Xiegu X6200 — CAT (CI-V) reference vs implementation¶

Manual: Radioddity "XIEGU X6200 CI-V implementation V1.0.6" (Firmware V1.0.6, dated 2025-06-20, 13 pp.) — official, tabular, CONFIRMED for its documented subset. Cross-referenced against Hamlib rigs/icom/xiegu.c (x6100_priv_caps, reused by X6200; RIG_MODEL_X6200 = 3091). Sources + verbatim extract: manuals/x6200.txt (+ committed manuals/x6200_civ_v1.0.6.pdf). Confidence: high for documented opcodes; Hamlib X108G_LEVELS/FUNCS over-advertise — the Radioddity table wins. Driver: no x6200/ dir — X6200 routes through Ic705SerialRadio (src/rigplane/backends/ic705/serial.py) → _IcomSerialRadioBase → CoreRadio (src/rigplane/runtime/radio.py, all get_/set_ live here). Factory: backends/factory.py:85 maps model in ("IC-705","X6200") to Ic705SerialRadio; personality comes from rigs/x6200.toml (model="X6200"), not a distinct class. Profile: rigs/x6200.toml (CI-V addr 0xA4, 19200 bps SERIAL-B). Validation: src/rigplane/validation/registry/*. Live run: not run — live serial access unavailable to this audit. All statuses are doc-vs-code; every "live" column is not run.

Protocol: CI-V-derived (FE FE A4 00 .. FD), narrower than Icom. Disambiguated from IC-705 (same 0xA4) by Xiegu-only 0x1D 0x19 model-ID → 6200. CH342 dual-UART. 0x18 power on/off intentionally absent (the documented wedge trigger — MOR-170).

Routing fact that drives the gaps below: CoreRadio setters/getters build CI-V frames from hardcoded opcodes in commands/* builders, not from the profile [commands] map (the cmd_map= path is never passed for these). The profile gate is _require_capability(<feature>). So a declared [capabilities] feature activates a method whose opcode is fixed in code — which may not match a Xiegu-documented opcode. The [commands] opcode table in x6200.toml is largely descriptive, not wired for the legacy freq/mode/func paths.

Command matrix (operator-facing)¶

Legend — Documented: ✅ in Radioddity V1.0.6 table · ⚠️ narrative-only / uncertain · ❌ absent. Validation: RMVR = round-trip readback check · presence = <cap>.presence only · — = no check / cap absent · structural = capability="".

CAT (X6200 opcode)	Documented	Backend method (CoreRadio)	Profile cap / cmd	Validation check_id · live
`0x25 0x00/0x01` selected/unselected freq	✅	`get_selected_freq`/`set_selected_freq`/`_unselected_` (dual-RX mixin)	`get/set_selected_freq`, `get/set_unselected_freq`	`freq.write`,`freq.reverse_sync` (structural) · not run
`0x03`/`0x05` freq (legacy)	⚠️ narrative only	`get_freq`/`set_freq` → main RX path uses hardcoded `0x03/0x05`	`get_freq=[0x03]`,`set_freq=[0x05]` (declared, not wired)	via `freq.*` structural · not run
`0x26 0x00/0x01` selected/unselected mode	✅	`get_selected_mode`/`set_selected_mode`/`_unselected_`	`get/set_selected_mode`,`get/set_unselected_mode`	`mode.set` (structural) · not run
`0x04`/`0x06` mode (legacy)	❌ (not in table)	`get_mode`/`set_mode` → main path hardcoded `0x04/0x06`	`get_mode=[0x04]`,`set_mode=[0x06]` (declared, not wired)	`mode.set` · not run
`0x02` RX freq range	✅	`get_band_edge_freq`	`get_band_edge_freq=[0x02]`	—
`0x07 0x00/0x01` VFO A/B select; `0x07 0xB0` swap	✅	`get/set_vfo_slot` (DualRxRuntimeMixin), `swap_vfo_ab`	`get/set_vfo=[0x07]`; `[vfo] scheme="ab"`	`vfo_slot.set` (structural RMVR) · not run
`0x0F 0x00/0x01` split	✅ (set only)	`set_split`/`get_split` (hardcoded `0x0F`)	`split`; `set_split=[0x0F]`	`split.set` RMVR · not run
`0x11` attenuator on/off	✅	`get/set_attenuator`	`attenuator`; `get/set_attenuator=[0x11]`	`attenuator.set` RMVR · not run
`0x16 0x02` preamp	✅	`get/set_preamp`	`preamp`; `get/set_preamp=[0x16,0x02]`	`preamp.set` RMVR · not run
`0x16 0x12` AGC	✅	`get/set_agc`	`agc`; `get/set_agc=[0x16,0x12]`	`agc.set` RMVR · not run
`0x16 0x22` noise blanker	✅	`get/set_nb`	`nb`; `get/set_nb=[0x16,0x22]`	`nb.set` RMVR · not run
`0x16 0x40` NR on/off	✅	`get/set_nr` (hardcoded `0x16 0x40`)	`nr`	`nr.set` RMVR · not run
`0x16 0x44` COMP on/off	✅	`get/set_compressor` (hardcoded `0x16 0x44`)	`compressor`	presence-only · not run
`0x16 0x50` key/dial lock	✅	`get/set_dial_lock`	`dial_lock`; `get/set_dial_lock=[0x16,0x50]`	`dial_lock.set` RMVR · not run
`0x14 0x01` AF	✅	`get/set_af_level`	`af_level`; `[0x14,0x01]`	`af_level.set` RMVR · not run
`0x14 0x02` RF gain	✅	`get/set_rf_gain`	`rf_gain`; `[0x14,0x02]`	`rf_gain.set` RMVR · not run
`0x14 0x03` squelch	✅	`get/set_squelch`	`squelch`; `[0x14,0x03]`	`squelch.set` RMVR · not run
`0x14 0x06` NR level	✅	`get/set_nr_level`	`nr`; `[0x14,0x06]`	— (cap `nr` covered by `nr.set` on/off only)
`0x14 0x09` CW pitch	✅	`get/set_cw_pitch`	`cw`; `[0x14,0x09]`	—
`0x14 0x0A` TX power	✅	`get/set_rf_power`	`tx`; `[0x14,0x0A]`	— (no `rf_power` RMVR check)
`0x14 0x0B` mic gain	✅	`get/set_mic_gain`	`[0x14,0x0B]`	—
`0x14 0x0C` keyer speed	✅	`get/set_key_speed`	`cw`; `[0x14,0x0C]`	`key_speed.set` RMVR · not run
`0x14 0x0D` DNF/notch FC	✅	`get/set_notch_filter` (hardcoded `0x14 0x0D`)	`notch` (no cmd string)	(see notch row below)
`0x14 0x0F` QSK hang	✅	`get/set_break_in_delay` (Icom uses `0x14 0x10`?)	`break_in` (no cmd string)	—
`0x14 0x12` NB level	✅	`get/set_nb_level`	`nb`; `[0x14,0x12]`	—
`0x14 0x15` MONI level	✅	`get/set_monitor_gain`	`[0x14,0x15]`	—
`0x15 0x02/0x11/0x12/0x15` S/pwr/SWR/Vd meters	✅	`get_s_meter`/`get_power_meter`/`get_swr`/`get_vd_meter`	`meters`; `[0x15,..]`	`meters.read` RO · not run
`0x1A 0x01` band/spectrum + band-stacking set	✅	`get_band_spectrum_display` / `set_bsr`	`get_band_spectrum_display=[0x1A,0x01]`	`bsr.select` (cap `bsr` not declared → resolves UNSUPPORTED) · not run
`0x1A 0x03` IF filter width (read)	✅	`get_filter_width` (`set_filter_width` exists)	`get_filter_width=[0x1A,0x03]`; cap `filter_width` not declared	— (`filter_width.set` skipped: cap absent)
`0x1A 0x05 (00 62)` LCD/key LOCK	✅	(no dedicated method; overlaps `dial_lock` 0x16 0x50)	—	—
`0x1C 0x00` PTT	✅	`set_ptt`/`get_transceiver_status`	`tx`; `ptt_on/off=[0x1C,0x00]`	`tx.ptt` (MANUAL) · not run
`0x1C 0x01` ATU (off/on/auto-tune)	✅	`get/set_tuner_status`	`tuner`; `get/set_tuner_status=[0x1C,0x01]`	`tuner.tune` (MANUAL) · not run
`0x19 0x00` radio ID	✅	`get_transceiver_id`	`get_transceiver_id=[0x19,0x00]`	`discovery.identify` (structural) · not run
`0x1D 0x19` Xiegu model ID → 6200	✅	(discovery probe `probe_xiegu_model_id`)	`get_xiegu_model_id=[0x1D,0x19]`	discovery-only · not run
DNF on/off `0x16 0x41`	✅	no method — `notch` cap routes to `set_manual_notch` (`0x16 0x48`, NOT X6200)	`notch` declared	`notch.set` RMVR → wrong opcode (D)
break-in (QSK) on/off	✅ only as `0x14 0x0F` hang time	`get/set_break_in` → hardcoded `0x16 0x47` (NOT X6200)	`break_in` declared	presence-only · not run (D)
VOX	❌ no CI-V	`get/set_vox` → hardcoded `0x16 0x46` (NOT X6200)	`vox` declared	`vox.read`/`vox.set`(TX-blocked) · not run (D)
repeater tone / TSQL / tone freq	❌ no `0x1B`/`0x16 0x42-43`	`get/set_repeater_tone`/`_tsql`/`_tone_freq` → `0x16 0x42/43`, `0x1B 0x00` (NOT X6200)	`repeater_tone`,`tsql` declared	`repeater_tone.set`,`tone_freq.set`,`tsql.set`,`tsql_freq.set` RMVR → all wrong opcode (D)
RIT / XIT `0x21`	❌ no CI-V	`get/set_rit_frequency`/`_tx_status` → hardcoded `0x21`	`rit`,`xit`; `[validation] write_only_controls=["rit","xit","notch"]`	`rit.set`,`xit.set` RMVR (write-only) · not run (D)
scope `0x27`	❌ no CI-V	scope mixin (gated)	cap `scope` not declared (intentional)	`scope.*` resolve UNSUPPORTED · correct
`pbt` (PBT 0x14 0x08/0x09)	❌ no CI-V	`get/set_pbt_inner/outer` (hardcoded `0x14 0x08/09`)	`pbt` declared; `[controls.pbt_inner/outer]`	no registry check → presence-only (D)
`data_mode` `0x26` data-flag / `0x1A 0x06`	partial (via 0x26 byte)	`get/set_data_mode` (hardcoded `0x1A 0x06`)	`data_mode` declared	no registry check → presence-only
`scan`	❌ no CI-V	none	`scan` declared	no registry check → presence-only

Band-stacking register set (0x1A 0x01 D0+D1) — band-recall UI, not a control round-trip (registry bsr exists but cap undeclared).
LCD backlight (0x14 0x19) and 0x1A 0x05 LOCK — front-panel display state.
No memory family, no scan-table CI-V documented for X6200 — nothing to audit.

Gap lists (priority-ordered)¶

A. UNDER-DECLARED — backend implements + opcode is X6200-documented, but profile/registry can't see it¶

CAT (X6200)	rigplane symbol	Fix	Ticket
`0x1A 0x03` filter width	`get/set_filter_width` exist + `get_filter_width=[0x1A,0x03]` in toml, but `filter_width` is not a declared capability → `filter_width.set` check skipped entirely	Declare `filter_width` cap once the segmented BCD index tables (Table 2-4) are confirmed on hardware; the profile comment already flags this as deferred pending HW. Until then: keep as-is.	NEW (low)
`0x14 0x0A` TX power · `0x14 0x0B` mic gain · `0x14 0x09` CW pitch · `0x14 0x15` MONI	methods exist, opcodes documented & in toml, but no RMVR registry check for `rf_power`/`mic_gain`/`cw_pitch`/`monitor`	Optional: add level RMVR checks (these are clean read/set BCD-255 pairs). Low — operator-visible but already wired.	NEW (low)
`0x16 0x40` NR on/off · `0x16 0x44` COMP on/off	`set_nr`/`set_compressor` use exactly the documented opcodes; `compressor` is presence-only and `nr.set` covers NR	Add `compressor.set` RMVR (`get/set_compressor`, opcode is correct for X6200).	NEW (med)

B. VALIDATION GAPS — implemented + declared + opcode-correct, but presence-only (no round-trip)¶

CAT (X6200)	rigplane symbol	Why presence-only	Ticket
`0x16 0x44` COMP	`get/set_compressor`	`compressor` cap has no registry check → `compressor.presence`	NEW (med) — opcode is X6200-correct, safe to RMVR
`0x14 0x08/0x09` PBT	`get/set_pbt_inner`/`_outer`	`pbt` cap has no registry check; opcodes are generic-Icom and not in the X6200 table → presence masks a likely-unsupported control	NEW (med) — verify PBT on HW before promoting; may belong in C/D
`0x1A 0x06` data mode	`get/set_data_mode`	`data_mode` cap has no registry check → presence	NEW (low)
(n/a) scan	—	`scan` cap declared, no method, no CI-V	NEW (low) — drop cap or implement

C. MISSING BACKEND — X6200-documented operator command, no (correct) backend method¶

CAT (X6200)	Function	Value	Ticket
`0x16 0x41` DNF on/off	Digital notch toggle — the X6200's real notch. No method binds it; `notch` cap mis-binds to `0x16 0x48` (manual notch, Icom-only).	Med — primary noise-reduction control	NEW (med) → see D
`0x14 0x0D` DNF centre freq	Notch FC 100-3000 Hz. `get/set_notch_filter` exists at `0x14 0x0D` (correct!) but is not surfaced via the `notch` cap path.	Low	NEW (low)
`0x1A 0x05` (00 62) key/LCD LOCK	Distinct from `0x16 0x50` dial lock; documented separately.	Low — overlaps dial_lock	NEW (low)
`0x1C 0x01 0x02` ATU auto-tune	`set_tuner_status` likely only sends on/off; the `0x02` auto-tune action value may be unreached.	Med — verify auto-tune value path	NEW (med)

D. MISMATCH / WRONG — declared/checked but the opcode is NOT in the X6200 documented set¶

These are the high-value finds: a capability is declared, a registry RMVR check is generated (and would run on live HW), but the hardcoded backend opcode is one the X6200 firmware does not document → the check would likely time out / NAK, or worse, silently hit an unrelated register. All need a live capture to confirm true behavior.

CAT path	rigplane symbol	Issue	Ticket
tone/CTCSS family	`repeater_tone.set`,`tone_freq.set`,`tsql.set`,`tsql_freq.set` RMVR checks (caps `repeater_tone`,`tsql` declared)	Backend sends `0x16 0x42/0x43` + `0x1B 0x00`. None are in the X6200 table — no `0x1B`, no tone funcs. 4 RMVR checks point at undocumented opcodes.	NEW (high) — drop `repeater_tone`/`tsql` caps from `x6200.toml` (or gate behind HW confirm)
manual notch	`notch.set` RMVR (cap `notch`)	Routes to `set_manual_notch` = `0x16 0x48` (Icom manual notch). X6200's notch is DNF `0x16 0x41` + FC `0x14 0x0D`. Check targets the wrong register.	NEW (high) — repoint `notch` to DNF `0x16 0x41`; add `get/set` DNF methods
VOX	`vox.read`/`vox.set`/`vox_gain.set` (cap `vox`)	`get/set_vox` = `0x16 0x46`, vox-gain `0x14 0x16` — neither documented for X6200.	NEW (med) — drop `vox` cap, or confirm on HW
break-in	`break_in` cap (presence-only)	`get/set_break_in` = `0x16 0x47` (Icom CW BK-IN mode). X6200 documents only QSK hang time `0x14 0x0F`, not a BK-IN-mode register.	NEW (med) — remap to QSK-hang or drop cap
RIT/XIT	`rit.set`,`xit.set` write-only checks (caps `rit`,`xit`)	`0x21` RIT/XIT — not in the X6200 table. Already partly hedged via `[validation] write_only_controls=["rit","xit","notch"]` ("SET accepted but GET times out", MOR-179/207 — those were raw-CIV-confirmed on other Icoms, not X6200). For X6200 the SET itself is undocumented.	NEW (med) — confirm `0x21` on X6200 HW; if absent, drop `rit`/`xit` caps
legacy mode `0x04`/`0x06`	`get_mode`/`set_mode` main path (hardcoded `0x04/0x06`)	X6200 table documents mode only via `0x26`. Main RX mode read/write may not answer on `0x04/0x06`. The dual-RX `0x26` path is correct; the legacy main path is the risk.	NEW (high) — confirm `0x04` responds on HW; if not, route main mode through `0x26 0x00`
legacy freq `0x03`/`0x05`	`get_freq`/`set_freq` main path (hardcoded `0x03/0x05`)	`0x03` appears only in the page-4 narrative example, not Table 1. Likely works (example is explicit) but unconfirmed; `0x25 0x00` is the table-blessed path.	NEW (med) — confirm; prefer `0x25 0x00` for main if `0x03` is flaky
`pbt`	`get/set_pbt_inner`/`_outer` (`0x14 0x08/0x09`)	Generic-Icom PBT opcodes, absent from X6200 table; cap declared with `[controls.pbt_inner/outer]` ranges.	NEW (med) — confirm PBT on HW; drop cap if unsupported

Correctly handled (no gap)¶

0x18 power on/off — intentionally omitted from x6200.toml (MOR-170 wedge). power_control cap undeclared → no powerstat path. ✅
0x27 scope — scope cap undeclared; all scope.* checks resolve UNSUPPORTED. Matches doc (no scope CI-V). ✅
NB 0x16 0x22, NR 0x16 0x40, AGC 0x16 0x12, preamp 0x16 0x02, ATT 0x11, AF/RF/SQL 0x14 0x01/02/03, key-lock 0x16 0x50, meters 0x15, PTT/ATU 0x1C — opcodes match the X6200 table; checks are sound. ✅

Documentation quality note¶

The Radioddity V1.0.6 PDF is the single authoritative source and is high quality for what it covers: every opcode above marked ✅ is verbatim from Table 1. The audit's confidence is therefore high on presence/absence of opcodes but the following need a live X6200 capture before acting on the D-list:

Legacy 0x03/0x04 (main freq/mode): 0x03 is narrative-only; 0x04 is fully absent. Does the firmware still answer them (Icom-compat), or only 0x25/0x26? This decides whether the legacy main paths are silently broken.
0x16 0x41 DNF vs 0x16 0x48 manual-notch: confirm the X6200 NAKs 0x48 (so the notch.set check is a real failure, not a coincidental hit).
Tone/CTCSS, VOX, break-in, RIT/XIT (0x1B, 0x16 0x42/43/46/47, 0x21): confirmed doc-absent; a live NAK capture would let us safely drop the four caps (repeater_tone, tsql, vox, plus break_in/rit/xit review) from x6200.toml.
PBT 0x14 0x08/0x09: doc-absent but cap-declared with full control ranges — most likely a copied-from-IC-705 over-declaration.
0x1A 0x03 filter width + 0x1C 0x01 0x02 ATU auto-tune: documented but un-surfaced / partially wired; capture to confirm the segmented BCD index and the auto-tune action value.

Headline: the X6200 profile is over-declared — it inherited several IC-705-class capabilities (repeater_tone, tsql, vox, break_in, rit, xit, notch→manual, pbt) whose hardcoded backend opcodes are not in the X6200 CI-V table. On live hardware these generate RMVR checks that would target undocumented registers. The fix is profile-side (drop/remap caps) once a single live capture confirms the NAKs — no backend changes are needed for the correctly-wired majority.