TEAC FT-3010¶

The TEAC FT-3010 is a QIC-3010-capable floppy-tape drive that also reports QIC-80 mode. It needs a tighter command cadence than the Colorado drives, and it fails hard when a report is left unfinished.

Identification¶

Values as reported by a bench unit (label on the unit: TEAC):

The vendor word is shared: TEAC shipped this generation under the ID ftape lists for the Wangtek 3200. The ftape entry named "Teac FT3010TR" is a different word, 0x03CF (make 15). Trust the label on the unit.

tw currently labels configuration 0x50 as "QIC-40 mode", because bit 7 (the QIC-80 mode flag) is clear. On a QIC-3010 drive with a QIC-3010 cartridge inserted, that bit carries no meaning; read the rate bits only.

Profile¶

Use --profile teac. It inherits generic and uses the colorado wake (Phantom Select + Enter Primary Mode), with tighter timings:

With this profile the bench unit passed 600 of 600 reports in 60 s and two clean load points.

Formats and rates¶

Quirks¶

• The command time-out is about 2.2 ms. The drive acknowledges a report 2.2 ms after the last Step pulse, the bottom of Rev J's TTIMEOUT range (2.2–2.9 ms). At a 2.2 ms Step period a pulse train can split: the drive sees a stray single pulse (Soft Reset) or a stray five (Alternate Command Time-out). Use the teac profile's 1.5 ms period.
• Alternate Command Time-out moves the ACK to about 6.25 ms. If a split train sends command 5, the drive's time-out becomes Rev J's alternate 6.2–6.9 ms, and reports acknowledge too late for a short ACK window. Only a Soft Reset or a power cycle clears that state. The generic 10 ms ACK window tolerates it.
• Never leave a report half-clocked. An ACK timeout, tw drive scope --pulses N with a report command, or a Soft Reset in the middle of a report leaves the drive in the report subcontext. Later commands are "illegal in report subcontext", and then every output goes dark: no cue Index, no Track Zero, no Write Protect, no response to any select. Only a power cycle brings it back.
• Automatic Seek Load Point waits for the first status report. After a cartridge is inserted, the drive holds its automatic load point until the first Report Drive Status (Rev J permits this, command 14). It then runs for about 30–50 s, during which it is not Ready and often does not acknowledge reports at all. Report Error Code (7) is rejected until it is Ready (Rev J Table 2a), so tw drive status shows it as not supported. wake() waits this out.
• New-cartridge latch. The first load point after insertion can end Ready with the new-cartridge and error bits still set, and a Seek Load Point sent while the latch is set is rejected. Read the error code, then run the load point again.
• Cue Index free-runs. Index pulses arrive every 2.5 ms, 0.9 ms wide, in every state, including during pulse trains and reports. Index-edge report timing and the firmware's wait-ready mean nothing on this drive; the profile uses fixed settle times.
• Unexplained software-reset error. After 11 idle minutes with nothing sent, the bench unit came up with error 27 (Software Reset Occurred) latched and ran its automatic load point on the first status. The cause is not known. wake() absorbs it.

See also¶

• Drives
• Drive profiles
• QIC-117 command interface