Colorado Jumbo 350¶
The HP/Colorado Jumbo 350 is a QIC-80 floppy-tape drive and the drive Tapewyrm was brought up on. It is the reference drive: every QIC-80 tape recovered so far was dumped on one.
Identification¶
Values as reported by a bench unit (jumperless, internal):
| Report | Raw value | Decoded |
|---|---|---|
| Vendor ID (32) | 0x0047 |
Legacy whole-word ID 71, Colorado Memory Systems |
| ROM version (9) | 0x58 |
ROM 88 |
| Drive configuration (8) | 0x90 |
QIC-80 mode, 500 kbit/s |
| Drive status (6) after power-on | 0x77 |
Ready, error, cartridge present, new cartridge, referenced, at BOT |
| Error code (7) after power-on | 0x001A |
Error 26, Power On Reset Occurred |
QIC-117 Rev J lists Colorado's make as "4 & 71". Split into make and model
fields, 0x0047 would read as make 1, model 7, so tw treats it as the
legacy whole-word ID 71. The Colorado 1400 reports a split word with make 71
instead (Colorado 1400).
Profile¶
Use --profile colorado.
| Setting | Value | Why |
|---|---|---|
| Wake | Phantom Select (46) + argument 0, then Enter Primary Mode (30) | The drive is a phantom drive: it ignores the drive-select lines until phantom-selected. |
| Step pulse width | 200 µs | Inside Rev J's 1–1100 µs. |
| Step gap | 2000 µs (2.2 ms period) | Over Rev J's 2.1 ms TSTEP maximum; this drive tolerates it. |
| Terminate gap | 3000 µs | Ends a pulse train past the 2.9 ms command time-out. |
| Report settle | 900 µs | Rev J TBIT. |
Formats and rates¶
| Format | Rate | Status |
|---|---|---|
| QIC-80 | 500 kbit/s | Tested. |
| QIC-40 | 250 or 500 kbit/s | Untested. |
Quirks¶
- A bare Phantom Select does nothing. The drive ignores command 46 unless
it is followed by its N+2 unit argument (unit 0, two pulses). Without it the
drive stays silent and every report times out. The
coloradoprofile sends the argument. - Cue Index pulses confirm selection. Once selected and ready, the drive
sends cue Index pulses about 2.9 ms apart (QIC-117 Rev J §1.3). If you see
them under
tw drive scope, the drive is selected; look elsewhere before blaming timing. - The ACK arrives about 3 ms after the last Step pulse. This is within Rev J's TACK of the command time-out.
- Do not end the wake sequence with Soft Reset. A reset deselects a
phantom drive (Rev J command descriptions, Soft Reset). The profile ends
with Enter Primary Mode instead, and
twreads and clears the latched power-on error afterwards. - No Report Tape Status or Report Format Segments. The drive never acknowledges commands 33 or 37. Tapewyrm takes the format from the tape's header segment instead, so this does not affect recovery.
- Do not send a report while one is in progress. If a report is left half-clocked, the next report command is not acknowledged. Finish or abandon the report with Report Next Bit pulses first.
- Error 41 after reselect. After a Phantom Deselect and reselect the drive latches error 41 (Drive Wakeup Reset Occurred). It is benign; reading the error code clears it.
- Dirty heads look like bad tape. On the bench, a patch of one track that read badly on one pass read cleanly on a later pass after cleaning. Clean the head and retry a bad track before you conclude the tape is damaged.