random characters or graphics at switch-on means the CPU has skipped its start-up routine. The cause can be in the RAM, the ROM, the video divider chain or the CPU itself, so Radio Shack isolates it by removing parts:[5]

  1. Take the unit apart (see TRS-80 Model I General Maintenance), connect the video and power DIN plugs and switch on.
  2. If the garbage has gone, suspect a short on the interconnect cable between the keyboard and logic boards, or a loose solder ball that has fallen out during disassembly. Check the cable for shorting conductors and test again.
  3. If it is still there, switch off, wait about ten seconds for the start-up logic to reset, and switch on again. If READY now appears, suspect the power-up reset circuit around reset switch S2 and capacitor C42.
  4. If there is still garbage, switch off, remove DIP shunt Z71 and switch on. This takes the RAMs out of the system. A screen of large colons now points to a RAM fault.
  5. If there are no large colons, check whether the keyboard is still connected. With the keyboard disconnected, a flickering display points to a data-line fault. With it connected, a steady display points to a keyboard fault.
  6. Otherwise, switch off, remove the ROMs and switch on. With no program to run, the CPU fills the screen with @9 characters in the 64-character format, flickering as it accesses the video RAM. A full screen of @9 points to the ROMs; a partial screen points to the video divider chain or video RAM; no @9 at all points to the CPU or something stopping it.

The handbook warns that the method gives a starting point and not a verdict: a unit can have a ROM fault and still show large colons, or show @9 and still have a CPU fault.[4]

Reading the @9 screen

[edit | edit source]

With the BASIC ROM pulled, the @9 screen also shows video RAM addressing faults. If whole horizontal rows of @9 are missing, check the row address inputs V6-V9. If vertical columns are missing, check the column address inputs V0-V5.[6] The trs-80.com repair notes, which produce the same @9 screen by pulling DIP shunt Z3, add that a repeating pattern of @9 and garbage identifies a stuck address line: four @9 alternating with four garbage characters means A2 (22 = 4), and a full line of 64 @9 alternating with a line of garbage means A6.[2]

Symptom table

[edit | edit source]
Model I faults and where to look
Symptom Probable cause Action
Screen fills with garbage that changes to different garbage as you type One of the seven video RAMs Test or swap the video RAMs (Z45-Z48, Z61-Z63).[2]
Some characters wrong, the same bit always wrong A video RAM with a stuck bit Work out which bit differs between the character expected and the one shown. The trs-80.com notes give Z46 for bit 2 and Z45 for bit 3.[2]
Dot rows missing on every character Line-count data to character generator Z29, or Z29 itself Check the line-count inputs to Z29; if they are good, replace Z29.[6]
Dots missing in vertical columns Shift register Z10 Check the dot inputs to Z10; if they are good, replace Z10.[6]
Unstable or flashing dots Z10 missing data when hot If freezer spray on Z10 changes the fault, replace Z10.[6]
Parts of graphics cells missing Graphics generator Z8 or a broken track round it Missing vertical parts of a cell usually point to Z8. A cell with vertical streaks points to shift register Z11.[6]
Unstable or flashing graphics Z11 Replace Z11.[6]
No blanking between character lines L8 from the divider chain, or latch Z27 Check the divider chain and Z27.[6]
Severe interference on the display Oscillation on the +5 V bus Look at the +5 V bus on a scope; if it oscillates, check regulator compensation capacitors C12 and C13.[6][7]
Words spelt wrongly on screen Bad data in video RAM, or a short near latch Z28 corrupting data to Z29 Check the video RAM and the area round Z28.[6]
Dim display Z41 failing; it runs hot Replace Z41 if it is hot.[6]
Display slowly dims and fades away Cassette relay K1 or its diode CR3 shorted Replace the relay or diode; see the cassette section.[8]
Vertical roll or horizontal tearing Video divider chain Check for 60 Hz at Z32 pin 11, then work up the chain (table below).[9]
Picture cannot be centred Trimmers R20 or R21 damaged by heat; sync generator Check the timing references at Z6 and Z57, and check C21 and C27 for shorts. Readers of trs-80.com also name Z5 alongside Z6 and Z57.[6][2]
Garbage at power-on clears after a second power cycle Power-up reset: reset switch S2 or capacitor C42 Check S2 and C42.[5]
?SN ERROR after answering MEM SIZE, or nothing can be typed at READY Memory fault Fit a known good set of RAMs.[2]
Part of a CLOADed program correct and part garbage RAM bit error Run a memory test, or swap RAMs until the bad one is found.[8]
32-character display will not clear Latch Z59 or its control from D3 See the cassette section below.[8]
Recorder motor will not stop, or will not run Reed relay K1 worn out Replace K1, matching its index mark to the silkscreen.[8]
A character appears by itself after the ">" prompt Shorted key or short on the keyboard PCB Find and clear the short.[10]
Keyboard completely dead No power to the keyboard, broken interconnect cable, or no KYBD* from the address decoder Check the cable, then KYBD* activity.[10][11]
+12 V and +5 V both missing +12 V supply in fold-back because of a short Measure across R18: about 0.6 V means fold-back. Find the short on the +12 V bus.[12]
−5 V missing Short on the −5 V bus, or CR2 fitted backwards Check for negative voltage on the adapter side of R19.[12]
+12 V bus dead A loose transistor heat sink turned against the transistor's leads Check and retighten the heat sink hardware.[12]

Power supply faults

[edit | edit source]

The regulators current-limit with fold-back, so a short rarely damages the supply itself; it shuts the rail down instead. The short does not have to be in the power section.[12]

  • Warning: do not test RAMs for heat with a finger. Radio Shack points out that a supply short inside a RAM can heat it to soldering temperature.[12]
  • +12 V and +5 V both missing: measure the voltage across R18, the +12 V current-sense resistor. About 0.6 V means the +12 V supply is in fold-back and has shut itself off. The +5 V regulator is referenced to the +12 V output, so it is off too. Find and clear the short on the +12 V bus before anything else will work.[12]
  • −5 V missing: check for ample negative voltage on the adapter side of R19. If R19 is dropping all of it, the −5 V bus is shorted. Check that zener CR2 has not been fitted backwards.[12]
  • The +12 V and −5 V supplies are used by the system RAM, so a fault on either rail suggests a shorted RAM. Pull all the RAMs and retest. If the rails recover, refit the RAMs one at a time, switching on after each, until the supply crashes. There may be more than one shorted RAM.[12]
  • A short on the +5 V bus has to be found by cutting tracks to isolate sections. Repair every cut with 22 gauge or heavier wire, because the supply tracks carry a lot of current, and never bridge a cut with solder alone.[12]
  • A dead +12 V bus can also be a loose transistor heat sink. The handbook (which names Q1 at this point) describes the hardware holding the transistor and heat sink together loosening until the heat sink turns and shorts against the transistor's base or emitter lead, or shears the leads from the board. Retighten any loose heat sink.[12]
  • An AC adapter with an open primary fuse or broken leads at the DIN plug is replaced, not repaired. If its fuse has blown, check bridge rectifier CR8 on the logic board for shorted diodes before connecting the new adapter.[12]
Regulator pin voltages in a working unit (handbook Figure 19, p. 75)
Pin Z1 (+5 V regulator) Z2 (+12 V regulator)
1 0.00 0.00
2 5.30 10.60
3 5.00 11.99
4 5.00 6.92
5 5.00 6.92
6 7.46 6.92
7 0.00 0.00
8 0.00 0.00
9 0.33 5.72
10 5.89 12.31
11 11.99 21.16
12 11.99 21.69
13 7.05 13.48
14 0.00 0.00

The figures were taken with a digital voltmeter referenced to the right-hand side of C9, with the supplies set to 12.00 V and 5.00 V.[12] Z1 and Z2 are both 723 regulators; Q6 is the +12 V pass transistor, and Q3 drives the +5 V pass transistor Q4.[7]

CPU, ROM and RAM

[edit | edit source]
  • For the CPU, check for activity on the address and data lines. With none on the address lines, go to the CPU's support circuits: the clock, reset switch S2 and the power-up clear logic. To disable the CPU for short checks, ground pin 25 of the Z80 (Z40) or the end of R58: the address lines float on the CPU side and the address buffer goes tri-state. Swapping in a known good Z80 is a quick check, but the handbook warns that the fault is usually elsewhere.[4]
  • ROM faults are addressing, data or chip-select faults. Early boards may have jumper wires on the solder side that have come loose, and vibration can lift a ROM partly out of its socket. Removing DIP shunt Z3 tri-states the ROMs so that their data pins can be checked for a floating state. The chip selects ROM* and MEM* come from DIP shunt Z71, whose shorting bars can crack.[4]
  • For the RAM, check that all three supplies reach every RAM and that every RAM is fully in its socket. If the unit runs but will not load a test program, fit a known good set of eight RAMs; if that cures it, put the original RAMs back one at a time, switching on after each, until the faulty one shows itself. RAM faults are most likely after a 4K-to-16K conversion: check that DIP shunt Z3 is programmed for the RAM fitted.[13]
  • In the address decoder, check Z73 pin 6 first. It enables Z21, which decodes A12 to A14; if Z21 is never enabled the decoder is dead. Both A15 and RAS* must be low for Z73's output to go low. Z21 has open-collector outputs that rely on pull-ups through DIP shunt Z3, so they show nothing with Z3 removed. Pressing the reset switch makes the CPU busy enough that every decoder output should show at least a few pulses.[11]

The Model I/III diagnostic ROM by Dave Giller, Frank IZ8DWF and Adrian Black tests the video RAM and up to 48 KB of system RAM without needing any working RAM, and reports through the cassette port as well as on screen. Program it into a 2716 and fit it in the left-most ROM socket on the logic board, the one the Level II ROM ribbon cable plugs into. Each bank test ends with a long middle tone (one for the first bank, two for the second, three for the third); then a long high tone means all bits good, a long low tone means all bits bad, and otherwise a short high or low tone is played for each bit from bit 7 down to bit 0. The stock Model I has only seven bits of video RAM, with bit 6 faked from bits 5 and 7, and the ROM recognises this and does not report it as a fault. Clip a test lead to the cassette port or use the cassette lead to hear the tones; the keyboard is not needed.[14]

Video divider chain

[edit | edit source]

The divider chain produces every video timing signal, and loss of the vertical or horizontal reference usually traces back to a failed counter or reset gate. Because the master oscillator is in this section, a completely dead system can also end up here. The main test point is Z32 pin 11, which should read 60 Hz on a frequency counter; if it does not, work up the chain towards the oscillator until a correct signal is found.[9]

Divider chain frequencies (handbook Figure 18, p. 69)
Signal Source Frequency Comment in the handbook
VDRV Z32 pin 11 60.00 Hz Vertical reference
R8 Z32 pin 8 60.00 Hz Character row
R4 Z32 pin 9 180.0 Hz Character row
R2 Z32 pin 12 360.0 Hz Character row
R1 Z65 pin 12 660.0 Hz Character row
L8 Z12 pin 11 1.319 kHz Character/scan line
L4 Z12 pin 8 2.639 kHz Character/scan line
L2 Z12 pin 9 3.959 kHz Character/scan line
L1 Z12 pin 12 7.917 kHz Character/scan line
HDRV Z50 pin 11 15.835 kHz Horizontal reference
C32 Z50 pin 8 31.670 kHz Character column
C16 Z50 pin 9 63.340 kHz Character column
C8 Z50 pin 12 110.840 kHz Character column
C4 Z65 pin 8 221.69 kHz Character column
C2 Z65 pin 9 443.38 kHz Character column
C1 Z43 pin 7 886.756 kHz Shown for the 64-character format; held low in 32-character format
Chain input Z43 pin 9 886.756 kHz Divider chain input

The figures assume a master clock of 10.641099 MHz; production units ran about 0.03 % slow, so small deviations are normal and only large errors point to a failed part. A missing signal does not prove that the chip producing it has failed. If everything after HDRV is missing but C32 is good, Z50, Z66, Z49 and Z6 are all suspects: cut the track from Z50 pin 11 and retest. If the pin is still dead, replace Z50; if it is now active, repair the cut, cut the track at Z66 pin 5, and carry on until the shorted track is found.[9] The chain's counters are 74LS93s, whose outputs often show a step of about half a volt while high. That is normal as long as the lowest step stays above 3.0 V.[15]

For a blank screen, check the supplies and the master oscillator first, then look for video at the emitter of Q1 and work backwards. Check whether character generator Z29 is receiving data; activity at Z26 pin 8 shows the system is trying to output characters.[6] The sync generator chips Z6 and Z57 are CMOS. A floating CMOS input can look high or low on a scope, and a finger across a broken track can be enough to make the circuit work for a while.[6]

Cassette

[edit | edit source]

Most cassette faults are the recorder motor not being controlled. K1, a reed relay, wears out, especially in Level II machines, where it chatters during the long power-up routine. A relay stuck closed can be freed by tapping it, but replace it anyway, as it is likely to stick again. K1 fits either way round: match its index mark to the silkscreen. Fitted backwards, it shorts Z41 pin 3 to VCC, Z41 overheats quickly, and damage to its relay side will probably take out the video side of Z41 as well. A shorted relay or diode CR3 shows as a display that slowly dims and fades away.[8]

  • For a CLOAD fault, load a long program while watching Z4 pin 10. It should sit high and go low on the tape pulses, with at least two pulses in each bit time. If it sits low and goes high on the pulses, suspect a shorted CR4 or CR5, or a level detector that has not charged C39 because CR6, CR7 or C39 has failed. With good data reaching flip-flop Z24, check activity at Z24 pin 8 and Z44 pin 15. During a CLOAD, OUT* resets Z24: if it stays set, suspect the gate at Z25 pin 8.[8]
  • CSAVE faults usually point to the ROM or RAM, or to latch Z59. Save without a tape and look for the output waveform on pin 5 of J3. If it is missing, check R53 to R56, OUTSIG* at Z59 pin 9, data lines D0 and D1, and look for a solder short pulling Z59 pin 1 low.[8]
  • If the display is stuck in 32-character mode: Z59 sets the display width from D3. Short Z59 pin 1 to ground for a second or two. If the display changes to 64 characters and stays there, suspect a ROM or CPU software fault. If it changes back to 32 characters each time the short is released, OUTSIG* is faulty or noise is triggering Z59 through an open track.[8]

One trs-80.com reader's loading fault was cured by resoldering joints near Z4, an LM3900.[2]

Shorts

[edit | edit source]

The handbook lists five common shorts: solder splashes, solder balls left by the wave-soldering machine, fine hairs of solder raised by scraping a track, etch shorts where heat has made a pad or track slide, and failed components. A component short between a supply bus and ground usually sends the supply into fold-back.[15]

  • A gate output shorted to VCC reads about 5 V; a normal TTL high is about 3.7 V. Follow the track and look where it passes close to a +5 V track.[15]
  • To tell a ground short from a short to another output, watch the node on a scope while switching off. A trace that does not move is shorted to ground; one that moves up and then down may be shorted to another TTL output.[15]
  • An unresponsive OR or NOR gate can be checked by shorting one input to ground; the other input's signal should then pass through. Never short a TTL output to 5 V.[15]
  • Data-bus shorts are the hardest. Ground the TEST* signal at the expansion connector to tri-state every data and address buffer, then look for a data line that is not floating. Pull one suspect line up to 5 V through a 4.7 kΩ resistor and look for another line going high with it.[15]

Intermittent faults

[edit | edit source]

Radio Shack's answer to a unit that fails only away from the bench was to burn it in: run the memory test of its SCQATS diagnostic tape for several hours with the unit in its case, and if a single bad bit cannot be pinned down, replace the RAMs as a set. Otherwise, write a program that makes the fault happen every time. Flex and tap the board with a non-conductive rod to find cold joints and loose solder balls. Look under the RAM sockets for leads folded under the socket body, and for plated-through holes where the solder has not come all the way through; these open feedthroughs come in groups, so check the holes round any that you find.[16]

[edit | edit source]

References

[edit | edit source]
  1. ↑ Radio Shack, TRS-80 Micro Computer Technical Reference Handbook, second edition, second printing, 1982. Hosted here as File:Radio Shack TRS-80 Micro Computer Technical Reference Handbook (2nd edition, 1982).pdf (see Radio Shack TRS-80 Micro Computer Technical Reference Handbook). Page numbers on this page are the printed ones; the PDF page number is one higher.
  2. ↑ 2.0 2.1 2.2 2.3 2.4 2.5 2.6 Ira Goldklang, "TRS-80 Troubleshooting Guide and Repair Q&A", TRS-80 Revived (retrieved 2026-10-21). Part one is Ken Brookner's general troubleshooting guide; part two is a collection of readers' questions and answers about specific faults. Community source.
  3. ↑ 3.0 3.1 Technical Reference Handbook, p. 60, "Power Supply Checks And Adjustments"; p. 61, Figure 15; parts list p. 94 (J1-J3 5-pin DIN sockets).
  4. ↑ 4.0 4.1 4.2 4.3 Technical Reference Handbook, pp. 64-66: activity, steady state and floating signals; the CPU flowchart (Figure 17); disabling the CPU for short checks; ROM addressing, data and chip-select faults.
  5. ↑ 5.0 5.1 Technical Reference Handbook, pp. 62-64, "Section Isolation Flowchart" and Figure 16.
  6. ↑ 6.00 6.01 6.02 6.03 6.04 6.05 6.06 6.07 6.08 6.09 6.10 6.11 6.12 Technical Reference Handbook, pp. 69-71: "Video RAM" and "Video Processing", and the list of display faults on p. 71.
  7. ↑ 7.0 7.1 Technical Reference Handbook, pp. 52-54, "System Power Supply": AC adapter, the three supplies, C12 and C13 as the compensation capacitors of Z1 and Z2.
  8. ↑ 8.0 8.1 8.2 8.3 8.4 8.5 8.6 8.7 Technical Reference Handbook, pp. 73-74, "Cassette Problems", "Cassette Audio Processing", CLOAD and CSAVE faults, and the 32-character display latch Z59.
  9. ↑ 9.0 9.1 9.2 Technical Reference Handbook, pp. 68-69, "Video Divider Chain" and Figure 18.
  10. ↑ 10.0 10.1 Technical Reference Handbook, pp. 67-68, "Keyboard" and "Interconnect Cable".
  11. ↑ 11.0 11.1 Technical Reference Handbook, p. 72, "Address Decoder Expanded Discussion".
  12. ↑ 12.00 12.01 12.02 12.03 12.04 12.05 12.06 12.07 12.08 12.09 12.10 12.11 Technical Reference Handbook, pp. 74-75, "Power Supply" and Figure 19.
  13. ↑ Technical Reference Handbook, pp. 66-67, "RAM" and "Address Decoder".
  14. ↑ Dave Giller, Frank IZ8DWF and Adrian Black, "TRS-80 Diagnostic ROM", GitHub repository README (retrieved 2026-10-21). Features, beep codes and installation on the Model I and III. Community source.
  15. ↑ 15.0 15.1 15.2 15.3 15.4 15.5 Technical Reference Handbook, pp. 78-81: the static output gate, input shorting, types of shorts, logic shorts, 74LS93 output steps, address and data line shorts, bent pins, DIP shunts.
  16. ↑ Technical Reference Handbook, pp. 76-77, "It Doesn't Work Sometimes": burn-in with SCQATS, cold joints, solder balls, bent socket leads and open feedthroughs.