Skip to content

helix-lt.md

Latest commit

 

History

History
119 lines (93 loc) · 5.8 KB

helix-lt.md

File metadata and controls

119 lines (93 loc) · 5.8 KB
Raw

Helix LT — device survey

What we know about the Helix LT, from a user's unit on Linux (2026-08-18, firmware string a7e2585). Unlike the Floor survey, this one rests on no USB captures and no device backup: everything below was read by fretwire itself over MI_00, with this PR's device entry applied.

Bottom line: the LT needs no protocol change. The handshake, the preset read, the snapshot decode, and the setlist and preset-list browses all work unmodified once the PID is in the table. The catalogue resolved every block and parameter of the loaded preset by name.

Nothing has been written to this device — no edit, no save, no backup, no restore. That is why the entry is Support::Reported rather than Support::Verified: the reads are real traffic checked against real parsers, but Verified means the edit builders have been reconciled against a unit's own bytes too, and here they have not. (The PR proposed Support::Untested; that tier means "only the USB IDs are known", which undersells this survey. Reported was added on 2026-08-20.)

USB identity [solid]

Helix Floor Helix LT
VID / PID 0x0E41 / 0x4248 0x0E41 / 0x424A
bcdDevice 0x0200 0x0200
product string — HELIX

Interface layout, read from Linux sysfs — the same six-interface shape the Floor has, with the vendor control channel at interface 0:

iface class kernel driver
0 Vendor (0xFF) none — this is MI_00
1 Audio control snd-usb-audio
2 Audio streaming snd-usb-audio
3 Audio streaming snd-usb-audio
4 Audio / MIDI snd-usb-audio
5 HID usbhid

CONTROL_INTERFACE, EP_IN and EP_OUT needed no change, and claim_interface(0) succeeded on the first try.

Identity and firmware [solid]

The handshake identity reply reports P21 — the Floor's model code, not a code of its own. The pulled preset agrees, at key 7:

key value
36 "P21\0"
37 "a7e2585\0"
35 57737248 = 0x03710020

Key 35 is recorded raw on purpose. docs/helix-floor.md notes device_version = 0x03800000 for firmware 3.82, which would make 0x03710020 read as ~3.71 — but that encoding has never been pinned down against two known firmware versions, so this is data, not a conclusion.

Preset model [solid]

Reading the loaded preset (8326 bytes, declared length matched):

  • Both DSPs. Preset key 1 is populated and blocks came back in slots ≥ 20 as well as in the first group, i.e. the global slot = dsp * 20 + index numbering the Floor established. The unit reported DSP1 71.0% used and DSP2 43.0%.
  • 8 snapshots, SNAPSHOT 1…SNAPSHOT 8, and the stored active index agreed with the live scene.
  • Every block resolved to a .models definition by name, with device-ordered named parameters — no unmatched model, no unmatched parameter key.

Maintainer note (2026-08-20). The survey as submitted gave two different block counts — a slot list of 21–28 plus 3–6 (twelve slots) in one bullet and "all 8 blocks" in the next — and the question went unanswered, so the exact figure is not recorded here. Nothing rests on it: dsps: Some(2) needs only that some block occupies a slot ≥ 20, which both readings agree on, and which preset key 1 being populated says independently. If you have an LT, fretwire pull prints the slot list and settles it.

Setlists [arity solid; names not corroborated]

Browsing banks 0, 1, 2, 6, 7 and 8 (3–5 were not reported, and it is not recorded whether they were browsed and dropped in editing or never browsed):

bank result
0 128 presets — US Double Nrm, Essex A30, Brit Plexi Brt (factory)
1 128 presets
2 128 presets — the user's own presets; read-info on the loaded preset reported bank: 2, index: 0, name: "WIP"
6 128 presets
7 128 presets — Quick Start, Parallel Spans, SNP:4-Amp Spill (templates)
8 refused, code -3

So: eight banks of 128, bank 0 factory and bank 7 templates — the Floor's layout. Bank 8 being refused bounds the arity at eight on its own, so the three unreported banks do not weaken that. These browses ran with this PR's entry applied, i.e. against its explicit setlist_size: Some(128); they corroborate that value rather than exercising setlist_stride()'s fallback, which never ran.

The names in DEVICES are therefore the Floor's. Worth being explicit about the difference in evidence: the Floor's names came from the eight L6Setlist streams of a real .hxb, whereas here only the arity and the character of the two end banks were observed. If an LT backup ever turns up and disagrees, this is the field to fix.

Not observed

  • preset_device_id. The handshake identity reply carries no 0x0021xxxx device id, and the wire preset stream has no such field — the Floor's 0x210001 came from a .hxb. Left None rather than copied across.
  • Every write path. Nothing was sent to this unit beyond reads and the handshake.
  • How the screen banks presets. Device::presets_per_bank — the 3 in an HX Stomp's 01A/01B/01C/02A — is None for the LT, so its presets are listed by slot number. One look at the unit's screen settles it. The Floor's is unknown too, so there is nothing to inherit.

Reproducing

With the LT connected and the udev rule installed:

fretwire detect                 # Helix LT: present (reported working, unverified)
fretwire connect                # handshake OK — device reports "P21"
fretwire pull                   # the preset above, blocks and params resolved
fretwire setlists               # 8 setlists
fretwire presets 7              # the templates bank
fretwire dump-raw lt.raw        # then: fretwire tree lt.raw