TRS-80 Model 6000 Troubleshooting Guide

Fault finding on the Tandy 6000 follows Tandy's 1985 service manual. The boot ROM on the Z80 I/O processor runs its own tests at switch-on and reports failures as two-letter codes; the 68000 CPU and memory boards are checked with Tandy's diagnostics after the Model II diagnostics have cleared the I/O side; and the 140 W power supply has its own bench procedure.[1] Tandy's procedure isolates faults to field-replaceable parts such as the power supply, the I/O processor board and socketed ICs.[1]
The power supply runs from the mains and stores a high DC voltage on its input capacitors, and the CRT holds a charge after switch-off. Tandy tells technicians to bench test the supply through an isolation transformer and to run the oscilloscope from one as well. See the CRT Discharge Procedure. |
Setting up
[edit | edit source]Tandy's set-up for fault finding: unplug the unit, remove the top cover (see TRS-80 Model 6000 General Maintenance), check that the monitor is still connected to the video/keyboard board and the power supply, take any disks out of the drives, check that the power switch under the left lip of the base is off, and then reconnect the mains.[1]
Start-up sequence
[edit | edit source]At switch-on the Z80 runs the boot ROM, whose tests run in order, each only if the previous one passed:[1]
- Interrupts are disabled and all RAM except page 0 is deselected.
- The CRT controller is set up and the whole screen is turned white.
- The ROM checksum is checked ("BOOT ERROR CK" on failure).
- The Z80 registers are checked by passing a bit pattern between them ("BOOT ERROR Z8").
- Lower RAM, from 1000H to 7000H, is checked.
- The keyboard buffer is flushed.
- The hard disk (if fitted) and floppy controllers are reset; the heads seek to track 5 and then restore to track 0.
If the machine is to boot from floppy, "INSERT DISKETTE" appears in the middle of the white screen. When a disk is inserted and the latch closed, the screen blanks with a blinking cursor in the middle while the system loads. The boot program read from the disk then tests the DMA, PIO and the memory from 2000H to FFFFH, and shows "32K MEMORY" or "64K MEMORY" when it passes, followed by "LOADING" or "INITIALIZING" at the top left.[1]
Boot error codes
[edit | edit source]| Code | Meaning |
|---|---|
| DC | Controller or drive error: busy not reset, not at track 0, or seek error |
| DO | Drive not ready |
| TK | Record not found reading track 0 (no readable formatted data) |
| LD | Lost data on read |
| SC | CRC error: data read incorrectly |
| RS | Not a Radio Shack format disk ("BOOT" and "DIAG" not found) |
| Code | Meaning |
|---|---|
| HT | Timeout waiting for ready |
| HC | CRC error in data |
| HI | CRC error in ID |
| HN | ID not found |
| HA | Aborted command |
| HO | Track 0 error |
| HM | Data address mark not found |
| HD | Any other error |
| Code | Number | Meaning |
|---|---|---|
| CK | Bad ROM checksum | |
| Z8 | CPU failure | |
| MF | RAM failure | |
| ML | 01 | Low RAM failure (0-3FF) |
| MH | 02 | High RAM failure |
| DM | 03, 13, 23 | DMA data failure; DMA gave no interrupt; DMA gave an extra interrupt |
| PI | 04, 14, 24 | PIO data failure; PIO gave no interrupt; PIO gave an extra interrupt |
| CT | 05, 15, 25, 45 | CTC gave no interrupt on channel 0, 1 or 2; CTC gave an extra interrupt |
Source: [1] The boot program's memory test also reports "BOOT ERROR LM" for lower memory and "BOOT ERROR HM" for higher memory.[1]
Hangs at "LOADING"
[edit | edit source]TRSDOS 4.0 and later need 16 KB of memory above the first 64 KB, from Z80 memory or the 68000's memory. The boot diagnostics do not test it, so if it is missing or faulty the machine stops with "LOADING" at the top left of the screen and no error message.[1]
I/O processor tests
[edit | edit source]| Program | What it tests |
|---|---|
| SYSTEM (29 November 1982) | Serial channels A and B, PIO (printer and external floppy interfaces), DMA, Z80 interrupts and bus; needs loop-back test cables |
| MEMII (version 1.3) | Video refresh RAM, shadow ROM checksum, checkerboard, modified address, coincidence and bank select |
| DISKDG8 | Thinline floppy drives |
| DOVIDAL | Video alignment and centring |
Source: [1]
CPU and memory boards
[edit | edit source]To isolate a fault on the 68000 side, Tandy's order is:
- Run all current Model II diagnostics. If any fail, refer to the Model II Technical Reference Manual; see the TRS-80 Model II Troubleshooting Guide.
- If the 68000 CPU or memory board is still suspected, remove both and run the Model II diagnostics again.
- Once the fault is placed on the 68000 CPU board or the 512K/1 MB memory board, use the charts below.[1]
| Result | Possible fault | Action |
|---|---|---|
| Test locks up | Z80A to 68000 interface | Check interface controller U36 and its select logic |
| Test locks up | Bus arbitration controller U6 | Check latch U7 and the bus arbitrator |
| Test locks up | Refresh logic | Check all the refresh circuit logic |
| Bad RAM reported | RAM | Work out which chips from the report and replace them |
| Bad RAM reported | Z80A to MC68000 interface | Check the interface circuit |
| Bad RAM reported | Memory board | See the memory board chart below |
| No errors | The chart refers to the memory board section; the Model 16 manual sends the technician to the CPU board charts at this point |
| Test | Result | Possible fault | Action |
|---|---|---|---|
| Interrupt | Locks up on a software interrupt | Interrupt circuit; MC68000; clock logic | Check interrupt controller U15 and U34, U40, U26; replace the MC68000; check the clock circuit |
| Interrupt | Locks up on a hardware interrupt | Interrupt circuit; MC68000; clock logic | Check U15 and every hardware interrupt input to it; replace the MC68000; check the clock circuit |
| Interrupt | Software interrupt received was not the one generated | Interrupt controller | Replace U15 |
| Interrupt | Hardware interrupt received was not the one generated | Interrupt circuit | Check the interrupt input lines for shorts |
| Offset/limit | OBERROR in the user range | Offset/limit registers | Check the output address lines for shorts |
| Offset/limit | Protected memory modified | U15, U41 | Check those ICs |
| Offset/limit | OBERROR not generated | 4-bit full adders | Check U23, U24, U38 |
| Offset/limit | Random errors | I/O decoding and strobes | Check the I/O and strobe circuit |
| Result | Possible fault | Action |
|---|---|---|
| Bad RAM reported | RAM | Work out which chips and replace them |
| Bad RAM reported | Data buffers | Check buffers U5 to U8 |
| Bad RAM reported | Address buffers | Check U3, U4, and U9 for A17 |
| Bad RAM reported | Memory control | Check U9 and the signal control logic and multiplexers |
Source: Tandy's charts, pp. 52-54.[1] The CPU board's clocks come from a 48 MHz crystal oscillator, divided down to 24, 12, 8 and 4 MHz; the charts' "check clock circuit" means this logic.[3]
Power supply
[edit | edit source]The supply is an Astec AA11082 (Tandy 8790047), 140 W. Bench test it only with at least the minimum load on every output; Tandy warns that running it unloaded can cause damaging oscillation. Load values and output pins are on the TRS-80 Model 6000 General Maintenance page.[2]
Start-up test
[edit | edit source]- Inspect the board for broken, burnt or damaged parts, and check the fuse with an ohmmeter if in doubt.
- With the minimum load connected, bring the mains up slowly on a variac while watching the +5 V output on an oscilloscope and a meter. The supply should start at about 40-60 V AC and regulate by 95 V AC.
- If +5 V appears, run the performance test below. If not, follow the no-output steps.[2]
No output
[edit | edit source]- If the fuse has blown, replace it but do not power up again until the cause is found.
- Check the diode bridge DB1, power transistor Q7 and its catch diode for shorted junctions, and replace any that are shorted.
- With an ohmmeter, check each output for a short. If +5 V is shorted, check the crowbar SCR (SCR1) and zener Z1 as well.
- Check B+: with the X10 probe's ground on the cathode of D5, turn the input up slowly and measure at the cathode of D9. At 95 V AC input it should be 130-170 V DC. If not, check DB1 and the fuse F1, then R5, D12, D13 and the input capacitors C5, C6, C7 and C8.
- Check that Q7 is switching, with the X10 probe on its TO-3 case (collector). If not, check Q7 for shorts, then look at its base (under the board, the centre pin nearest the board edge). A correct base waveform means the fault is on the secondary side; a wrong one means the control section, and Tandy's instruction is to return the supply to the repair depot.[2]
The manual's figures 7-16 and 7-17 show the expected collector and base waveforms.[2]
Performance test
[edit | edit source]| Test | Input | Loads on all outputs |
|---|---|---|
| 1 | 95 V AC | Maximum |
| 2 | 135 V AC | Maximum |
| 3 | 135 V AC, then varied over the full range to look for instability | Maximum |
| 4 | 135 V AC | Minimum |
| 5 | 95 V AC | Minimum |
| Output | Minimum | Maximum | No load | Ripple (max.) |
|---|---|---|---|---|
| V01, +5 V | 4.90 V | 5.10 V | 50 mV p-p | |
| V02, +12 V | 11.40 V | 12.60 V | 100 mV p-p | |
| V03, +12 V | See note | See note | 100 mV p-p | |
| V04, +24 V | 21.20 V | 26.40 V | 30.0 V | 250 mV p-p |
| V05, −12 V | −11.40 V | −12.60 V | 50 mV p-p |
V03 must stay within 100 mV of its starting value when the +24 V load steps from 0 to 2.5 A (decaying to 0.92 A within 350 ms) or from 0.55 A to 3.05 A (decaying to 1.45 A), and within 5% under all other conditions.[2] The same switching-supply approach, for the Model II's AA11080, is on the TRS-80 Model II Troubleshooting Guide.
Floppy drives
[edit | edit source]- Of the boot errors, DC and DO come from the drive status, and TK, LD and SC from reading track 0.[1] Tandon's manual for the TM848 drives lists the causes of soft errors as airborne contaminants, electrical noise, excessive temperature and other external causes, and its first checks are grounding between supply, drive and host, freedom from conducted and radiated noise, temperature, and repeating the fault on a second diskette.[4]
- Tandon also says to check the supply voltages are within specification before working through its troubleshooting table.[4]
- If drives have been changed, or the controller has drifted, the data separator and pulse width adjustments on the I/O processor board may need resetting; see the TRS-80 Model 6000 General Maintenance page.[5]
Related pages
[edit | edit source]- TRS-80 Model 6000
- TRS-80 Model 6000 General Maintenance
- TRS-80 Model 6000 Capacitor Replacement Guide
- TRS-80 Model 16 Troubleshooting Guide
- TRS-80 Model II Troubleshooting Guide
- Tandy 6000 Service Manual
References
[edit | edit source]- ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 1.11 Tandy Computer Products, Service Manual, Tandy 6000/6000-HD, catalogue numbers 26-6021/2, 1985, hosted as File:Tandy 6000 Service Manual (26-6021, 26-6022, 1985).pdf. Section 6, "Troubleshooting Procedures", pp. 47-54.
- ↑ 2.0 2.1 2.2 2.3 2.4 2.5 Service Manual, Tandy 6000/6000-HD (1985), Section 7.6, "Power Supply Board", pp. 193-200.
- ↑ Service Manual, Tandy 6000/6000-HD (1985), Section 7.2, CPU board, "Clock Logic".
- ↑ 4.0 4.1 Service Manual, Tandy 6000/6000-HD (1985), Appendix B, Tandon TM848-1E and TM848-2E Thinline Flexible Disk Drives OEM Operating and Service Manual (1983), section 6, "Troubleshooting Guide and Replacement Procedures".
- ↑ Service Manual, Tandy 6000/6000-HD (1985), Section 4, "Adjustments", pp. 19-21.