Jump to content

Tandon TM100-2 Head Alignment

From RetroTechCollection
Tandon TM100 head module retaining and cam screws (manual figure 5-7)
Tandon's Figure 5-7. Left, the drive from above with the back at the top: the module retaining screw and the cam screw at the rear of the head module. Right, the drive from below: the other two module retaining screws beside the spindle pulley.

The Tandon TM100-2 moves its heads with a four-phase stepper motor. Each phase turns the motor 3.6 degrees, and a capstan and split band turn that into a movement of one track of the head carriage.[1] Radial track alignment sets the position of the whole head module so that the heads sit at the standard distance from the hub centre. Tandon calls for it only after servicing the drive or when diskette interchange problems are suspected.[2] The module is held by three retaining screws and moved by a cam screw at its rear.[3]

Tandon's manual covers the single-sided Tandon TM100-1 with the same procedure. Its check of the upper head applies "if applicable", so on a TM100-1 check and adjust head 0 only.[4]

Tandon's own procedure needs a Dysan 224/2A alignment diskette and an oscilloscope.[5] Philip Avery published a method that uses an ordinary diskette formatted on a known-good drive instead, and RickNel added a variant that reads the drive with ImageDisk.[6][7]

Symptoms

[edit | edit source]

Minuszerodegrees.net describes alignment drift as something that happens over many years, but not on every drive. It is indicated when both of these are true: diskettes formatted in the drive work in the same drive, and the drive will not read good 360 KB diskettes formatted on a different, known-good drive. The site's author strongly advises against attempting an alignment without that evidence, because the attempt can leave the drive worse than before.[8]

Terry Stewart found the same pattern on his IBM PC (5150). The machine dropped into ROM BASIC instead of booting from drive A:, and booted normally with the drives swapped, which cleared the controller and the boot diskette. The faulty drive could format, read, write and boot from diskettes it had prepared itself, but could not read his existing diskettes, and the good drive could not read diskettes prepared on the faulty one.[9] Tandon's troubleshooting table gives "alignment off" as a cause of a drive that will not read, with the action "Verify Cats Eye lobes, index burst, and Track 0 switch adjustment".[10]

Rule these out first, because they give similar symptoms:

  • A dirty or misadjusted Track 0 switch. Chuck(G) on the VCFed forum has found that many "alignment" problems on Tandon 100 drives are a dirty track 0 sensor.[11] On the cctalk mailing list, Chuck Guzis advised cleaning the sensor before any re-alignment, because everything is keyed off the track 0 reference.[12] See Tandon TM100-2 Track 0 Switch Repair.
  • A loose split-band screw on the stepper motor. Avery says these commonly come loose and cause inconsistent tracking errors.[6]
  • Spindle speed out of tolerance. Tandon's index check starts by checking the motor speed.[13] See Tandon TM100-2 Spindle Speed Adjustment.
  • A diskette that buckles between the heads. RickNel found one head not reading at all, and the diskette buckled where it passed between the heads. The plastic frame on top of the aluminium chassis may have shrunk, or been pressed back too firmly after cleaning, so that the diskette slot was no longer level with the fixed head. Levering one or two of the frame's holding posts about 1 mm out of their holes cured it. A diskette should slide in without pushing, shoving or twisting.[7]

Avery adds that the drive must first format, verify, read and write reliably on diskettes it has written itself. All an alignment does is let the drive exchange diskettes with other drives.[6] For other drive faults see Tandon TM100-2 and Floppy Drive Repair.

Tools and parts

[edit | edit source]

For Tandon's procedure:[5]

  • A dual-channel wideband oscilloscope, Hewlett-Packard 1740A or equivalent.
  • An exerciser (AVA Model 103C or equivalent), or software that can step the drive to any track, select the upper or lower head and write a 1F or 2F pattern.
  • A power supply for the drive.
  • A certified alignment diskette, Dysan 224/2A or equivalent.
  • A 7/64-inch Allen wrench for the module retaining screws, and a flat-blade screwdriver for the index sensor.[3][13]

For Avery's method:[6][7]

  • An oscilloscope. RickNel did without one by using ImageDisk's read tests.
  • A computer that can drive the TM100-2 and software that seeks to a chosen track and reads it continuously, ignoring errors. Avery used Floppy Doctor on a TRS-80 Model III and notes that a short BASIC program can do the same job; RickNel used ImageDisk, Dave Dunfield's free DOS utility, on a PC.
  • A strong magnet and a diskette, plus a known-good drive to format it.
  • Insulating material for the logic board pins above the cam, or a screwdriver with an insulated shank.

If the heads fail Tandon's azimuth check, the head module has to be replaced: 171113-001 "Module DSR" for the TM100-2 or 171113-002 "Module SSR" for the TM100-1.[14][15]

Procedure

[edit | edit source]

Before adjusting

[edit | edit source]
Tandon TM100-2A split band screw on stepper motor (Philip Avery article)
Underside of a TM100-2A. The arrow marks the screw that clamps the split band to the stepper motor capstan, under its plastic cover. (Photo: Philip Avery's article on Terry Stewart's blog)
  1. Inspect the drive with the power off. Check for loose or missing hardware. Check that the front latch opens and closes and that the head arm rises when the door is opened. Turn the hub by hand; it should turn freely. Check that the circuit boards are secure, the connectors firmly seated and that a diskette goes in and comes out easily.[16]
  2. Check the spindle speed. See Tandon TM100-2 Spindle Speed Adjustment.[13]
  3. Clean and check the Track 0 switch. See Tandon TM100-2 Track 0 Switch Repair.[12]
  4. Check that the split-band screw on the stepper motor is tight. Avery reaches it by removing one plastic pin and swinging the plastic cover aside.[6]

Checking alignment with the Dysan 224/2A

[edit | edit source]
Tandon TM100 logic board test points (manual figure 5-1)
Tandon's Figure 5-1: the logic board test points. TP1, TP2 and TP10 are along the top of the board; TP7 is beside the terminator socket and TP6 above connector P11.
Tandon TM100 cats eye patterns (manual figure 5-6)
Tandon's Figure 5-6: cat's eye patterns from even amplitude down to one lobe at 40 % of the other.
  1. Leave the alignment diskette and the drive at room temperature for one hour before starting.[4]
  2. Set up the oscilloscope: channel A on Test Point 1, channel B on Test Point 2, ground on Test Point 10. Read differentially, A plus B with B inverted. Set the time base to 20 milliseconds per division and trigger externally from Test Point 7 on the positive edge. Adjust the amplitude for at least four divisions, AC coupled.[4] Test Point 7 carries the index pulse, which comes round every 200 ms on a soft-sectored diskette.[17]
  3. Apply power to the drive and select it.[4]
  4. Insert the Dysan 224/2A alignment diskette.[4]
  5. Select head 0, the lower head, and read track 16. Track 16 lies 1.9167 inches from the centre of the hub; the other tracks are spaced from it at 48 tracks per inch.[4]
  6. Adjust the oscilloscope to show the cat's eye pattern (Figure 5-6). Check that the smaller of the two lobes is not less than 75 % of the amplitude of the other. Tandon's 75 % figure applies to an alignment diskette verified against a standard alignment diskette.[4]
  7. Step the drive to track 0, then back to track 16, and check the pattern again.[4]
  8. Step the drive to track 26 or higher, then back to track 16, and check the pattern again.[4]
  9. Switch to head 1, the upper head, read track 16 and make the same three checks: at track 16, after stepping to track 0 and back, and after stepping to track 26 or higher and back. A TM100-1 has no upper head; skip this step.[4]
  10. If every check passes, the radial alignment is acceptable. If any check fails, adjust the module.[4]

Adjusting with the Dysan 224/2A

[edit | edit source]
Tandon TM100-2A head module retaining screws underneath (Philip Avery article)
The two module retaining screws on the underside of a TM100-2A (arrows). (Photo: Philip Avery's article on Terry Stewart's blog)
Tandon TM100-2A top module retaining screw beside cam (Philip Avery article)
The rear of a TM100-2A, servo board in front: the arrow marks the module retaining screw on top of the head module, beside the cam. (Photo: Philip Avery's article on Terry Stewart's blog)
  1. Loosen the three module retaining screws half a turn with a 7/64-inch Allen wrench. Two are underneath the module and one is at the rear of the module (Figure 5-7).[3]
  2. Turn the cam screw counterclockwise.[3]
  3. Watch the cat's eye pattern of the head that is farthest out of alignment.[3]
  4. Turn the cam screw until the two lobes are equal in amplitude.[3]
  5. Tighten the module retaining screws with the 7/64-inch Allen wrench.[3]
  6. Repeat the whole alignment check on each head.[3]

Tandon gives no figure for how far to turn the cam. Avery turns it no more than half a turn either way.[6] RickNel found that tightening the screws can move the alignment slightly, and tightened them one at a time while still testing.[7]

If the cat's eye pattern cannot be verified or adjusted, Tandon's troubleshooting table lists four causes: the drive cannot read (go to "will not read"), it cannot step (go to "will not seek or restore"), the Track 0 sensor is misadjusted (do the Track 0 switch adjustment), or the alignment diskette is defective (check it and replace it if necessary).[10]

Aligning without an alignment diskette

[edit | edit source]
Tandon TM100-2A turning the alignment cam screw (Philip Avery article)
Turning the cam screw of a TM100-2A with a long flat-blade screwdriver passed down beside the logic board. (Photo: Philip Avery's article on Terry Stewart's blog)

Avery's method aligns the drive to a diskette written by a well-aligned drive. He describes it as less accurate than the official method with an alignment diskette.[6]

  1. Bulk-erase a diskette by passing a strong magnet within about 20 mm of it, several times over the whole surface. Erasing it avoids aligning to old, possibly misaligned tracks, and lets you check later that the drive is not one track out.[6]
  2. Format the erased diskette in a drive that is known to be well aligned, one that reads and writes diskettes from other drives. On a two-drive system, use that system's good drive.[6]
  3. Put the diskette in the drive under test, seek to a middle track such as track 20, and read a sector there continuously, ignoring errors.[6]
  4. Connect the oscilloscope to Test Point 1 or Test Point 2, with its ground on Test Point 6. Avery sets 0.1 V per division and 20 µs per division and expects a signal of 0.3 to 0.6 V peak to peak. If there is no signal, insert a diskette that the drive under test wrote itself to confirm the settings and connections and to see the best-case signal, then go back to the diskette from the good drive.[6]
  5. Loosen the two retaining screws underneath the drive and the one on top beside the cam. Cover the logic board pins above the cam with insulation, or use an insulated-shank screwdriver on the cam screw.[6]
  6. Turn the cam slowly, no more than half a turn in either direction, while watching the oscilloscope, and stop at the maximum signal.[6]
  7. Tighten the top retaining screw only mildly. Step across the diskette and check that every track gives a usable signal. Above about track 30 the signal falls off, because the inner tracks are shorter; on a good drive it ranges from 0.6 V down to nearly 0.3 V. Avery sometimes needs several attempts to get a good signal across the whole diskette.[6]
  8. Tighten all three retaining screws.[6]
  9. Check that the drive has not been set one track out. Step to track 40, one beyond the drive's 0 to 39, where there should be no signal. Step back to track 39, where the signal should first appear. Keep stepping down to track 0: there should be a good signal there, the Track 0 sensor should operate, and the drive should not step below track 0.[6]

RickNel's variant with ImageDisk

[edit | edit source]

RickNel aligned a TM100-2 from his S-100 machine this way and found the read and write tests enough, although he notes that the oscilloscope check gives a more fundamental alignment.[7]

  1. For better access, take the logic board off. Unplug all its cables, remove its two screws, then slide and wiggle it free.[7] Tandon's own removal steps are to unplug every connector, noting where each goes, remove the two mounting screws, slide the board about half an inch towards the back of the drive and lift it off.[18]
  2. Unlace the two head cables, which come from the front of the drive, from their cradles in the plastic frame. Reconnect the board standing on its end at the back of the drive, leaving slack in the head cables so that they do not drag on the head carriage.[7]
  3. Use the ImageDisk alignment function to move the heads to a chosen cylinder and read one head at a time. It reads the same track repeatedly until told to stop.[7]
  4. Loosen the retaining screws and turn the cam back and forth until both heads read consistently at track 20.[7][6]
  5. Step through all the tracks on both sides, making finer adjustments until every track reads.[7]
  6. Use ImageDisk's test reads and writes to confirm at least the outermost, innermost and centre tracks.[7]
  7. Keep the drive running the tests while tightening the screws gently, one at a time.[7]

Follow-up checks

[edit | edit source]

Tandon's index check should be made after the drive has been aligned. The index adjustment sets the time from the index pulse to the start of data, which matters for hard-sectored diskettes.[13]

  1. Set up the oscilloscope: channel A on Test Point 1, channel B on Test Point 2, ground on Test Point 6, A plus B with B inverted, 50 µs per division, external trigger from Test Point 7 on the negative edge, at least two divisions of amplitude.[13]
  2. Apply power, select the drive, check the motor speed and insert the alignment diskette.[13]
  3. Seek to track 1, select head 0 and read the time from the trigger point to the start of the first data pulse. The tolerance is 200 ± 100 µs. On a TM100-2, check head 1 as well.[13]
  4. When track 1 is within tolerance, check the index on track 34, on both heads of a double-sided drive.[13]
  5. If any reading is out of tolerance, loosen the index sensor's retaining screw a quarter turn counterclockwise from the bottom of the chassis, move the sensor with a flat-blade screwdriver until the data starts 200 ± 100 µs after the trigger, tighten the screw and recheck every index reading.[13]

Tandon also adjusts the Track 0 stop after the cat's eye pattern has been adjusted. That procedure, and the Track 0 switch check, are in Tandon TM100-2 Track 0 Switch Repair.[19]

The head azimuth is checked on track 34 with the third azimuth burst of the alignment diskette and must be within ± 18 minutes on each head. It cannot be adjusted; a head module that fails is replaced.[14] After a head module replacement, Tandon's table of checks calls for the radial track alignment, index, azimuth, Track 0 switch, head output and cone centring checks, and on single-sided drives the compliance check as well.[20]

Testing

[edit | edit source]
  • With Tandon's procedure, every cat's eye check passes on each head, and the index readings are within 200 ± 100 µs.[4][13]
  • With Avery's method, the track count check passes: no signal at track 40, a signal from track 39 down to track 0, and the Track 0 sensor operating at track 0.[6]
  • Exchange diskettes with a known-good drive in both directions. After Avery aligned Terry Stewart's drive, it booted his existing PC-DOS diskettes, diskettes it prepared read in his second drive and the reverse, and it ran a heavy workout in the 5150 without errors.[9]

Revision notes

[edit | edit source]

Tandon's 1983 manual describes the TM100-1 and TM100-2.[1] Avery's article and photographs are of a TM100-2A, with the two lower retaining screws, the top retaining screw and the cam in the places Tandon's Figure 5-7 shows them.[6][3]

[edit | edit source]

References

[edit | edit source]
  1. ↑ 1.0 1.1 Tandon Corporation, OEM Operating and Service Manual, TM100-1, TM100-2 Disk Drives, 48 Tracks Per Inch, 179022-001 Rev. B (1983), section 4.4 "Read/Write Head Positioner", p. 4-5: unipolar four-phase stepper motor, 3.6 degrees per phase; the capstan and split band translate this into a one-track movement of the head carriage. Hosted as File:Tandon TM100-1 TM100-2 48TPI Operating and Service Manuals 1983.pdf.
  2. ↑ Tandon, TM100 manual (1983), section 5.6 "Radial Track Alignment Check and Adjustment", introduction, p. 5-4: the procedure locates the head at the proper radial distance on the hub centre line; adjustment is necessary only after servicing or if diskette interchange problems are suspected.
  3. ↑ 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 Tandon, TM100 manual (1983), section 5.6, "Radial Track Alignment Adjustment", steps 1 to 6, pp. 5-8 and 5-9, and Figure 5-7 "Head Module Retaining and Cam Screws".
  4. ↑ 4.00 4.01 4.02 4.03 4.04 4.05 4.06 4.07 4.08 4.09 4.10 4.11 Tandon, TM100 manual (1983), section 5.6, "Radial Track Alignment Check", steps 1 to 21, pp. 5-6 to 5-8, with Figure 5-5 (hub centre line and track locations) and Figure 5-6 (cat's eye patterns): one hour at room temperature, oscilloscope settings, Dysan 224/2A, track 16 at 1.9167 in, the 75 percent limit and its note, the checks at tracks 0 and 26, and the upper-head check "if applicable".
  5. ↑ 5.0 5.1 Tandon, TM100 manual (1983), section 5.2 "Equipment Required", pp. 5-1 and 5-2: oscilloscope, exerciser or software, power supply, Dysan 224/2A alignment diskette, Dysan 104/2D output diskettes and hand tools including 5/64-inch and 7/64-inch Allen wrenches.
  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 6.13 6.14 6.15 6.16 6.17 6.18 Philip Avery, "Adjusting the radial alignment a Tandon TM100-2A drive without an alignment disk", 23 June 2010, published on Terry Stewart's (Tezza's) blog, 28 June 2010: the split-band screw, the bulk-erased reference diskette, oscilloscope settings and expected signal, the three retaining screws and the cam, and the track count check.
  7. ↑ 7.00 7.01 7.02 7.03 7.04 7.05 7.06 7.07 7.08 7.09 7.10 7.11 RickNel, addendum of 2 November 2010 to Avery's article on the same page, classic-computers.org.nz: aligning with ImageDisk read and write tests, tightening the screws while testing, removing the logic board for access, and the buckled diskette caused by the plastic frame.
  8. ↑ Minuszerodegrees.net, "Tandon TM100-2 in IBM 5150/5155/5160 – Known Problems/Issues", section "Incorrect head alignment": drift over many years on some drives, the two signs that indicate it, and the advice not to attempt an alignment without them.
  9. ↑ 9.0 9.1 Terry Stewart, "Tandon TM100-2A disk drive alignment on an IBM-PC (5150)", 29 June 2010: diagnosis of the misaligned drive and the result after Avery's alignment.
  10. ↑ 10.0 10.1 Tandon, TM100 manual (1983), Table 6-1 "Troubleshooting Guide", pp. 6-3 and 6-4: "Will not read" and "Fails Cats Eye alignment check and adjustment".
  11. ↑ Chuck(G), post of 11 October 2023 in the thread "Floppy drive alignment disks", Vintage Computer Federation Forums: "many of the 'alignment' issues on Tandon 100s" are a dirty track 0 sensor. Quoted on the minuszerodegrees.net TM100-2 page.
  12. ↑ 12.0 12.1 Chuck Guzis, "Tandon Track 0 adjustment", cctalk mailing list, 28 August 2021: clean the track 0 sensor before re-aligning, since everything is keyed off the track 0 reference.
  13. ↑ 13.00 13.01 13.02 13.03 13.04 13.05 13.06 13.07 13.08 13.09 Tandon, TM100 manual (1983), section 5.7 "Index Checks and Adjustment", pp. 5-9 and 5-10, with Figures 5-8 and 5-9: when to check, oscilloscope settings, motor speed check as step 4, tracks 1 and 34, tolerance 200 ± 100 microseconds, and the index sensor adjustment.
  14. ↑ 14.0 14.1 Tandon, TM100 manual (1983), section 5.8 "Azimuth Check", p. 5-11: track 34, third burst, ± 18 minutes; the azimuth is not adjustable and a head that fails means replacing the module assembly.
  15. ↑ Tandon, TM100 manual (1983), Appendix A "Recommended Spares and Major Assemblies", p. A-1: 171113-001 Module DSR and 171113-002 Module SSR.
  16. ↑ Tandon, TM100 manual (1983), section 5.1 "Visual Inspection", p. 5-1.
  17. ↑ Tandon, TM100 manual (1983), Figure 4-5 "Soft-Sectored Index Pulse", p. 4-6: Test Point 7, 4 ms average pulse every 200 ms; Figure 5-1 "Logic Circuit Board with Test Points", p. 5-2.
  18. ↑ Tandon, TM100 manual (1983), section 6.4, "Logic Circuit Board" removal, pp. 6-7 and 6-9, with Figure 6-2.
  19. ↑ Tandon, TM100 manual (1983), section 5.10 "Track 0 Stop Adjustment", p. 5-13: the stop screw should be adjusted after the cat's eye pattern has been adjusted.
  20. ↑ Tandon, TM100 manual (1983), Table 6-2 "Guide to Checks and Adjustments", p. 6-6: checks after replacing the head module assembly.


Retrieved from "https://retrotechwiki.com/index.php?title=Tandon_TM100-2_Head_Alignment&oldid=32703"