RetroTechCollection

IBM PC XT Troubleshooting Guide

Revision as of 20:44, 30 September 2026 by Josh (talk | contribs) (Rewrite from IBM service manual and minuszerodegrees: remove invented beep codes, HelpPC code list (AT/PS2 codes), wrong SW1-2 and SW1-1 advice, inverted RESET level and C8000 beep claim; add minimum diagnostic configuration, memory code decoding, IBM connector limits)

The IBM Personal Computer XT (5160) with IBM's BIOS reports faults through beeps and on-screen messages from its power-on self test (POST). The IBM PC (5150) Troubleshooting Guide covers the 5150, which differs in its switches, board versions and supply.

IBM PC XT with a 5151 display and 83-key keyboard (Ruben de Rijcke, CC BY 3.0, via Wikimedia Commons)

The POST is a confidence test only. No 201 error does not prove the RAM good, no 301 does not prove the keyboard good, and no 601 does not prove the floppy subsystem good.[1]

Before you start

  • The XT is slow to start. IBM gives 13 to 90 seconds for a complete POST, depending on the options fitted; minuszerodegrees.net says it can take minutes before an error message appears.[2][3]
  • The POST writes no codes a POST card can read.[3]
  • After switching off, wait at least one second before switching on again, or the supply may not regenerate power good.[4]
  • Check SW1 against what is fitted. Switch 1 must be off (on makes the POST loop); switch 2 is off when an 8087 is fitted and on when it is not; 3 and 4 set the system board memory; 5 and 6 the display; 7 and 8 the number of diskette drives.[5][6]

Normal power-on sequence

IBM describes a good start as follows:[2]

  1. A cursor appears after about four seconds.
  2. The memory size appears in the top left corner and counts up in 64 KB steps while the memory is tested (the 5150 does not show this).
  3. One short beep when the POST completes.
  4. The machine boots from diskette or fixed disk, or shows the Cassette BASIC screen if there is nothing to boot.

Minimum diagnostic configuration

For a machine that appears dead, minuszerodegrees.net reduces it to the system board, power supply and speaker, with no cards, no keyboard and no drives connected, and SW1 switches 5 and 6 both off (monochrome). About five seconds after switching on, a working board gives one long and two short beeps, because the POST cannot find the monochrome adapter the switches ask for.[7] P8 and P9 can be forced on the wrong way round, so check them.[7]

If there are no beeps:

  • If the fan does not turn, or turns only for a moment, either the supply is faulty or the board has a short that shuts it down. Switch off, unplug P8, switch on again. If the fan now runs, look for a short on the board: C56 on +12 V first, then the +5 V tantalums (see IBM PC XT Capacitor Replacement Guide).[7][8]
  • If the fan keeps turning, that proves only part of the supply. Measure the rails with P8 and P9 still connected to the board.[7]
Checks at P8/P9 with the board running
Connector pin Wire Signal Expected
P9 pins 4, 5, 6 Red +5 V +4.8 to +5.25 V
P9 pin 3 White −5 V −4.6 to −5.5 V
P8 pin 3 Yellow +12 V +11.5 to +12.6 V
P8 pin 4 Blue −12 V −10.9 to −13.2 V
P8 pin 1 Orange Power good 2.4 to 5.2 V

Limits from IBM's service manual p. 4-3;[2] pin functions and wire colours from minuszerodegrees.net.[7] If all four rails are in limits but power good is below 2.4 V, the supply is faulty.[7]

IBM's dead machine checks add: disconnect all external cables, remove the 8087 if fitted, unplug the drive power connectors one at a time (a failing drive logic board can load the supply down), check the supply voltages at the board, and check the board's resistance.[2]

With the supply and voltages good and still no beeps, the clock and reset can be checked at any expansion slot. CLK on pin B20 is 4.77 MHz with a 33 % duty cycle; OSC on B30 is 14.31818 MHz with a 50 % duty cycle; RESET DRV on B2 is active high, so it should be high at power-on and low once the board is running. B1, B10 and B31 are ground.[4] minuszerodegrees.net also suggests a diagnostic ROM in socket U18 in place of the BIOS, preferring Ruud's diagnostic ROM to the Supersoft/Landmark ROM on the 5160.[1]

POST beeps and messages

IBM 5160 POST beeps and messages
Beeps or message Meaning and likely causes
Nothing: no display, no beep, no diskette access after two minutes Many causes. Rule out an 11/08/82 BIOS with a CGA card and SW1 set for MDA, then use the minimum diagnostic configuration.
Flashing cursor only, no beeps With the 11/08/82 BIOS, an MDA card with SW1 set for CGA. Also seen with a badly seated MDA card.
Flashing cursor, no RAM count, then an attempt to boot from A: An MDA card with SW1 set for EGA; with the 05/09/86 BIOS, also a CGA card with SW1 set for EGA.
1 short beep POST passed.
2 short beeps Accompanies an on-screen xx01 error and "ERROR. (RESUME = F1 KEY)".
1 long + 2 short SW1 set for MDA or CGA and the card is missing or faulty; or a board fault makes the POST think an MDA or CGA is fitted when SW1 is set for EGA/VGA; or a video card's own ROM found a fault.
1 long + 3 short The ROM on an IBM EGA card found a fault in the card.
Continuous tone Not produced by the POST. Faulty system board, a card interfering with the board, or a faulty supply; power good being present does not prove the supply.
101 8259 interrupt controller, 8253 timer or 8237 DMA controller, or their support logic.
201, e.g. "30000 10 201" RAM. See Memory errors.
301 or "xx 301" Keyboard. A preceding byte identifies a stuck key. Otherwise: no keyboard, an AT keyboard, a broken cable conductor, a bad joint on the DIN socket, or the board's keyboard circuit.
601 Diskette subsystem. See Diskette faults.
1701 Produced by the fixed disk controller's own ROM, not the system board. Covers controller, cables, drive, configuration and power.
1801 The POST found a 5161 extender card and could not talk to the expansion unit: cable, power, receiver card, extender card, or the 5161 not switched on first. An accelerator card in slot 8 can also cause it.
C8000 ROM, CA000 ROM, D0000 ROM An adapter's ROM at that address is faulty, corrupt or badly seated. C8000 is usually a fixed disk controller; CA000 is often a third-party floppy controller.
F6000 ROM 08/16/82 and 11/08/82 BIOS only: U19 missing, faulty or badly seated. The machine still boots; only Cassette BASIC is lost.
F8000, FA000 or FC000 ROM 1982 BIOS only: part of U18 is corrupt or badly seated.
PARITY CHECK 1 Parity error in system board RAM: a chip, a badly seated chip, refresh or parity logic, or a filter capacitor near the RAM.
PARITY CHECK 2 Parity error in RAM on an expansion card.
PARITY CHECK with no number Not produced by the IBM BIOS; the Parity_Boot virus shows it.

Source: minuszerodegrees.net.[1] On the 1986 BIOS, the POST checks U18 and U19 first and halts the processor without any message or beep if either is bad.[1]

Memory errors

The number before "201" locates the failing chip. IBM's service manual gives the decoding for each board:[2]

  • On the 64/256 KB board (5160, 5155 and XT/370) the code has four characters. The first character, 0 to 3, is the bank on the system board; the last two give the failing bit (00 is the parity chip, then 01, 02, 04, 08, 10, 20, 40, 80 for bits 0 to 7). IBM's example, 3040 201, is bank 3, bit 6.
  • On the 256/640 KB board the code has seven characters. The first character is the bank; the sixth and seventh give the bit. IBM's example, 3C000 40 201, is bank 3, bit 6.

IBM treats a first character of 0 to 3 as a system board failure. IBM also warns that a chip fitted the wrong way round can give a memory error, a lock-up or a blank screen.[2]

The 5160's POST sizes memory by writing to the start of each 64 KB block and stops at the first block that does not read back. A failed data chip therefore often shows as less memory counted, with no 201: minuszerodegrees.net pulled a chip from bank 3 of a 256 KB board and the machine counted 192 KB.[9] A failed chip in bank 0 leaves the board apparently dead.[9]

On a 256/640 KB board that counts only 256 KB, check jumper E-2; it must be fitted for 256 KB chips in banks 0 and 1.[2]

IBM's method for a memory fault the diagnostics do not catch: reduce the memory one bank at a time with the switches until the error stops. The last bank removed is the failing one; then restore the switches and swap chips in that bank until the error goes.[2]

Power supply faults

  • The supply shuts all outputs down on a DC overcurrent or overvoltage and stays down while the fault is there. The fan may still turn.[4][3]
  • It also shuts down if underloaded. A working system board alone is usually enough load.[3] To measure a supply on its own, IBM loads P9 pins 1 and 6 with 1.5 ohm, 25 W.[2]
  • Turning off the red switch does not remove mains voltage from everything inside the supply.[10]
  • Limits and sense levels are in the maintenance guide.

Fixed disk faults

  • For a 1701 error, reseat the 34-pin control and 20-pin data cables at both ends and the adapter in its slot, check +12 V at the drive (IBM's limits at the drive power connector are +11.5 to +12.6 V, and +4.8 to +5.2 V on +5 V), and check that the drive type is one the controller supports.[1][2][11]
  • The 10 MB cable (P/N 8529271) plugs into the adapter with the trace wires facing up, away from the system board; the 20 MB cable (P/N 6480086) with the trace wire facing down.[2]
  • The 10/20 MB adapter has a drive-type switch block for drives C and D; IBM's settings for types 1, 2, 13 and 16 are on p. 6-4 of the service manual. Type 1 is the IBM 10 MB drive.[5]
  • IBM treats a 10 MB drive as good if CHKDSK's total disk space less bytes in bad sectors is at least 10,240,000 bytes, and a 20 MB drive at 20,480,000 bytes.[2]
  • A replacement fixed disk or diskette drive needs the terminating resistor if the drive it replaces had one; the last drive on each cable must be terminated.[2]

Diskette faults

Causes of 601 listed by minuszerodegrees.net include a faulty or badly seated adapter, a dirty slot, SW1 7 and 8 set for more drives than are fitted (1986 BIOS), the wrong cable or a cable fitted wrongly, a missing or faulty drive, the drive select jumper, and termination.[1]

  • Check the signal cable for shorts between adjacent pins with the meter on ohms ×1K; IBM expects infinity.[2]
  • On IBM Type 2 drives whose file control card is stamped 010 on its left edge, read/write errors are fixed by removing capacitor C39.[2]
  • Do not clean the heads by hand; Tandon's manual forbids it.[12] Belt, speed and rail lubrication on the Tandon drives are in the maintenance guide.

Keyboard faults

For a 301 error, IBM's checks at the keyboard connector on the system board are:[2]

5160 keyboard connector (service manual p. 4-8)
Pin Signal Level
1 Keyboard clock +5 V
2 Keyboard data +5 V
3 Keyboard reset (not used by the keyboard)
4 Ground 0 V
5 +5 V +5 V

The XT needs a PC/XT-protocol keyboard; an AT-only keyboard gives 301.[1] The keyboard interface is discrete logic on the board, with no controller chip.[3]

Expansion card faults

  • A card in slot 8 must return the card-selected signal. IBM's Asynchronous Communications Adapter does so with jumper J13 fitted; a card that does not will not be recognised there.[2] An accelerator card in slot 8 can cause 1801.[9]
  • The async adapter (P/N 8529150) must be set for RS-232C, not current loop, when the diagnostics are run, or they report 1101 errors.[2]
  • IBM notes that diagnostics do not necessarily fail with two cards on one interrupt level even though the operator keeps getting errors; list the cards by level and remove one of any pair that share a level.[2]
  • IBM reserved C0000 to C7FFF for video ROMs, and XT fixed disk controllers usually put their ROM at C8000. Microsoft's article Q63588 says a few VGA ROMs extend past C8000.[9]
  • Some 16-bit cards sold as working in an 8-bit slot mean an 8-bit slot in an AT, or need reconfiguring for 8-bit use.[9]
  • For communications faults on a machine with a Display Station Emulation Adapter, IBM checks the DMA chip at U28. If it is an AMD 8237 or 9517, a 74LS32 must be fitted at U90 or the spare position; boards without it are to be replaced (ECA005).[2]

Known problems

  • If the POST keeps restarting with no error, SW1 switch 1 is on.[9][5]
  • With the 01/10/86 and 05/09/86 BIOS, some third-party floppy controllers read 1.44 MB diskettes but cannot boot from them.[13]
  • IBM does not support the 5161 Expansion Unit on models 088, 089, 267, 268, 277 and 278.[14]
  • Early boards are wired for Mostek MK38000-series mask ROMs in U18 and U19. A 27256 or 27C256 EPROM usually works there, but not always.[9]
  • Replacing the 8088 with a V20 sometimes causes problems.[9]
  • A POST card shows nothing useful on a 5160.[3]

References

  1. ↑ 1.0 1.1 1.2 1.3 1.4 1.5 1.6 IBM 5160 – POST – Some errors (on-screen, or otherwise), minuszerodegrees.net. Meanings and causes of the POST beeps and messages, the continuous tone, 1701, 1801, the ROM messages and PARITY CHECK 1 and 2, the SW1 video-setting symptoms, and the note that the 1986 BIOS halts without any output if U18 or U19 fails its early check.
  2. ↑ 2.00 2.01 2.02 2.03 2.04 2.05 2.06 2.07 2.08 2.09 2.10 2.11 2.12 2.13 2.14 2.15 2.16 2.17 2.18 2.19 IBM, Personal Computer Family Service Information Manual, SA38-0037-00, First Edition, January 1989, hosted as File:IBM Personal Computer Family Service Information Manual SA38-0037-00.pdf. Chapter 4, "5150/5160-Based System Diagnostics": POST sequence and timing, dead machine tips and connector voltage limits (p. 4-3); dummy load (p. 4-4); memory error-code decoding (pp. 4-5 to 4-7); E-2 jumper, mislabelled boards and interrupt conflicts (p. 4-7); keyboard connector voltages (p. 4-8); Type 2 diskette drive C39 and terminators (p. 4-9); drive signal cable and power checks (p. 4-11); CHKDSK capacity check (p. 4-12); fixed disk cable orientation, slot 8 and 1101 errors (p. 4-13); DSEA and down-level system boards, ECA005 (p. 4-14).
  3. ↑ 3.0 3.1 3.2 3.3 3.4 3.5 IBM 5160 – Basics, for 5160 beginners, minuszerodegrees.net.
  4. ↑ 4.0 4.1 4.2 IBM, Personal Computer XT and Portable Personal Computer Technical Reference, part 6280089, March 1986, hosted as File:IBM PC XT 5155 5160 Technical Reference Mar86.pdf. I/O channel pin assignments (p. 1-17) and signal descriptions for CLK, OSC and RESET DRV (pp. 1-21 to 1-23); SW1 bit map (p. 1-27); power supply outputs, sense levels and power good (pp. 3-3 to 3-5).
  5. ↑ 5.0 5.1 5.2 IBM, Personal Computer Family Service Information Manual, SA38-0037-00 (1989), chapter 6, "5160 PC XT", p. 6-3 (SW1 functions and math coprocessor setting) and p. 6-4 (display, diskette and fixed disk adapter switch settings).
  6. ↑ IBM 5160 – Motherboard Switch/Jumper Settings, minuszerodegrees.net. On/off combinations for each SW1 setting.
  7. ↑ 7.0 7.1 7.2 7.3 7.4 7.5 IBM 5150/5155/5160 – Minimum Diagnostic Configuration, minuszerodegrees.net, and the pages it leads to: no beeps, fan not turning and fan turning.
  8. ↑ Short-circuit on the +12 volt line of an IBM 5150/5155/5160 motherboard, minuszerodegrees.net.
  9. ↑ 9.0 9.1 9.2 9.3 9.4 9.5 9.6 9.7 IBM 5160 – Known Problems/Issues, minuszerodegrees.net. The RAM sizing flaw, shorted tantalums, bank 0 failure, looping POST, slot 8, 16-bit ISA cards, the C8000 ROM conflict, MK38000 ROM sockets, the V20, and the 1986 BIOS notes.
  10. ↑ IBM 5160 – Power Supply Unit, minuszerodegrees.net.
  11. ↑ IBM Fixed Disk Adapter, minuszerodegrees.net. Controller versions and the drive geometries each supports.
  12. ↑ Tandon, TM100-1 / TM100-2 48 TPI Operating and Service Manuals (1983), p. 6-1, hosted as File:Tandon TM100-1 TM100-2 48TPI Operating and Service Manuals 1983.pdf.
  13. ↑ IBM 5160 – BIOS Revisions, minuszerodegrees.net.
  14. ↑ IBM, Personal Computer Family Service Information Manual, SA38-0037-00 (1989), chapter 7, p. 7-1.


Retrieved from "https://wiki.retrotechcollection.com/index.php?title=IBM_PC_XT_Troubleshooting_Guide&oldid=16477"