Vantage 2.0 m deck

[mbutawan]

VantageDeck(size=2.0) raises NotImplementedError. This PR adds the 2.0 m deck, and makes VantageDeck serialization round-trip.

2.0 m deck

• 84 tracks on the same 22.5 mm raster as the 1.3 m deck (size_x 1912.5, size_y 653.5). The instrument reports 84 tracks in its machine configuration.
• The trash sits at the waste position a 2.0 m Vantage uses when initializing its pipetting channels, captured from a VENUS initialization on the instrument: A1PMDI xp13845 & yp3891 3623 3355 3087 2819 2551 2283 2016 tp2450 & tz1235 ..., i.e. eject x 1384.5 mm, tip-end deposit z 123.5 mm. The trash's y footprint is the 1.3 m deck's; the eight channels eject between y 201.6 and 389.1, inside it.
• For reference, the same initialization ejects the 96 head at A1HMDI xp14089 yp2721 tz2175 te2450. Not modelled here, since the Vantage backend only drops 96 tips into TipRacks.

Serialization round trip

Deck.deserialize(deck.serialize()) raises TypeError for every VantageDeck, 1.3 m included: serialize() keeps num_tracks, with_trash96 and core_grippers from HamiltonDeck.serialize(), which VantageDeck.__init__ does not take, and __init__ cannot take back the model a 54-track deck is serialized with. serialize() now drops those keys and __init__ passes model through, as STARDeck does.

Testing

• vantage_decks_tests.py: geometry and track positions for both sizes, trash position, unknown sizes refused, serialize round trip for both sizes.
• Values confirmed against VENUS/VStarTranslator traces from a Microlab Vantage 2.0 (Venus on Vantage 4.5.0.5217; 8× 1000 µL channels, 2× 5 mL channels, CO-RE 96 head, autoload, gripper arm). No PLR-driven hardware run yet — we are validating three 2.0 m instruments with PLR and will report back; draft until then.

🤖 Generated with Claude Code

[rickwierenga]

do you have a vantage that you tested this with?

[rickwierenga]

we do not need tests for declarations like this